Dive into the transformative world of AI with hosts Matt and Scott on the Power Time Show. This episode explores AI's role in revolutionizing various life aspects and highlights tech innovations like Microsoft's Wave 2 and Project Sophia. Special guest Adam Gernon from Microsoft discusses the practical applications of AI, emphasizing Copilot's impact on business and addressing job replacement concerns. The conversation also covers the integration of AI in workplaces, offering insights on using AI tools effectively and anticipating future job interviews. A must-listen for those looking to leverage AI in their work.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Hi everyone, welcome to the Power Time Show with Matt and Scott. And on today's show...
You know, shiny toys, it's a bit of a common theme in sort of where I work at
the moment, that I like all the new things that are coming through and I have to get stuck in.
There isn't a problem that we've found so far that can't be solved with the full stack.
But there was a journey we all needed to go on.
(00:22):
To me, that's when we'll hit, if you will, the holy grail.
I'm sure Adam's basically a celebrity now. The answer, no ads, just get an answer.
Yep. And absolutely loving that app at the moment. Matt's just sitting there downloading an app.
It's power time. Get ready, let's go. Unlocking secrets leaders should know.
(00:44):
Music.
Hi, everyone. Welcome to Power Time. I'm joined here with the sweet prince,
a.k.a. Matt Noble, to have a bit of a chat about Copilot and AI today.
(01:05):
And we've got a very special guest today joining us later in the show.
But before we get into that, Matt, how has your week been?
It's been a very exciting couple of weeks in AI.
And how have you been going? Busy?
Yeah, I've been very busy. I'm not too sure about this nickname going a bit viral as well.
But I have to thank our prior guest, John Kelly, for, I guess,
sharing some of our little secrets.
(01:27):
It. So yeah, I guess that's out in the wilderness now, some would say.
And like I say, very interesting in AI. So much stuff is happening.
I know you're always spamming me with all these sort of crazy AI things.
I think I sent you one last night about how you could turn a photo into actual
recording and presenting based off that. So there's so many cool things happening.
Like I said, looking forward to unpacking it today on this episode.
So yeah, let's start on that. So Scott, you sent me something about Udio.
(01:50):
And you also mentioned and you might be creating our new Udeo jingle for our
preamble for our podcast.
Can you explain that to firstly me and then also our guests on the show?
Yeah, sure. I was listening to a different podcast, actually,
and they spoke about UDIO.
UDIO.com is the URL, and we'll put that in the chat notes.
(02:11):
But basically, it creates songs, right?
So this is nothing new, and there's quite a few sites that do this.
But what is new is it creates music in a way that it sounds very realistic,
and you can actually do full musicals, so just from a prompt.
So if you get really good with your prompt engineering, Darian can create some
(02:31):
amazing intros. And I thought, oh, well, on this episode, I mean,
we use AI for everything, right, Max?
I use it for transcribing when all the mastering for this podcast is done through
AI. There's a fair bit of AI going on already.
And we use a different tool to be able to do the power time jingle.
But UDI looks really, really good and really promising.
And it does blow my mind to think that you can do a simple jingle.
(02:54):
And I mean, I created a song for my daughter and maybe I'll play that as well
in a few seconds and tell me what you think.
But I created a song about our
dog Cookie and I just thought it was phenomenal. So let me roll it now.
Music.
Play, play, play away Play Ivy with Cookie in the sun today They'll laugh and bark in parks so bright.
(03:41):
That's amazing. So how did you go about creating that? You just put in a couple
of sentences or anything like that?
And how easy would someone have access to doing that? Like, what did you do to get started?
And basically, how can I play around with it? Because again,
and I'm a bit of a curious soul. I like playing around with, you know, shiny toys.
It's a bit of a common theme in sort of where I work at the moment that I like
all the new things that are coming through and I have to get stuck in and very
(04:03):
curious. So when it comes to that, but talk to me about how you actually started creating that.
It was pretty easy. I just created a prompt and then I went into ChatTube BT
actually and just said, create a song.
And then I used that song after it created the music to add to the chorus lines inside the tool.
(04:24):
And then it just basically created the entire song for me. So just amazing result.
I couldn't believe how this is an AI using AI to create stuff.
And it reminds me, we've set up a project within my team at Microsoft called
Hearts and Minds and Hearts and Minds was really there to be able to showcase
to customers using different techniques with audio and video,
(04:48):
how our products really will fit the customer.
So we'll go talk to a customer, we'll do discovery, then we'll grab that content
and then we'll turn that into something that's really snackable for them.
And we've had tremendous success. We've had CEOs of major four banks look at these things.
And for us, because we're a smaller part of Microsoft, it gives us a lot of
(05:12):
reach into really seeing people in different accounts. And it's been a great pre-sales tool.
But the thing is, if you think about when I started at Microsoft seven years ago,
the way we would do this, and we still had that concept, but the way we would
do it is we'd go see a third-party company that charge a lot of money to go
and create this content for customers and it'd be like a three to six month process,
(05:33):
and it'd be really onerous and in the end, you kind of lost what you're doing now, right?
Now, because we've sort of, we're dancing to it three or four years ago,
the team was able to create some really good assets three or four years ago
that we were able to turn out in probably about two months of just work and
duration and all this kind of stuff without the use of any other third-party companies.
Now we fast forward on three years later and with all the tooling that we've
(05:56):
got around AI and the video editing stuff that you can access,
we can turn out these videos in about three or four days.
And the quality is better than some of the videos coming out of corp for us.
So, and we can just do that locally within my team. I mean, it does help that
somebody in my team, Manny has a PhD in video editing.
That does help, I will say. And we have a sound engineer as well.
(06:19):
So we do have a bit of an advantage there. But having said that,
you know, the quality of this material is excellent.
I wish I could share it, but it's really specific to customers. So we can't.
But tools like UDIO are going to make a huge difference moving forward in marketing
and things like that. So I was really impressed when I saw the quality of this.
And another thing that really dropped this week was, you know,
(06:43):
looking at Microsoft Research and this Vassar One.
I don't know if anyone's seen Vassar One, but the ability to be able to put
emotions in a virtual avatar, it's just, it's been fantastic.
Yeah, yeah, yeah. So just, if anyone wants to Google Vassar One,
so pretty much Microsoft Research is working on realistic, avatars look like
(07:05):
people and have real emotion and things like that.
It just really, it's about half a dozen sort of, you know, previews.
And it blew my mind as well.
And, you know, I think, I really think it's not going to be long before it's
going to be very difficult to distinguish between kind of a virtual avatar and
a real person, as good and bad as that sounds.
(07:28):
And, you know, when we started talking about this, obviously use case,
contact centers, all these kind of things.
So for all of us who are leading teams out there, pretty soon you'll be leading
an AI team and a group of people that people will be monitoring this AI team
as well, which is a really interesting concept.
And everyone goes on to this topic of jobs, right, when you talk about AI.
(07:50):
And, you know, Matt, you and I were at Steve Bartlett's session the other Friday,
and, you know, he spoke about if you look at every technology curve,
you know, you're going to have a lot of people
It doesn't matter how you look at it, how you cut it, it always increases a huge amount of jobs.
And the topic at the moment is if I've got an AI-assisted bot that kind of knows
(08:12):
all this stuff, you know, that
opens the scope to the jobs that I could do that I couldn't do before.
So, you know, maybe highly skilled jobs that are around today,
maybe you can do like a two-year course and have an assistant that kind of does
that heavy lifting for you.
And a lot more people can be GPs or a lot more people can be pathologists and
(08:34):
these kind of roles, right?
So it's really going to change the way we work and I think in a really positive
way. It's easy to go to the negative.
But let's face it, no one likes working in a contact center sometimes.
It can be rough and customers quite often don't like calling into them.
So if we can modernize and make it a better experience for customers,
I think it's a great thing. I think it is. Yeah, I completely agree.
(08:55):
And similar to yourself, we're looking at all these different technologies and
all these different use cases. And again, some of the things we're doing is
really groundbreaking and really exciting.
And similar to your sentiment regarding Steve Bartlett, that was fantastic.
We got a lot out of that together. But again, his big message was really about
embracing AI, getting to learn to work with AI.
And that really is sort of that next tech sort of forefront that we're going
(09:16):
to be embarking on together.
And again, the whole concept of this is not going to take your job,
learn how to embrace it, learn how to use it, and then really maximize your
efficiency on your day-to-day role within your sort of workplace. place.
So on the AI topic, I know Wave 2 is coming out soon, Scott.
I'm wondering if you've seen that and whether you can potentially talk at some
(09:37):
level without giving away too much about what's coming on with that.
Yeah. So for those listening to the show that don't know how,
particularly the leaders that don't know how Microsoft releases,
I guess, feature updates, it's usually two waves across our financial year.
We have Wave 1, Wave 2, the kind of April, May, September, October over the time frame.
(09:57):
I mean, with all the features that are coming out, you could say it's continuous,
except for Christmas and the financial.
But that's how it works. And we publish the waves, and these waves started off,
I remember a few years ago, like two or three pages. Now they're 150.
So the acceleration and innovation has really been tremendous.
But that's the time where we really stand up and go, because a lot of things
(10:18):
get up and they don't get funded.
