April 8, 2025 • 24 mins
Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM
FULL SHOW NOTES
https://www.microsoftinnovationpodcast.com/673
Microsoft appears to be settling into a more deliberate AI strategy with increasing focus on data governance, security, and Fabric as a central component for AI implementation. The recent clarity around Microsoft's vision suggests a transition from reactionary AI deployment to a more considered approach for long-term value.
TAKEAWAYS
• Microsoft Fabric and OneLake emerging as an "AI landing zone" for consolidating multi-cloud data
• Free mirroring technology for AWS and other platforms reduces barriers to cross-cloud data integration
• Business Applications division potentially facing reorganization within the next five years
• Dataverse likely moving to Fabric while Copilot Studio shifts to Microsoft 365
• Traditional UI paradigms may disappear as AI interfaces replace menus and navigation elements
• Legal involvement becoming critical in AI development for the first time in software history
• Increased focus needed on the social and governance aspects of ESG in AI implementation
• Serious concerns about AI's potential to facilitate cybercrime and human rights violations
Microsoft 365 Copilot Adoption is a Microsoft Press book for leaders and consultants. It shows how to identify high-value use cases, set guardrails, enable champions, and measure impact, so Copilot sticks. Practical frameworks, checklists, and metrics you can use this month. Get the book: https://bit.ly/CopilotAdoption
If you want to get in touch with me, you can message me here on Linkedin.
Thanks for listening 🚀 - Mark Smith
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Mark Smith (00:01):
Welcome to the Ecosystem Show.
We're thrilled to have you withus here.
We challenge traditionalmindsets and explore innovative
approaches to maximizing thevalue of your software estate.
We don't expect you to agreewith everything.
Challenge us, share yourthoughts and let's grow together
.
Now let's dive in.
It's showtime.
Welcome back, welcome back.
(00:22):
Welcome back.
We're on air again, mpp Summitis behind us, the world has
changed dramatically for a lotof us and we're on the Ecosystem
Show.
Andrew, how are you?
Andrew Welch (00:38):
The world has always changed.
For you, mark, it's always likeevery couple of weeks that we
go without recording, recordingan episode and then, yeah, the
world has changed I'm in massiveabsorption mode at the moment
what was your highlight of?
This is a leading question, but, uh, what was your highlight
from mvp summit is, to theextent that you can share yeah,
hanging out with you, of course.
(00:59):
I mean the extent that you canshare.
Mark Smith (01:01):
Yeah.
Andrew Welch (01:04):
Hanging out with you.
Of course, I mean, that's noteven that.
So what was?
What's your second choice?
Mark Smith (01:10):
Hanging out with you , of course.
Andrew Welch (01:13):
Oh, great Okay.
Mark Smith (01:14):
Well, I think we can ?
Andrew Welch (01:15):
I think we can move on Okay.
Mark Smith (01:27):
So I suppose the highlight for me was the clarity
that I feel Microsoft issettling into an AI rhythm that
is not reactionary.
Andrew Welch (01:31):
I feel this.
Mark Smith (01:32):
You know they've gone through two years of
jamming copilot into everythingand getting their branding and
go to market kind of.
It's kind of definitely been aassemble the plane while it's
flying feel I remember, do youremember, mvp Summit two years
(01:52):
ago?
Andrew Welch (01:52):
And that was very much like that was the feel
.
Though MVP Summit two years ago, I will say best MVP Summit
I've been to, because that waswhen we learned about Fabric,
but anyway, yeah, sorry, Iinterrupted you.
Mark Smith (02:06):
Yeah, the stabilization and and I feel
that you know I don't know ifit's confirmation bias, but you
and I have been talking, all ofus have been talking on this
podcast for some time now, maybeeven a year the extreme
importance of purview in the aistory, the extreme importance of
fabric in the ai story and,yeah, which is all about data
(02:32):
governance, security I meandon't get me started on intra,
id etc.
Because that all plays a partin the story as well.
Co-pilot for security got a lotof push at, I feel it, at
summit.
I just think there's amicrosoft's vision in the space.