The rule is if they don't make it into the wave documents, then they're not
going to go to market. So, we keep an eye on those documents to really make
sure what product is coming and is actually going to land in product.
And my team analyzes it all the time. And, of course, we get early access to it before anyone else.
So, I can't – there's a lot of stuff I can't comment on. Come on. This is what I will say.
(10:43):
I can't believe how we've created AI-enabled products and put them to product so fast. Yeah.
This is a standout in this wave.
And there was so many things that I just went, wow, that's amazing.
That's going to change the life of this person. Or that's going to make a big
(11:06):
difference to this admin.
Or this person is going to be so much more effective. or that messaging is really
fantastic around that, right?
So I think we'll get into this. As soon as I'm allowed to talk about it,
we'll get into it a bit deeper.
But, yeah, I've never been so excited about a technology wave before than seeing
what's going into our product and how people are using it.
(11:28):
It's kind of like, you know, we've got a candy store and the kids have now gone
in and just gone, all right, what can I, you know, or a Lego store,
more like a Lego store, what can I build?
Now I've got all this Lego available to me and we're just seeing these pots
of innovation come out of everywhere, which is really interesting.
So I think Project Sophia as well, you know, that's... Oof. Yeah.
(11:51):
We could spend a whole podcast talking about that. That's very exciting.
Maybe we should. Maybe that's one of the next topics we'll talk about,
Scott. I think that's very interesting and that's going to basically unpack
and hopefully bridge a lot of this stuff together.
Yeah, I think I've got the perfect guest to come on, Greg Nash,
who I talk with regularly, specializing in data and AI.
Let's bring Greg on to really talk about Project Sophia and get his thoughts
(12:13):
because he's deep in the weeds on this stuff all the time.
But to put it in perspective, I reckon I saw this week on our internal calls
probably about six or seven different projects like Project Sophia that are
going to really change the way that we work.
So everything's got a project name, as you know.
But yeah, that was really exciting stuff. So all these projects are great.
(12:37):
But that one in particular, I think everyone's excited about to see.
And for those that haven't seen it, you can go on and we'll put links again in the chat.
But effectively, instead of constructing a,
I guess, using Power BI and constructing a data visualization now,
we just use natural language to go and prompt data and it creates the visual
(12:57):
that suits the data, if that makes sense.
So hey, grab me this and I'll just go and do it. and then you can click on it
and go, can you change this from a pie graph to a bar graph and so on and so forth.
So it's really a conversation with your data. I really like it. It's good.
Yeah, there's just so much cool things happening at the moment and there's so
many new sort of technology options and streams and all of this kind of stuff.
(13:18):
And I agree with you when you mentioned before about going fast.
I remember I caught up with Aaron Bjork, I think a couple of weeks ago in Sydney
and he was talking about the Microsoft strategy about going fast,
going fast with this and making sure they can roll out as much,
as quickly as possible and embed this into their products. And again,
from a partner standpoint, I've never.
Seen anything get adopted and pushed out so fast. And it's really amazing.
(13:39):
And it's really incredible. And I guess a lot of the learnings we're seeing
as part of that sort of initial push is really beneficial to Microsoft as part of this.
And the amount of use cases we've seen in the sort of last couple of weeks is really incredible.
Again, we caught up with Dwayne Robinson, the co-pilot, I think the product owner of that.
And again, that was a phenomenal session. I really enjoyed that.
And again, we're seeing a lot of really fantastic co-pilot studio examples come
(14:03):
come out, particularly the last couple of weeks as customers get more familiar
with it. I think that's really exciting.
And again, that's a really good segue into, I guess, maybe our guest for today.
And Scott, maybe you want to talk about Adam in a bit more detail and kind of
what he does within Microsoft.
Yeah. So I've been working with Adam for a long time now.
And Adam took on the project of spearheading, I guess, enterprise program for
(14:25):
early adopters for CodePilot and program.
And he works with a lot of customers to be able to drive that adoption.
So he was Just kind of in it from the beginning, Adam's done a lot of different
podcasts and YouTube sort of videos. He's shifted on with Lisa Crosby recently.
It was good to have him on the show and really have a bit of a chat to him about
where he sees the practical use cases for Copilot.
(14:46):
And look, really timely for us as well because Copilot's getting GPT 4.5 in April.
So that means it'll go multimodal, which opens a whole bunch of more use cases
right there for the product.
And also our co-pilot for sales in the biz app side of things and co-pilot service
is starting to get really mature now.
(15:08):
So effectively, we have, I mean, to demystify it a little bit,
we have three co-pilots in my world that really make sense.
We have co-pilot studio, which is, you know, how to create a chatbot and really
build upon. And that is super popular, like we spoke about at the AI tour.
We have co-pilot for service, which works across Dynamics products as well as
Salesforce, ServiceNow, and others.
(15:29):
And then we have Co-Pilot for Sales, which does the same, works across similar products.
So, and, you know, they're really there to optimise the people that use those
products, the salesperson or a customer service agent or whatnot.
Yeah, I agree. And Adam, similar to your sentiment, Adam's been doing almost a road show.
So if you see a little co-pilot booth and you're located at one of the AI events,
(15:50):
normally Adam will be kind of running the show and he knows his stuff,
obviously. He's a really good guy.
And also, I actually got a lot of benefit out of the conversation because I
found all these great examples that I hadn't heard about, about utilizing M365
co-pilot. So thanks again, Adam, for saving me about maybe four to five hours
a week and on the road to 20.
That's all I want to say. I really want to leverage AI and I'm going to be picking
(16:13):
Adam's brain consistently to give me all these new shiny examples of how we can use Copilot.
So yeah, we really hope the listeners enjoy that chat and get a lot out of it.
I actually downloaded a mobile application, which Adam told me to on the podcast
during the podcast again.
So you'll hear that come up and I recommend everyone also downloading that,
but I won't give that away.
Way we want the listeners to listen to it and then obviously get the benefit
(16:34):
of that so yeah awesome all right well we might take a short break matt and
then we'll bring on adam onto the show sounds good.
Music.
(16:54):
Thanks, Matt, and good to be here.
Thanks, Scotty, for inviting me. Yeah, look, my background is customer,
project, program, portfolio, delivery, then moving into Microsoft,
really focused on the adoption and change as a CSM, customer success manager from our work.
(17:14):
Because what I've realized through delivery was it doesn't matter how well you
do a project from a project triangle perspective, you know, scope,
scale, and budget, it really is about our people getting the value from it.
So being able to drive the value from your investment is where I then focused on that.
Matt and I had a great experience working together on Parkland,
(17:35):
really seeing it not from the Microsoft helicopter or 30,000 feet there,
but boots on the ground in terms of just the challenges and the opportunities
that people have when adopting technology.
And then, yeah, then came back to Microsoft and had.
I had a great experience over the last nine months taking all of our early access
program or the private preview customers in enterprise commercial through the M365 co-pilot,
(18:01):
which if you had asked me back in October about this, I probably would have
said to you a very different answer,
but very grateful now to be able to reflect back on just what it was like to ride the crest of a wave.
That is co-pilot and I look forward to sharing more about that today.
So maybe let's click a little bit more into that early access program because
(18:21):
effectively you were the first sort of group of people to really take Copilot to market.
How was that? How did you find that? I mean, no one knew what it was.
And here's this program that we have to talk about with customers.
Yeah, look, without revealing too much, it was the closest to working in partner
land that I've ever had at Microsoft because I moved out of being in the field into Corp.
(18:42):
And we were part of the CSAs globally that were looking after any single early
access program participants across the globe.
Initially, we were looking at around 600. I think it ended up being around 1300.
And it was, how do we take them through the journey of phone purchase through
(19:04):
to establishing the right governance framework, framework,
making sure that the adoption was strong so that you can drive towards business
value and then how do you capture that business value for something so new to
be able to feed into your business days.
Very much like in partner land, you might have put in a strategy for a customer
and then they say, no, can you execute to value?
(19:25):
But there was a journey we all needed to go on. So we fast forward kind of a
year on, I guess, from that point,
customers had a chance to user what's the
feedback on office copilot there the
answer is like you would expect and it's
kind of a parallel to your adoption to curve for technology
(19:46):
there are still customers that are part of the emp that have really started
are still stuck at their security governance compliance risk management stage
and then there are those that are uh that have customers i got an email today
with one saying don't try and take my co-pilot license from me to their project team
adoption team so the answer really comes down to how you've ran it and some of these are not,
(20:13):
you know ai or generative ai related it's basic program and change management around,
drawing sponsorship giving people the support capacity to play to learn to grow to experiment,
because ultimately the biggest change with co-pilot is moving from let's call
it that kind of waterfall deterministic thinking of try to pass the LES no,
(20:37):
co-founder doesn't work to.
More of an agile kind of way of thinking, experimentation, feedback,
come back mindset, try it later.
And those organizations that did that well, typically partnering with an innovation
function that was driving and supporting them, really embraced that approach
and supported their users to get to capturing alpha.
(21:01):
Really something I'm really interested in having a chat because in the introduction
before you kind of alluded to the way that you really look at making sure the
organization get value from the product.
And it's something that I'm working with you for a period of time definitely
resonated with. I think both of us, and that's why we've always had a really good synergy.
But talk to me about how organizations specifically are getting value from the
various co-pilot suites of products because there are a number of them.
(21:23):
There's hundreds of co-pilots.
So maybe talk to me a bit about your definition of co-pilot and maybe some examples
of the co-pilot business value returns that you're seeing from your customers
that you've been working with.
Yeah. Microsoft marketing don't always make it easy for customers.
If I aggregate it up to a higher level, the co-pilot brand is about having an
(21:44):
assistant in whatever you're working in, a digital assistant with you, whatever you're doing.
So for example, co-pilot for the web, aka Bing Chat, Bing Chat Enterprise, etc.
Is a digital assistant across finding and generating information or whatever
you need to do that's based on the course of information that's available out in the world one way.
(22:06):
So it really comes down to helping you where you are, whatever you're doing.
And that might be, you'll use different co-pilots depending on your role.
You might be, for example, in service, customer service,
and I've got some great examples of where they have a lot of value out of M365
co-pilot, but obviously think about Dynamics or Sales co-pilot,
(22:27):
and also those extensibility players into your world around Power Platform and Co-pilot Studio.
Because we don't just do one thing, like we don't just have effective meetings.
We also then typically have to communicate out about that. We also have to convert
that into maybe different languages, different audiences, different segments
(22:48):
of who we're targeting with this message, and we'll also need to capture and measure all that.
Not everything will be done out of the box with CodeFiler Frame 365,
but what you'll find is that there isn't a problem that we've found so far that
can't be solved with the full stack, whether it's Copilot extensibility into
low code or into to Procode and the whole Copilot stack.
(23:09):
It's really interesting, Adam, because I know my team's been doing a lot of
work at the moment around MTC,
so Microsoft Technology Center sessions, where we can connect Copilot with Azure
to grab customers' data and then turn that into actionable outcomes and having
a lot of success around that.
So I think we can see the platform now is really coming mean together,
(23:35):
and particularly around guardrails.
So, what's your feedback around, you know, governance and guardrails and things
like that that you're seeing in the platform and being able to support CodePol?
I think it's a unique proposition from us. Yeah, absolutely.
I think this has been like the eternal kind of fight between good and evil, if you will,
between security versus usability and the slider of where do we actually accept
(24:01):
risk risk and innovate, or lock down and manage risk to zero.
And so for me, it's about having the right sponsorship and awareness of really
business risk management rather than trying to manage risk to zero.
So obviously what we're able to do in your world to set up those guardrails
(24:22):
so people have that experience of being able to experiment within guardrails
is critical to innovation.
How do we have that right balance and having
the right setup to support devs in
a low-code way can only accelerate development but
i think we miss a trick when we don't then take that to really the end users
(24:45):
and to me that's when one hit if you will the holy grail of how gen i can really
be enabled because think about the likes of well one of our own now nathan backers and what Well,
he was able to do a Telstra by having the SME knowledge to be able to implement that.
That's the opportunity across all of the Copilot world and universe and what
(25:08):
NAI can do with people that are in the Microsoft stack.
But we have to give them the opportunity to do it.
So Adam, you've been running a lot of, I guess, you've been on the conference tour, I've seen.
And for anyone that has been looking at the Microsoft socials,
Adam's basically a celebrity now.
There's lots of videos going on where Adam's talking to customers.
What I'm interested to hear is what type of questions you've been getting from
(25:32):
customers during your basically your conference road tour that you've been going on.
And I guess how have you been directing and giving advice and all those kind of things? user.
And will AI take my job, for example?
And through the Early Access program where we are working with those 300 users
that had a license to be able to do this, my answer was, okay,
(25:53):
I may not take your job, but people that can use AI will.
So this is your opportunity to really get and build the muscle and the skills
to be able to interact in a new way, in an agile experimentation model with
learning a new language?
How do you talk to a generator of AI, Copilot, whatever Copilot you're using?
(26:15):
And also, how do you go on that change manager journey?
Because personally, I was frustrated as hell sometimes with Copilot,
but candidly, Copilot didn't care.
It was only me that got frustrated. So how did I then have to manage myself,
to grow, to learn how to use Copilot? You've got to be nice.
What was that about? You've got to be nice to AI, so yeah.
(26:39):
Absolutely. And the answer to your question comes down to who I was talking
to, what they did, and what they were doing, and where they were taking a value-based lens.
So the best engagement I had with COOs typically that were taking a full lens
across the business, rather than just a cost view, or this view,
or that view, or a personal risk view.
(27:02):
I think customers are still kind of kind of trying to
work out you know where this all fits and and
how how that can be more productive where's the
return on investment all of these things yeah copilot's a
licensed option so maybe just talk us through
how people get started with office co-pilots and
and you know what what you're seeing how people
(27:23):
are purchasing this and what are the kind of all the quiet yeah absolutely
and look there's been a lot of learning certainly at the enterprise level
that and i also have a lot of friends of
mine that are in smc space or smb space as
a microsoft would look at small to medium businesses or across
that whole kind of lens and given the
majority of australian businesses are more in the
(27:45):
are not enterprise it is amazing how much value there is that space for co-pilot
and especially now we've opened it up to anyone one effectively with basic business
premium license because one of my friends we tested this with during EAP was RFP responses.
(28:05):
It consumed so much time, Matt.
You and I have had the joy of that on both sides of the fence, right?
It makes me cringe. I saw you shiver.
And so how do you actually run an RFP process to generate that on the outside?
How do you then on the other side of the fence take your PowerPoint presentation
that you need to now present based on the submission that you had.
(28:30):
That's kind of the basics, like just drafting content creation and co-file. I'm working with it.
It's how do you then kind of build on top of that? How do you take users towards what else can it do?
And it's that proven model of the longer the adoption, the stronger the business
value is generated because you have more time to think about what can it do for me?
(28:50):
Apply it to what I do. And so, like, if I take that RFP example,
can you now compare this response with this response? What's it missing?
Or these are the key criteria, co-pilot, this is my response,
have I addressed all of the key criteria?
In the past, you'd typically have, Matt would produce something,
(29:13):
he'd send it to IC to review before you then would send it out more broadly,
co-pilot is now your first reviewer.
So we're missing a trick around productivity, productivity, productivity,
but what about quality? How do we measure that?
Because Copilot can review your document.
Does this conclusion match their body? Does it have a call to action?
(29:36):
Copilot, can you help me to generate an inspirational speech based on my annual
report and reference that file?
Include some famous quotes tied to this content.
Oh, and by the way, there's a joke about AI. Could you also include a funny quid.
It's that kind of using it in how you do your job.
And the last thing that I typically do when I'm talking to an exec is also,
(29:57):
so how did you go in your annual?
So when you presented this, how did you go?
Did you watch the video? Nah. What did your team say? Oh, they said it was great.
Of course they did. Did you ask about it?
Copilot won't actually care about your role. It'll give you the true reflection.
So when I ran one of those sessions that you were talking about,
(30:18):
I asked Copilot afterwards, how did I present? Did the message land?
Did it sound like I was reading the slides? You know, you can get that feedback
because it's independent without any fear of it losing its job and managing up.
I love the RFP example, and I love deep thinking through these things, right?
(30:42):
So if you think about an RFP process, we've got Copilot creating the RFP.
We've got Copilot responding with the RFP.
We've got Teams calls throughout the RFP process that are summarized by Copilot,
so people don't actually need to attend the meetings.
So where does that leave, when you think about it, where does it leave pre-sales
for us and our engagement with our customers? Matt, you probably got some comments
(31:05):
around that as well, given that you guys are from the services point of view,
right? That's another story altogether.
Yeah. I just want to thank Adam for saving me 20 hours a week, based off that .
I think every other partner that's listening is of a similar sentiment to myself,
because that's a fantastic use case. It's a common pattern.
We should be using AI and co-pilot to do those repeatable tasks.
(31:29):
Again, thanks Adam for sharing that. And I want to know if there are any other
sort of specific examples that you are utilizing Copilot for considering your experience.
Are there sort of some quick tips, quick tricks?
We do Teams meeting notification summaries, all that kind of stuff.
But I wonder from the master, what other things is Copilot doing that I should
(31:49):
be doing and the listeners should be doing at home?
Unfortunately, Matt, Copilot can't help you with what you shouldn't be doing.
So we just focus on what you still need to do.
How's that for you? You can ask, but I don't know.
Yeah, look, I think it's really interesting because I try and tell the stories
(32:09):
with users so it resonates with them in a storytelling fashion.
So for me, we have, and I think every Every customer of ours at this stage has seen Project Falcon.
And it's a demo that shows how you can capture the sentiment of the meeting,
understand the decision that was made, and save yourself an hour.
(32:30):
And in three minutes, you can take that information and understand the pros
and cons for the decision to delay the go-to-go by two weeks.
But then how do you take that and turn that into what you do?
So I was a program manager. So how do you then use GoPilot to take that information
and put it into your steer code deck?
And can you turn it into an audience for an exec?
(32:53):
So be more sustained and rewrite the language for me without having to manually
do that, without having to go back through the tape multiple times and capture that.