I just feel it's on par rightwith particularly the purview
(02:52):
and the the um fabric story,which enables then everything
else yeah, well, so so it youknow.
Andrew Welch (03:03):
Interesting that you that you mentioned that
because it is this week isFabCon, or the Fabric
Microsoft's worldwide Fabricconference, and and I couldn't
be there for family reasons, butI'm quite sad to have missed it
I would love to try to make itnext year.
One of the things that I thinkis emerging and this is a story
(03:27):
that Microsoft has been kind oftrying to tell for a couple of
years since Fabric was released.
But let me go back to lastsummer.
I was working with a client,working with an organization
that is they're a big enterprisefinancial services organization
.
They are the extreme ofmulti-cloud.
(03:50):
I mean they've got everythingin there.
They have a data platformthey've built with AWS, they use
Microsoft 365.
They have Copilot for M365.
They dabble in power platform,they have ServiceNow, they use
SAP.
I mean these guys have they'rebuilding AI workloads in the
(04:13):
Google Cloud.
So this thing is thisorganization is.
You know that's a decision youcan make, right.
But the problem that they werereally wrestling with is that,
wow, we've gone, we've backedinto this multi-cloud world.
Yeah, how do we bring all ofthis together?
How do we bring all thistogether for use with AI?
(04:34):
So the one thing that I knew Icouldn't do, the recommendation
I couldn't make, and for goodreason, right, I couldn't say
well, let's rebuild your dataplatform, right, they've just
spent the last several yearsspending millions of dollars to
build, to platform their dataplatform in AWS.
So that was sort of aconstraint that I had
architecturally speaking.
(04:54):
So I pitched to them this ideaof using OneLake, which is a
part of Fabric, and I called itthe AI landing zone, which is a
part of Fabric, and I called itthe AI landing zone, nice.
And my thought was that whatwe're going to do with OneLake,
primarily, wherever we could,through shortcuts, and OneLake,
obviously or not obviously, butif you don't know, shortcuts
(05:16):
data not just from Microsofttechnologies like Dataverse or
Azure SQL, but it shortcuts datafrom Amazon S3 or from Google
Cloud, know from from Googlecloud.
So my concept was let's use anand Oracle, right, so?
So let's use off the top of myhead, do you know, I don't know
if there's a shortcut fromOracle off the top of my head.
Mark Smith (05:38):
Well, the mirroring technology came out of GA just
recently.
There's that option and I don'tknow if you've seen that the
mirroring option for a lot ofthese workloads AWS et cetera as
well is that it's free.
Yeah, have you seen that, whichis kind of like it breaks down
that barrier of no.
I've already done this in aws,already done this somewhere else
(05:58):
, so I could take advantage ofthis free method of still
bringing that data in for yourai landing zone.
Andrew Welch (06:06):
I like it yep, yep and and I I just think that
this idea of using one lake asthe sort of the final point of
data consolidation and thetechnologies that Microsoft has
developed from a cutting andfrom a mirroring perspective,
treating OneLake as that finalpoint, right, and thinking of it
(06:27):
as the AI landing zone for yourdata, or maybe the data landing
zone for AI, however you wantto couch it, or maybe the data
landing zone for AI, however youwant to couch it, it's becoming
more and more clear that thatis Microsoft's vision and that
that is a complete vision, or anincreasingly complete vision.
The world and the technology ismoving so quickly today that
(06:49):
calling anything complete is alittle tough.
But yeah, so I don't know howwe got on this riff, but it is
Fabcom this week.
It'll be really interesting tosee what comes out of it.
But I just think that the otherthing that I think makes Fabric
and OneLake specifically sounique is that there is no
competitor to it.
Yeah, there's lakes and there'sdata platforms, but this is one
(07:11):
where, for all of its faultsand to me, the only faults I
have with Fabric are that it isstill a work in progress, right,
it's like Power Platform five,six years ago.
Right, but for all of itsfaults.
It has no competitor, it has nopeer, and I think that that's
as someone who has been justobsessed with data platform tech
, like when AI really hit thescene two and a half years ago
(07:35):
and we we had chat, gpt, thatthat entered the market, which I
cannot believe.