Then I now need to update the project schedule tied to that,
right? Because the milestones, and it's now two weeks delayed.
I can't even remember where that file was. Was it Matt sent me it?
Is it in the project governance folder on Sharecoin? No, it isn't.
(33:17):
And I've got to get this done now.
We spend 27 minutes a day or three weeks a year wasting time on things we know
we have, we just can't find.
Copile. What is the project milestones for project file then? It'll find them for you.
Then, one of the other things that I got burnt on this before is it execs about 400 emails a day.
(33:38):
I was told by one of the EGMs and one of the banks.
How much time they spend in meetings and how much of that they actually read
is very hard right and when i was in
portfolio delivery one of the milestones had been
changed on the project that i was providing an update on and i didn't know but
some of the other people in the room had just found out before the meeting so
i used topo to actually find look in my emails and teams chats over the last
(34:02):
reporting period say two weeks find anything related to project Falcon and it will find,
because I've sent it or someone sent it for me,
an email that will show there was a milestone file was delayed by five days.
So I'm finding things I didn't even know I had. Then don't stop there.
Okay, so Copilot, can you now update my milestone list with the email I just
(34:28):
received and output that in a table so I can copy that into my Stereo Codec
and it'll do that for you.
So it's how do you actually kind of just have the time to play experiment and
see what it can do for you in what you do.
That was kind of one of the let's say the demo scenarios.
You kind of just need to give yourself the time to play to experiment to learn
(34:50):
and grow and I always assume I'm the problem if I'm not getting the output from
Copilot until I kind of get what it is and get where I want to get to.
So there's some of the small examples. The last one I'll give you is I had to
play with Copilot Studio because with all the hype and Copilot focus,
I haven't had any time really to focus on teams and what we used to do in our old world.
(35:14):
And I had to renew my associate certificate.
So I built myself a Copilot Studio.
Pointed it at the learn.microsoft.com URL and did my test where I had the best
imaginable search engine to find everything I needed to answer some of the questions
(35:36):
that might have come up in that test.
It's very interesting that you've opened that can of worms now because I think
we're either going to generate a whole bunch of people that appear really,
really smart and certainly going through interviews and things like that,
people should be really prepared.
If I was interviewing someone for a role in my team, my expectations have now
(35:58):
gone up. You should know who the team are, what the technology is.
You should even know a little bit about Microsoft, the company,
around what its goals are, because it's just a chat, right?
So I think everyone's kind of leveling up now.
You know, this is the new expectation. If you're not on board,
then you're going to be, you're going to struggle, right?
And I think the same goes for tests, the exams all
(36:19):
these sort of things where they're open but when the natural
thing is to use language model for this stuff but suddenly you
know i can see a podcast matt top 10 hacks for copilot or top 10 use cases you
could probably do that every three months and that's gonna build this list of
you know these great ideas because they just keep flooding in all every day
out here to something crazy and we build out like all the csa caption those
(36:42):
scenarios and they're now going to get put
into the scenario library that you get as part of CodePilot.
I know. Because, you know.
Really what you're trying to do is there's the innovators and the people that will,
experiment and there's the you know early adopters and then there's the followers
that we want but they probably won't spend the time to do it themselves they
need to be shown or taught and that's really the key is how do we build up this
(37:04):
library for people to be able to capture.
The value from it so that people can then just apply it and see for themselves and what they do,
and each organization is capturing that because you'll be able to put that into
co-pilot lab and filter on your own organization all those prompts and use cases
that are working really well for what you do.
(37:25):
You know, we spoke about this with Nathan where we were talking about,
you know, when you first join Microsoft, someone sends you an Excel spreadsheet with 200 links.
We're going to kind of figure out how you navigate Microsoft.
And we think the future is probably going to be here's 200 prompts.
Go punch that in the copilot and you'll get a lot more sort of information.
(37:47):
So, you know, maybe there's a Power App there, Matt, a prompt Power App where
you just click on a button, I want to do this, and it just sends a prompt to a language model.
So you've saved me 20 hours and a new Power App idea. Adam, you're welcome to
show it anytime you want, mate.
So that's really good advice as part of it.
I agree with the 10 tips and tricks. As Adam was talking, I was just sitting
(38:09):
there nodding my head because I was like, that's amazing.
Yeah. So, Adam, you're obviously quite mature and using CoPilot for someone
that's looking to kind of level up and start to navigate it.
You kind of mentioned before, just get started.
But what are some other things you can do to think about how to interact with
Copilot? Because I think that's the thing, get started.
But yeah, just some advice from you about navigating it through and having it
(38:29):
as that, you know, basically your assistant to help guide you. So, yeah.
So probably two things there. For M365 Copilot, for any of the adoption people
out there that are trying to drive this and someone's just giving you a few
licenses and you're like, how do I, what do I do with it?
It's really important that people understand that CoViolet, as we talked about,
(38:51):
is everywhere, and it's different in how it will behave, because deliberately
tied to the scope of what you do.
So if I'm running a meeting like this and we have transcription on, I can interact with it.
Matt's asked a question, I didn't quite catch his name. I don't want to be rude.
What was Matt's name? Or Matt's mentioned some stochastic interest rate that
(39:13):
I remember from my uni days, but no idea what that means.
And I don't want to look like the dumb IT guy, quickly ask Cobol.
So remember, there's no dumb questions around what Cobol have been doing for
you and it never gets bored.
It'll never get cranky with you or short tempered because you've asked the same
thing three times. So it is really there for you no matter what you're doing.
That might be sending an email and I myself have ADHD and sometimes I can be
(39:38):
very direct when under pressure and conscious about how the tone might come across an email,
which is not actually how I feel, it's just how that can be written.
So using coaching by a co-pilot to put itself into the shoes of the recipient,
give me feedback on, does this have empathy?
Does this have clarity of the ask?
Those are really important kind of upskilling for me.
(40:03):
Because it also then helps me learn how to write it better the next time and
hopefully less kind of coaching required as we all upskill.
And so you'll see that as a pattern across the tools. It's there to upskill.
Excel will tell you and help you with multiple things, but it'll also explain
the steps so that you can actually learn as you go by osmosis, right?
(40:24):
So I use the analogy of one room at a time or introducing a kitten to a house,
you know, one room at a time.
Do the same at Copilot. that, take people through the adoption around effective
meetings or some sort of hero scenario like finding information,
whatever it is, because if we don't do that, we see a big drop off in that hype
cycle of Gartner, if you're familiar with that.
(40:46):
Very soon people go from, what's this all about, straight into the trough of
disillusionment, wonder what this, like just giving up.
To help with that is how do we actually show them value so they'll stay with
the journey, because it is a journey, it's a change journey.
When Judson was out a few weeks back, a couple of weeks ago,
he said to everyone in the room when he was in New Zealand, I'm thinking you
(41:09):
should use the CodePilot app.
And I was one of those people that hadn't used the CodePilot app before.
I was surprised about it.
Download it now. I've used it every day since. So I think anyone getting started
out there, if they really just want something, then you can just, it's just so useful.
The amount of times I just punch into that CodePilot app before i google or
(41:31):
bing something now it's it's becoming that first place to find things out and
you just get the answer no ad just get an answer yep and we're absolutely loving
that app at the moment matt's just sitting there downloading it actually yeah.
Well hopefully there's like 5 000 people listening to this podcast now like
(41:54):
grabbing that app and downloading.
Maybe time to wrap it up, Matt. I mean, it's been amazing, Adam,
having you on the podcast.
I love your feedback as always and a lot of experience around this,
having run that program and being the first to market.
But thank you for joining the show and chatting about all things CodePilot and
use cases. Matt, any final words?
Yeah, thanks again, Adam. And can CodePilot help my golf game?
(42:17):
Because I think we do a game of golf soon. So let's get onto that.
It can at least help you to set a reminder for the alarm to go off so that you
are there. Yeah, start there.
Anything to tell me next time.
Anyway, we're digging a whole lot of little dark secrets like the sweet prince
and everything from last show. So yeah.
(42:38):
Sweet prince is out there. I think that's how we're starting the next podcast.
But massive thanks, Adam. And thanks for sharing a lot of that.
I got a lot out of it and I can imagine everyone that's listening is part of it.
And I think definitely getting you you off on a list of potential to chat more
about tips and tricks because I can imagine in 12 months, the scenarios we're
(42:59):
going to understand and be utilizing with Cokepilot is going to be phenomenal.
So it's a continually learning, evolving journey for everyone.
So looking forward to continuing the conversation with yourself.
And thanks again for coming on the show.
Cheers, guys. Thank you so much for having me. Thanks, Adam.
Music.
(43:33):
Hey, that's a wrap for today. Thanks so much for everyone for listening to our podcast.
Just a few reminders. So we've set up a LinkedIn page. So we'll endeavor to
put some updates on that LinkedIn page. So go on there and click follow for us.
Also, don't forget to go to powertime.au and subscribe to our newsletter.
And lastly, if you could please click follow on this podcast so you get the
(43:55):
latest episodes as they come out.
We're trying to get the podcasts out every two to three weeks.
So there should be a good cadence with a whole bunch of new guests coming on.
Now, lastly, don't forget that the views on this podcast are not the views of
NCS or Microsoft, but they're the views of Matt and I and our guests.
Thanks again. See you. Can't wait to see you on the next podcast.
Hi everyone, welcome to the Power Time Show with Matt and Scott. And on today's show...
You know, shiny toys, it's a bit of a common theme in sort of where I work at
the moment, that I like all the new things that are coming through and I have to get stuck in.
There isn't a problem that we've found so far that can't be solved with the full stack.
But there was a journey we all needed to go on.
(00:22):
To me, that's when we'll hit, if you will, the holy grail.
I'm sure Adam's basically a celebrity now. The answer, no ads, just get an answer.
Yep. And absolutely loving that app at the moment. Matt's just sitting there downloading an app.
It's power time. Get ready, let's go. Unlocking secrets leaders should know.
(00:44):
Music.
Hi, everyone. Welcome to Power Time. I'm joined here with the sweet prince,
a.k.a. Matt Noble, to have a bit of a chat about Copilot and AI today.
(01:05):
And we've got a very special guest today joining us later in the show.
But before we get into that, Matt, how has your week been?
It's been a very exciting couple of weeks in AI.
And how have you been going? Busy?
Yeah, I've been very busy. I'm not too sure about this nickname going a bit viral as well.
But I have to thank our prior guest, John Kelly, for, I guess,
sharing some of our little secrets.
(01:27):
It. So yeah, I guess that's out in the wilderness now, some would say.
And like I say, very interesting in AI. So much stuff is happening.
I know you're always spamming me with all these sort of crazy AI things.
I think I sent you one last night about how you could turn a photo into actual
recording and presenting based off that. So there's so many cool things happening.
Like I said, looking forward to unpacking it today on this episode.
So yeah, let's start on that. So Scott, you sent me something about Udio.
(01:50):
And you also mentioned and you might be creating our new Udeo jingle for our
preamble for our podcast.
Can you explain that to firstly me and then also our guests on the show?
Yeah, sure. I was listening to a different podcast, actually,
and they spoke about UDIO.
UDIO.com is the URL, and we'll put that in the chat notes.
(02:11):
But basically, it creates songs, right?
So this is nothing new, and there's quite a few sites that do this.
But what is new is it creates music in a way that it sounds very realistic,
and you can actually do full musicals, so just from a prompt.
So if you get really good with your prompt engineering, Darian can create some
(02:31):
amazing intros. And I thought, oh, well, on this episode, I mean,
we use AI for everything, right, Max?
I use it for transcribing when all the mastering for this podcast is done through
AI. There's a fair bit of AI going on already.
And we use a different tool to be able to do the power time jingle.
But UDI looks really, really good and really promising.
And it does blow my mind to think that you can do a simple jingle.
(02:54):
And I mean, I created a song for my daughter and maybe I'll play that as well
in a few seconds and tell me what you think.
But I created a song about our
dog Cookie and I just thought it was phenomenal. So let me roll it now.
Music.
Play, play, play away Play Ivy with Cookie in the sun today They'll laugh and bark in parks so bright.
(03:41):
That's amazing. So how did you go about creating that? You just put in a couple
of sentences or anything like that?
And how easy would someone have access to doing that? Like, what did you do to get started?
And basically, how can I play around with it? Because again,
and I'm a bit of a curious soul. I like playing around with, you know, shiny toys.
It's a bit of a common theme in sort of where I work at the moment that I like
all the new things that are coming through and I have to get stuck in and very
(04:03):
curious. So when it comes to that, but talk to me about how you actually started creating that.
It was pretty easy. I just created a prompt and then I went into ChatTube BT
actually and just said, create a song.
And then I used that song after it created the music to add to the chorus lines inside the tool.
(04:24):
And then it just basically created the entire song for me. So just amazing result.
I couldn't believe how this is an AI using AI to create stuff.
And it reminds me, we've set up a project within my team at Microsoft called
Hearts and Minds and Hearts and Minds was really there to be able to showcase
to customers using different techniques with audio and video,
(04:48):
how our products really will fit the customer.
So we'll go talk to a customer, we'll do discovery, then we'll grab that content
and then we'll turn that into something that's really snackable for them.
And we've had tremendous success. We've had CEOs of major four banks look at these things.
And for us, because we're a smaller part of Microsoft, it gives us a lot of
(05:12):
reach into really seeing people in different accounts. And it's been a great pre-sales tool.
But the thing is, if you think about when I started at Microsoft seven years ago,
the way we would do this, and we still had that concept, but the way we would
do it is we'd go see a third-party company that charge a lot of money to go
and create this content for customers and it'd be like a three to six month process,
(05:33):
and it'd be really onerous and in the end, you kind of lost what you're doing now, right?
Now, because we've sort of, we're dancing to it three or four years ago,
the team was able to create some really good assets three or four years ago
that we were able to turn out in probably about two months of just work and
duration and all this kind of stuff without the use of any other third-party companies.
Now we fast forward on three years later and with all the tooling that we've
(05:56):
got around AI and the video editing stuff that you can access,
we can turn out these videos in about three or four days.
And the quality is better than some of the videos coming out of corp for us.
So, and we can just do that locally within my team. I mean, it does help that
somebody in my team, Manny has a PhD in video editing.
That does help, I will say. And we have a sound engineer as well.
(06:19):
So we do have a bit of an advantage there. But having said that,
you know, the quality of this material is excellent.
I wish I could share it, but it's really specific to customers. So we can't.
But tools like UDIO are going to make a huge difference moving forward in marketing
and things like that. So I was really impressed when I saw the quality of this.
And another thing that really dropped this week was, you know,
(06:43):
looking at Microsoft Research and this Vassar One.
I don't know if anyone's seen Vassar One, but the ability to be able to put
emotions in a virtual avatar, it's just, it's been fantastic.
Yeah, yeah, yeah. So just, if anyone wants to Google Vassar One,
so pretty much Microsoft Research is working on realistic, avatars look like
(07:05):
people and have real emotion and things like that.
It just really, it's about half a dozen sort of, you know, previews.
And it blew my mind as well.
And, you know, I think, I really think it's not going to be long before it's
going to be very difficult to distinguish between kind of a virtual avatar and
a real person, as good and bad as that sounds.
(07:28):
And, you know, when we started talking about this, obviously use case,
contact centers, all these kind of things.
So for all of us who are leading teams out there, pretty soon you'll be leading
an AI team and a group of people that people will be monitoring this AI team
as well, which is a really interesting concept.
And everyone goes on to this topic of jobs, right, when you talk about AI.
(07:50):
And, you know, Matt, you and I were at Steve Bartlett's session the other Friday,
and, you know, he spoke about if you look at every technology curve,
you know, you're going to have a lot of people
It doesn't matter how you look at it, how you cut it, it always increases a huge amount of jobs.
And the topic at the moment is if I've got an AI-assisted bot that kind of knows
(08:12):
all this stuff, you know, that
opens the scope to the jobs that I could do that I couldn't do before.
So, you know, maybe highly skilled jobs that are around today,
maybe you can do like a two-year course and have an assistant that kind of does
that heavy lifting for you.
And a lot more people can be GPs or a lot more people can be pathologists and
(08:34):
these kind of roles, right?
So it's really going to change the way we work and I think in a really positive
way. It's easy to go to the negative.
But let's face it, no one likes working in a contact center sometimes.
It can be rough and customers quite often don't like calling into them.
So if we can modernize and make it a better experience for customers,
I think it's a great thing. I think it is. Yeah, I completely agree.
(08:55):
And similar to yourself, we're looking at all these different technologies and
all these different use cases. And again, some of the things we're doing is
really groundbreaking and really exciting.
And similar to your sentiment regarding Steve Bartlett, that was fantastic.
We got a lot out of that together. But again, his big message was really about
embracing AI, getting to learn to work with AI.
And that really is sort of that next tech sort of forefront that we're going
(09:16):
to be embarking on together.
And again, the whole concept of this is not going to take your job,
learn how to embrace it, learn how to use it, and then really maximize your
efficiency on your day-to-day role within your sort of workplace. place.
So on the AI topic, I know Wave 2 is coming out soon, Scott.
I'm wondering if you've seen that and whether you can potentially talk at some
(09:37):
level without giving away too much about what's coming on with that.
Yeah. So for those listening to the show that don't know how,
particularly the leaders that don't know how Microsoft releases,
I guess, feature updates, it's usually two waves across our financial year.
We have Wave 1, Wave 2, the kind of April, May, September, October over the time frame.
(09:57):
I mean, with all the features that are coming out, you could say it's continuous,
except for Christmas and the financial.
But that's how it works. And we publish the waves, and these waves started off,
I remember a few years ago, like two or three pages. Now they're 150.
So the acceleration and innovation has really been tremendous.
But that's the time where we really stand up and go, because a lot of things
(10:18):
get up and they don't get funded.
The rule is if they don't make it into the wave documents, then they're not
going to go to market. So, we keep an eye on those documents to really make
sure what product is coming and is actually going to land in product.
And my team analyzes it all the time. And, of course, we get early access to it before anyone else.
So, I can't – there's a lot of stuff I can't comment on. Come on. This is what I will say.