It's been two and a half yearsnow.
I became a data platform geekfar more than I became an ai
geek.
Mark Smith (07:45):
So, um, interesting, yeah, I'm very excited about it
do you know, and, by the way,I'm going to couch what I'm
going to say is this was not anannouncement, nobody even
alluded to it, but one of mystrengths is be as is has been
able to connect futuristic dotsin the space and draw
conclusions about what I see.
(08:06):
Yeah, my observation was thisis that I think business
applications is going away atmicrosoft in the next five years
as we know it now.
I think that the dataversecomponent naturally will go
across to fabric right and andand, and that will become part
of their suite.
I then see copilot studio,which I feel is the last thing
(08:29):
standing as look what we haveinside microsoft now, and Apps
is going to go to M365.
And therefore, you know, itseems to be that's the only you
know.
The big thing that is oneveryone's radar is the value
from Biz Apps, is the CopilotStudio story, which is that, and
(08:51):
the AI mix.
So I'm just wondering.
I feel in my other observationsacross Microsoft that M365, a
lot of modern work folks alreadythought that Copilot Studio was
theirs.
They've already for some timethought that Power Automate was
right Because of the SharePointyou know history and stuff like
that.
They thought Power Apps wasbecause it lit up in SharePoint
(09:14):
with that little purple buttonthat said let's turn this into
an app.
And so I just see that there's.
You know, as microsoft goesthrough a reinventing itself as
an ai company, which it clearlyis that you know you've seen
mustafa come into the businessand it is the ceo of um ai
inside microsoft.
The whole re-engineering partof the whole engineering org
(09:37):
massively got reassigned to himin that move coming into
Microsoft.
And I'm just wondering ifthere's this, as they
rationalize their business,going forward and AI really
blends everything that I thinktheir organizational structural
models might start to change.
And hence for me you know, Isee my career, you know,
(10:00):
although it's been a powerplatform for 21, 22 years now,
it's really starting to leanmuch more to the data and AI
kind of story and the work we'redoing.
Andrew Welch (10:11):
That's a really fascinating excuse me,
fascinating question, and let'sdefinitely so.
I have some thoughts that I'llshare, but let's definitely
explore that on a future episodewhen we have more of the posse
here, because I'd love to knowwhat Anna, some combination of
Anna, chris and Will think aboutthat.
(10:32):
But know, listen, I haveconversations with I have
conversations with you know acombination.
Sometimes it's with a you knowkind of a C-level decision maker
in a Microsoft customer andthey're trying to, they're
considering which technology toinvest in.
Sometimes it's with a Microsoftpartner that is considering
(10:55):
which lines of business toinvest them in, to pursue.
Sometimes it's actually withinvestors who are looking at
buying a Microsoft partner.
Potentially right, andinvariably we talk about biz
apps.
Business applications has beenan odd fit for Microsoft for a
(11:18):
long time.
Right, like, remember,microsoft is a platform company,
so we sort of saw this comingand going right.
Like sometimes, when Microsoftwas doing really well and really
wanted to expand beyond itsborders, you'd get a lot of
interest in Dynamics.
Many of those Dynamicsapplications were actually
acquired by Microsoft.
They were not originallydeveloped in-house because
(11:40):
Microsoft felt they needed acompetitor to SAP or to Oracle
or to Salesforce or to whatever.
So BizApps have always been areally interesting fit within
Microsoft.
First of all, I think that noneof the technology within what
we think of as businessapplications, dynamics and Power
Platform.
I don't think that technologyis going away.
I do think that Dynamics isgoing to be seriously, seriously
(12:06):
rethought and not re-platformedis not the right way to put it
but the whole concept of SaaSbusiness applications are going
to be very, very different infive years.
We've talked about that on otherepisodes.
You do raise an interestingpoint, though, about our
platform.
So I want to be clear in caseanyone from Microsoft is
(12:28):
listening to this.
I'm not advocating adismembering of that group or of
that team or of that technology.