(10:43):
I can't believe how we've created AI-enabled products and put them to product so fast. Yeah.
This is a standout in this wave.
And there was so many things that I just went, wow, that's amazing.
That's going to change the life of this person. Or that's going to make a big
(11:06):
difference to this admin.
Or this person is going to be so much more effective. or that messaging is really
fantastic around that, right?
So I think we'll get into this. As soon as I'm allowed to talk about it,
we'll get into it a bit deeper.
But, yeah, I've never been so excited about a technology wave before than seeing
what's going into our product and how people are using it.
(11:28):
It's kind of like, you know, we've got a candy store and the kids have now gone
in and just gone, all right, what can I, you know, or a Lego store,
more like a Lego store, what can I build?
Now I've got all this Lego available to me and we're just seeing these pots
of innovation come out of everywhere, which is really interesting.
So I think Project Sophia as well, you know, that's... Oof. Yeah.
(11:51):
We could spend a whole podcast talking about that. That's very exciting.
Maybe we should. Maybe that's one of the next topics we'll talk about,
Scott. I think that's very interesting and that's going to basically unpack
and hopefully bridge a lot of this stuff together.
Yeah, I think I've got the perfect guest to come on, Greg Nash,
who I talk with regularly, specializing in data and AI.
Let's bring Greg on to really talk about Project Sophia and get his thoughts
(12:13):
because he's deep in the weeds on this stuff all the time.
But to put it in perspective, I reckon I saw this week on our internal calls
probably about six or seven different projects like Project Sophia that are
going to really change the way that we work.
So everything's got a project name, as you know.
But yeah, that was really exciting stuff. So all these projects are great.
(12:37):
But that one in particular, I think everyone's excited about to see.
And for those that haven't seen it, you can go on and we'll put links again in the chat.
But effectively, instead of constructing a,
I guess, using Power BI and constructing a data visualization now,
we just use natural language to go and prompt data and it creates the visual
(12:57):
that suits the data, if that makes sense.
So hey, grab me this and I'll just go and do it. and then you can click on it
and go, can you change this from a pie graph to a bar graph and so on and so forth.
So it's really a conversation with your data. I really like it. It's good.
Yeah, there's just so much cool things happening at the moment and there's so
many new sort of technology options and streams and all of this kind of stuff.
(13:18):
And I agree with you when you mentioned before about going fast.
I remember I caught up with Aaron Bjork, I think a couple of weeks ago in Sydney
and he was talking about the Microsoft strategy about going fast,
going fast with this and making sure they can roll out as much,
as quickly as possible and embed this into their products. And again,
from a partner standpoint, I've never.
Seen anything get adopted and pushed out so fast. And it's really amazing.
(13:39):
And it's really incredible. And I guess a lot of the learnings we're seeing
as part of that sort of initial push is really beneficial to Microsoft as part of this.
And the amount of use cases we've seen in the sort of last couple of weeks is really incredible.
Again, we caught up with Dwayne Robinson, the co-pilot, I think the product owner of that.
And again, that was a phenomenal session. I really enjoyed that.
And again, we're seeing a lot of really fantastic co-pilot studio examples come
(14:03):
come out, particularly the last couple of weeks as customers get more familiar
with it. I think that's really exciting.
And again, that's a really good segue into, I guess, maybe our guest for today.
And Scott, maybe you want to talk about Adam in a bit more detail and kind of
what he does within Microsoft.
Yeah. So I've been working with Adam for a long time now.
And Adam took on the project of spearheading, I guess, enterprise program for
(14:25):
early adopters for CodePilot and program.
And he works with a lot of customers to be able to drive that adoption.
So he was Just kind of in it from the beginning, Adam's done a lot of different
podcasts and YouTube sort of videos. He's shifted on with Lisa Crosby recently.
It was good to have him on the show and really have a bit of a chat to him about
where he sees the practical use cases for Copilot.
(14:46):
And look, really timely for us as well because Copilot's getting GPT 4.5 in April.
So that means it'll go multimodal, which opens a whole bunch of more use cases
right there for the product.
And also our co-pilot for sales in the biz app side of things and co-pilot service
is starting to get really mature now.
(15:08):
So effectively, we have, I mean, to demystify it a little bit,
we have three co-pilots in my world that really make sense.
We have co-pilot studio, which is, you know, how to create a chatbot and really
build upon. And that is super popular, like we spoke about at the AI tour.
We have co-pilot for service, which works across Dynamics products as well as
Salesforce, ServiceNow, and others.
(15:29):
And then we have Co-Pilot for Sales, which does the same, works across similar products.
So, and, you know, they're really there to optimise the people that use those
products, the salesperson or a customer service agent or whatnot.
Yeah, I agree. And Adam, similar to your sentiment, Adam's been doing almost a road show.
So if you see a little co-pilot booth and you're located at one of the AI events,
(15:50):
normally Adam will be kind of running the show and he knows his stuff,
obviously. He's a really good guy.
And also, I actually got a lot of benefit out of the conversation because I
found all these great examples that I hadn't heard about, about utilizing M365
co-pilot. So thanks again, Adam, for saving me about maybe four to five hours
a week and on the road to 20.
That's all I want to say. I really want to leverage AI and I'm going to be picking
(16:13):
Adam's brain consistently to give me all these new shiny examples of how we can use Copilot.
So yeah, we really hope the listeners enjoy that chat and get a lot out of it.
I actually downloaded a mobile application, which Adam told me to on the podcast
during the podcast again.
So you'll hear that come up and I recommend everyone also downloading that,
but I won't give that away.
Way we want the listeners to listen to it and then obviously get the benefit
(16:34):
of that so yeah awesome all right well we might take a short break matt and
then we'll bring on adam onto the show sounds good.
Music.
(16:54):
Thanks, Matt, and good to be here.
Thanks, Scotty, for inviting me. Yeah, look, my background is customer,
project, program, portfolio, delivery, then moving into Microsoft,
really focused on the adoption and change as a CSM, customer success manager from our work.
(17:14):
Because what I've realized through delivery was it doesn't matter how well you
do a project from a project triangle perspective, you know, scope,
scale, and budget, it really is about our people getting the value from it.
So being able to drive the value from your investment is where I then focused on that.
Matt and I had a great experience working together on Parkland,
(17:35):
really seeing it not from the Microsoft helicopter or 30,000 feet there,
but boots on the ground in terms of just the challenges and the opportunities
that people have when adopting technology.
And then, yeah, then came back to Microsoft and had.
I had a great experience over the last nine months taking all of our early access
program or the private preview customers in enterprise commercial through the M365 co-pilot,
(18:01):
which if you had asked me back in October about this, I probably would have
said to you a very different answer,
but very grateful now to be able to reflect back on just what it was like to ride the crest of a wave.
That is co-pilot and I look forward to sharing more about that today.
So maybe let's click a little bit more into that early access program because
(18:21):
effectively you were the first sort of group of people to really take Copilot to market.
How was that? How did you find that? I mean, no one knew what it was.
And here's this program that we have to talk about with customers.
Yeah, look, without revealing too much, it was the closest to working in partner
land that I've ever had at Microsoft because I moved out of being in the field into Corp.
(18:42):
And we were part of the CSAs globally that were looking after any single early
access program participants across the globe.
Initially, we were looking at around 600. I think it ended up being around 1300.
And it was, how do we take them through the journey of phone purchase through
(19:04):
to establishing the right governance framework, framework,
making sure that the adoption was strong so that you can drive towards business
value and then how do you capture that business value for something so new to
be able to feed into your business days.
Very much like in partner land, you might have put in a strategy for a customer
and then they say, no, can you execute to value?
(19:25):
But there was a journey we all needed to go on. So we fast forward kind of a
year on, I guess, from that point,
customers had a chance to user what's the
feedback on office copilot there the
answer is like you would expect and it's
kind of a parallel to your adoption to curve for technology
(19:46):
there are still customers that are part of the emp that have really started
are still stuck at their security governance compliance risk management stage
and then there are those that are uh that have customers i got an email today
with one saying don't try and take my co-pilot license from me to their project team
adoption team so the answer really comes down to how you've ran it and some of these are not,
(20:13):
you know ai or generative ai related it's basic program and change management around,
drawing sponsorship giving people the support capacity to play to learn to grow to experiment,
because ultimately the biggest change with co-pilot is moving from let's call
it that kind of waterfall deterministic thinking of try to pass the LES no,
(20:37):
co-founder doesn't work to.
More of an agile kind of way of thinking, experimentation, feedback,
come back mindset, try it later.
And those organizations that did that well, typically partnering with an innovation
function that was driving and supporting them, really embraced that approach
and supported their users to get to capturing alpha.
(21:01):
Really something I'm really interested in having a chat because in the introduction
before you kind of alluded to the way that you really look at making sure the
organization get value from the product.
And it's something that I'm working with you for a period of time definitely
resonated with. I think both of us, and that's why we've always had a really good synergy.
But talk to me about how organizations specifically are getting value from the
various co-pilot suites of products because there are a number of them.
(21:23):
There's hundreds of co-pilots.
So maybe talk to me a bit about your definition of co-pilot and maybe some examples
of the co-pilot business value returns that you're seeing from your customers
that you've been working with.
Yeah. Microsoft marketing don't always make it easy for customers.
If I aggregate it up to a higher level, the co-pilot brand is about having an
(21:44):
assistant in whatever you're working in, a digital assistant with you, whatever you're doing.
So for example, co-pilot for the web, aka Bing Chat, Bing Chat Enterprise, etc.
Is a digital assistant across finding and generating information or whatever
you need to do that's based on the course of information that's available out in the world one way.
(22:06):
So it really comes down to helping you where you are, whatever you're doing.
And that might be, you'll use different co-pilots depending on your role.
You might be, for example, in service, customer service,
and I've got some great examples of where they have a lot of value out of M365
co-pilot, but obviously think about Dynamics or Sales co-pilot,
(22:27):
and also those extensibility players into your world around Power Platform and Co-pilot Studio.
Because we don't just do one thing, like we don't just have effective meetings.
We also then typically have to communicate out about that. We also have to convert
that into maybe different languages, different audiences, different segments
(22:48):
of who we're targeting with this message, and we'll also need to capture and measure all that.
Not everything will be done out of the box with CodeFiler Frame 365,
but what you'll find is that there isn't a problem that we've found so far that
can't be solved with the full stack, whether it's Copilot extensibility into
low code or into to Procode and the whole Copilot stack.
(23:09):
It's really interesting, Adam, because I know my team's been doing a lot of
work at the moment around MTC,
so Microsoft Technology Center sessions, where we can connect Copilot with Azure
to grab customers' data and then turn that into actionable outcomes and having
a lot of success around that.
So I think we can see the platform now is really coming mean together,
(23:35):
and particularly around guardrails.
So, what's your feedback around, you know, governance and guardrails and things
like that that you're seeing in the platform and being able to support CodePol?
I think it's a unique proposition from us. Yeah, absolutely.
I think this has been like the eternal kind of fight between good and evil, if you will,
between security versus usability and the slider of where do we actually accept
(24:01):
risk risk and innovate, or lock down and manage risk to zero.
And so for me, it's about having the right sponsorship and awareness of really
business risk management rather than trying to manage risk to zero.
So obviously what we're able to do in your world to set up those guardrails
(24:22):
so people have that experience of being able to experiment within guardrails
is critical to innovation.
How do we have that right balance and having
the right setup to support devs in
a low-code way can only accelerate development but
i think we miss a trick when we don't then take that to really the end users
(24:45):
and to me that's when one hit if you will the holy grail of how gen i can really
be enabled because think about the likes of well one of our own now nathan backers and what Well,
he was able to do a Telstra by having the SME knowledge to be able to implement that.
That's the opportunity across all of the Copilot world and universe and what
(25:08):
NAI can do with people that are in the Microsoft stack.
But we have to give them the opportunity to do it.
So Adam, you've been running a lot of, I guess, you've been on the conference tour, I've seen.
And for anyone that has been looking at the Microsoft socials,
Adam's basically a celebrity now.
There's lots of videos going on where Adam's talking to customers.
What I'm interested to hear is what type of questions you've been getting from
(25:32):
customers during your basically your conference road tour that you've been going on.
And I guess how have you been directing and giving advice and all those kind of things? user.
And will AI take my job, for example?
And through the Early Access program where we are working with those 300 users
that had a license to be able to do this, my answer was, okay,
(25:53):
I may not take your job, but people that can use AI will.
So this is your opportunity to really get and build the muscle and the skills
to be able to interact in a new way, in an agile experimentation model with
learning a new language?
How do you talk to a generator of AI, Copilot, whatever Copilot you're using?
(26:15):
And also, how do you go on that change manager journey?
Because personally, I was frustrated as hell sometimes with Copilot,
but candidly, Copilot didn't care.
It was only me that got frustrated. So how did I then have to manage myself,
to grow, to learn how to use Copilot? You've got to be nice.
What was that about? You've got to be nice to AI, so yeah.
(26:39):
Absolutely. And the answer to your question comes down to who I was talking
to, what they did, and what they were doing, and where they were taking a value-based lens.
So the best engagement I had with COOs typically that were taking a full lens
across the business, rather than just a cost view, or this view,
or that view, or a personal risk view.
(27:02):
I think customers are still kind of kind of trying to
work out you know where this all fits and and
how how that can be more productive where's the
return on investment all of these things yeah copilot's a
licensed option so maybe just talk us through
how people get started with office co-pilots and
and you know what what you're seeing how people
(27:23):
are purchasing this and what are the kind of all the quiet yeah absolutely
and look there's been a lot of learning certainly at the enterprise level
that and i also have a lot of friends of
mine that are in smc space or smb space as
a microsoft would look at small to medium businesses or across
that whole kind of lens and given the
majority of australian businesses are more in the
(27:45):
are not enterprise it is amazing how much value there is that space for co-pilot
and especially now we've opened it up to anyone one effectively with basic business
premium license because one of my friends we tested this with during EAP was RFP responses.
(28:05):
It consumed so much time, Matt.
You and I have had the joy of that on both sides of the fence, right?
It makes me cringe. I saw you shiver.
And so how do you actually run an RFP process to generate that on the outside?
How do you then on the other side of the fence take your PowerPoint presentation
that you need to now present based on the submission that you had.
(28:30):
That's kind of the basics, like just drafting content creation and co-file. I'm working with it.
It's how do you then kind of build on top of that? How do you take users towards what else can it do?
And it's that proven model of the longer the adoption, the stronger the business
value is generated because you have more time to think about what can it do for me?
(28:50):
Apply it to what I do. And so, like, if I take that RFP example,
can you now compare this response with this response? What's it missing?
Or these are the key criteria, co-pilot, this is my response,
have I addressed all of the key criteria?
In the past, you'd typically have, Matt would produce something,
(29:13):
he'd send it to IC to review before you then would send it out more broadly,
co-pilot is now your first reviewer.
So we're missing a trick around productivity, productivity, productivity,
but what about quality? How do we measure that?
Because Copilot can review your document.
Does this conclusion match their body? Does it have a call to action?
(29:36):
Copilot, can you help me to generate an inspirational speech based on my annual
report and reference that file?
Include some famous quotes tied to this content.
Oh, and by the way, there's a joke about AI. Could you also include a funny quid.
It's that kind of using it in how you do your job.
And the last thing that I typically do when I'm talking to an exec is also,
(29:57):
so how did you go in your annual?
So when you presented this, how did you go?
Did you watch the video? Nah. What did your team say? Oh, they said it was great.
Of course they did. Did you ask about it?
Copilot won't actually care about your role. It'll give you the true reflection.
So when I ran one of those sessions that you were talking about,
(30:18):
I asked Copilot afterwards, how did I present? Did the message land?
Did it sound like I was reading the slides? You know, you can get that feedback
because it's independent without any fear of it losing its job and managing up.
I love the RFP example, and I love deep thinking through these things, right?
(30:42):
So if you think about an RFP process, we've got Copilot creating the RFP.
We've got Copilot responding with the RFP.
We've got Teams calls throughout the RFP process that are summarized by Copilot,
so people don't actually need to attend the meetings.
So where does that leave, when you think about it, where does it leave pre-sales
for us and our engagement with our customers? Matt, you probably got some comments
(31:05):
around that as well, given that you guys are from the services point of view,
right? That's another story altogether.
Yeah. I just want to thank Adam for saving me 20 hours a week, based off that .
I think every other partner that's listening is of a similar sentiment to myself,
because that's a fantastic use case. It's a common pattern.
We should be using AI and co-pilot to do those repeatable tasks.
(31:29):
Again, thanks Adam for sharing that. And I want to know if there are any other
sort of specific examples that you are utilizing Copilot for considering your experience.
Are there sort of some quick tips, quick tricks?
We do Teams meeting notification summaries, all that kind of stuff.
But I wonder from the master, what other things is Copilot doing that I should
(31:49):
be doing and the listeners should be doing at home?
Unfortunately, Matt, Copilot can't help you with what you shouldn't be doing.
So we just focus on what you still need to do.
How's that for you? You can ask, but I don't know.
Yeah, look, I think it's really interesting because I try and tell the stories
(32:09):
with users so it resonates with them in a storytelling fashion.
So for me, we have, and I think every Every customer of ours at this stage has seen Project Falcon.
And it's a demo that shows how you can capture the sentiment of the meeting,
understand the decision that was made, and save yourself an hour.
(32:30):
And in three minutes, you can take that information and understand the pros
and cons for the decision to delay the go-to-go by two weeks.
But then how do you take that and turn that into what you do?
So I was a program manager. So how do you then use GoPilot to take that information
and put it into your steer code deck?
And can you turn it into an audience for an exec?
(32:53):
So be more sustained and rewrite the language for me without having to manually
do that, without having to go back through the tape multiple times and capture that.
Then I now need to update the project schedule tied to that,
right? Because the milestones, and it's now two weeks delayed.
I can't even remember where that file was. Was it Matt sent me it?
Is it in the project governance folder on Sharecoin? No, it isn't.
(33:17):
And I've got to get this done now.
We spend 27 minutes a day or three weeks a year wasting time on things we know
we have, we just can't find.
Copile. What is the project milestones for project file then? It'll find them for you.
Then, one of the other things that I got burnt on this before is it execs about 400 emails a day.
(33:38):
I was told by one of the EGMs and one of the banks.
How much time they spend in meetings and how much of that they actually read
is very hard right and when i was in
portfolio delivery one of the milestones had been
changed on the project that i was providing an update on and i didn't know but
some of the other people in the room had just found out before the meeting so
i used topo to actually find look in my emails and teams chats over the last
(34:02):
reporting period say two weeks find anything related to project Falcon and it will find,
because I've sent it or someone sent it for me,
an email that will show there was a milestone file was delayed by five days.
So I'm finding things I didn't even know I had. Then don't stop there.
Okay, so Copilot, can you now update my milestone list with the email I just
(34:28):
received and output that in a table so I can copy that into my Stereo Codec
and it'll do that for you.
So it's how do you actually kind of just have the time to play experiment and
see what it can do for you in what you do.
That was kind of one of the let's say the demo scenarios.
You kind of just need to give yourself the time to play to experiment to learn
(34:50):
and grow and I always assume I'm the problem if I'm not getting the output from
Copilot until I kind of get what it is and get where I want to get to.
So there's some of the small examples. The last one I'll give you is I had to
play with Copilot Studio because with all the hype and Copilot focus,
I haven't had any time really to focus on teams and what we used to do in our old world.
(35:14):
And I had to renew my associate certificate.
So I built myself a Copilot Studio.
Pointed it at the learn.microsoft.com URL and did my test where I had the best
imaginable search engine to find everything I needed to answer some of the questions
(35:36):
that might have come up in that test.
It's very interesting that you've opened that can of worms now because I think
we're either going to generate a whole bunch of people that appear really,
really smart and certainly going through interviews and things like that,
people should be really prepared.
If I was interviewing someone for a role in my team, my expectations have now
(35:58):
gone up. You should know who the team are, what the technology is.
You should even know a little bit about Microsoft, the company,
around what its goals are, because it's just a chat, right?
So I think everyone's kind of leveling up now.
You know, this is the new expectation. If you're not on board,
then you're going to be, you're going to struggle, right?
And I think the same goes for tests, the exams all
(36:19):
these sort of things where they're open but when the natural
thing is to use language model for this stuff but suddenly you
know i can see a podcast matt top 10 hacks for copilot or top 10 use cases you
could probably do that every three months and that's gonna build this list of
you know these great ideas because they just keep flooding in all every day
out here to something crazy and we build out like all the csa caption those
(36:42):
scenarios and they're now going to get put
into the scenario library that you get as part of CodePilot.
I know. Because, you know.
Really what you're trying to do is there's the innovators and the people that will,
experiment and there's the you know early adopters and then there's the followers
that we want but they probably won't spend the time to do it themselves they
need to be shown or taught and that's really the key is how do we build up this
(37:04):
library for people to be able to capture.
The value from it so that people can then just apply it and see for themselves and what they do,
and each organization is capturing that because you'll be able to put that into
co-pilot lab and filter on your own organization all those prompts and use cases
that are working really well for what you do.
(37:25):
You know, we spoke about this with Nathan where we were talking about,
you know, when you first join Microsoft, someone sends you an Excel spreadsheet with 200 links.
We're going to kind of figure out how you navigate Microsoft.
And we think the future is probably going to be here's 200 prompts.
Go punch that in the copilot and you'll get a lot more sort of information.
(37:47):
So, you know, maybe there's a Power App there, Matt, a prompt Power App where
you just click on a button, I want to do this, and it just sends a prompt to a language model.
So you've saved me 20 hours and a new Power App idea. Adam, you're welcome to
show it anytime you want, mate.
So that's really good advice as part of it.
I agree with the 10 tips and tricks. As Adam was talking, I was just sitting
(38:09):
there nodding my head because I was like, that's amazing.
Yeah. So, Adam, you're obviously quite mature and using CoPilot for someone
that's looking to kind of level up and start to navigate it.
You kind of mentioned before, just get started.
But what are some other things you can do to think about how to interact with
Copilot? Because I think that's the thing, get started.
But yeah, just some advice from you about navigating it through and having it
(38:29):
as that, you know, basically your assistant to help guide you. So, yeah.
So probably two things there. For M365 Copilot, for any of the adoption people
out there that are trying to drive this and someone's just giving you a few
licenses and you're like, how do I, what do I do with it?
It's really important that people understand that CoViolet, as we talked about,
(38:51):
is everywhere, and it's different in how it will behave, because deliberately
tied to the scope of what you do.
So if I'm running a meeting like this and we have transcription on, I can interact with it.
Matt's asked a question, I didn't quite catch his name. I don't want to be rude.
What was Matt's name? Or Matt's mentioned some stochastic interest rate that
(39:13):
I remember from my uni days, but no idea what that means.
And I don't want to look like the dumb IT guy, quickly ask Cobol.
So remember, there's no dumb questions around what Cobol have been doing for
you and it never gets bored.
It'll never get cranky with you or short tempered because you've asked the same
thing three times. So it is really there for you no matter what you're doing.
That might be sending an email and I myself have ADHD and sometimes I can be
(39:38):
very direct when under pressure and conscious about how the tone might come across an email,
which is not actually how I feel, it's just how that can be written.
So using coaching by a co-pilot to put itself into the shoes of the recipient,
give me feedback on, does this have empathy?
Does this have clarity of the ask?
Those are really important kind of upskilling for me.
(40:03):
Because it also then helps me learn how to write it better the next time and
hopefully less kind of coaching required as we all upskill.
And so you'll see that as a pattern across the tools. It's there to upskill.
Excel will tell you and help you with multiple things, but it'll also explain
the steps so that you can actually learn as you go by osmosis, right?
(40:24):
So I use the analogy of one room at a time or introducing a kitten to a house,
you know, one room at a time.
Do the same at Copilot. that, take people through the adoption around effective
meetings or some sort of hero scenario like finding information,
whatever it is, because if we don't do that, we see a big drop off in that hype
cycle of Gartner, if you're familiar with that.
(40:46):
Very soon people go from, what's this all about, straight into the trough of
disillusionment, wonder what this, like just giving up.
To help with that is how do we actually show them value so they'll stay with
the journey, because it is a journey, it's a change journey.
When Judson was out a few weeks back, a couple of weeks ago,
he said to everyone in the room when he was in New Zealand, I'm thinking you
(41:09):
should use the CodePilot app.
And I was one of those people that hadn't used the CodePilot app before.
I was surprised about it.
Download it now. I've used it every day since. So I think anyone getting started
out there, if they really just want something, then you can just, it's just so useful.
The amount of times I just punch into that CodePilot app before i google or
(41:31):
bing something now it's it's becoming that first place to find things out and
you just get the answer no ad just get an answer yep and we're absolutely loving
that app at the moment matt's just sitting there downloading it actually yeah.
Well hopefully there's like 5 000 people listening to this podcast now like
(41:54):
grabbing that app and downloading.
Maybe time to wrap it up, Matt. I mean, it's been amazing, Adam,
having you on the podcast.
I love your feedback as always and a lot of experience around this,
having run that program and being the first to market.
But thank you for joining the show and chatting about all things CodePilot and
use cases. Matt, any final words?
Yeah, thanks again, Adam. And can CodePilot help my golf game?
(42:17):
Because I think we do a game of golf soon. So let's get onto that.
It can at least help you to set a reminder for the alarm to go off so that you
are there. Yeah, start there.
Anything to tell me next time.
Anyway, we're digging a whole lot of little dark secrets like the sweet prince
and everything from last show. So yeah.
(42:38):
Sweet prince is out there. I think that's how we're starting the next podcast.
But massive thanks, Adam. And thanks for sharing a lot of that.
I got a lot out of it and I can imagine everyone that's listening is part of it.
And I think definitely getting you you off on a list of potential to chat more
about tips and tricks because I can imagine in 12 months, the scenarios we're
(42:59):
going to understand and be utilizing with Cokepilot is going to be phenomenal.
So it's a continually learning, evolving journey for everyone.
So looking forward to continuing the conversation with yourself.
And thanks again for coming on the show.
Cheers, guys. Thank you so much for having me. Thanks, Adam.
Music.
(43:33):
Hey, that's a wrap for today. Thanks so much for everyone for listening to our podcast.
Just a few reminders. So we've set up a LinkedIn page. So we'll endeavor to
put some updates on that LinkedIn page. So go on there and click follow for us.
Also, don't forget to go to powertime.au and subscribe to our newsletter.
And lastly, if you could please click follow on this podcast so you get the
(43:55):
latest episodes as they come out.
We're trying to get the podcasts out every two to three weeks.
So there should be a good cadence with a whole bunch of new guests coming on.
Now, lastly, don't forget that the views on this podcast are not the views of
NCS or Microsoft, but they're the views of Matt and I and our guests.
Thanks again. See you. Can't wait to see you on the next podcast.
Popular Podcasts
1. Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
2. Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations.
3. Crime Junkie
If you can never get enough true crime... Congratulations, you’ve found your people.