But if you do think about it,power BI already has a foot in
both worlds right, it's brandedas Power Platform, but for all
intents and purposes it's ownedin Microsoft as part of Fabric
right, as part of the dataplatform team, dataverse.
(12:48):
I think that it's becomingincreasingly difficult to
implement Fabric or Dataverseindependent of one another right
, so it makes a ton of sense forDataverse to really be the
application, the transactionaldata orchestrator of the future
for Microsoft.
And then you know, for a longtime I've said that the other
(13:08):
Power Platform components, powerApps specifically, really are
just a collection of Azuremicroservices with a marketing
label wrapped around it.
Yes, yes.
So there's a very I mean this isa really interesting
(13:30):
hypothetical going away as as aiand ai's ability to do a lot of
the things that citizendevelopers previously turned to
low code for.
Yeah, but certainly doingmaking some moves there would
really help microsoft in thisfoot into world's issue that
(13:50):
it's had for a long time.
Is power platform for seriousapplication development, which
we have long said that it istime?
Is power platform for seriousapplication development, which
we have long said that it is, oris power platform a citizen
developer tool?
Right now, the answer is both,but microsoft has long had a
marketing problem there yeah, I.
Mark Smith (14:05):
I think that like well, one year away now from
power platform being 10 yearsold, right 2016, when it came
out, I know the nomenclaturecame out of power platform in
2019, but the tooling was it wascalled.
I know the nomenclature cameout of Power Platform in 2019,
but the tooling was there.
Andrew Welch (14:17):
It was called the BAP Mark, the BAP, the
Business Applications Platform,bap.
Yeah.
Mark Smith (14:22):
And I just think that, you know, will Microsoft
365 become a platform, maybe Inthe time ahead?
It's just a positioning.
I don't think any of the tech'sgoing to go away.
I think it's like it's becomingvery mature now.
It's a proven thing, you know.
It has massive scale, massiveum companies that use it at
(14:47):
phenomenal scale, and so it'sit's better down now and it's
not something we're trying tosell as the new thing.
And I do believe, though,within five years, the concept
of menus, of lists and thingsand the way we display data will
just not be a thing anymore,because you'll be able to get
directly you know, with AI andthings like that to what you
(15:08):
need, without having to navigateand go oh, it's in this table,
in this location, in this area,you know.
Andrew Welch (15:16):
As the hamburger once gave way to the ellipse,
now the ellipses shall give wayto nothing to the microphone to
the microphone yeah, it'sexciting times.
Mark Smith (15:31):
I had some good side discussions around red teaming
with red teaming folks insidemicrosoft, which I think is is
probably an under investigatedarea for a lot of companies.
And you know I spoke at theCanadian Power Platform Summit
and did really my first forayinto presenting on AI, as you
(15:52):
know, publicly, after two yearsof consumption in the space, and
we talked about trustworthy AIand really I referenced a lot
the white paper you know thatyou wrote.
You released in Vegas last yearand I just see once again that
is now really becoming into amature model in that
(16:12):
organizations need to understandthat it's not just about
building a use case or aworkload right, it's so much
more Rigor needs to be putaround this.
And another talk track thatcame out a lot was, for the
first time in softwaredevelopment, the need for legal
(16:35):
risk mitigation has never beenhigher or has never been like I.
I can think of all the projectsI've been involved in the last
21 years.
At no point did we go listen,we need to get a lawyer involved
in what we're doing Never Yetin AI.
I feel that is definitely aconversation that's happening
more.
Andrew Welch (16:56):
No, totally.
And I think that therelationship between lawyers and
technologists is really goingto be changed significantly,
really going to be changedsignificantly in the coming
years.
And, funny enough, I know Iknow a guy who is a lawyer, went
(17:19):
to law school lawyer by tradewho actually then went and
learned to write code and is nowa software engineer and he's a
fascinating guy, right, likefrom that perspective.
But one piece I have my advicefor lawyers, particularly if you
are, you know you might be apartner by now, right?
(17:41):
You know, I think probably thesweet spot are partners or very
senior associates, counsel, thattype of thing, that type of
area of time in your careerwhere you have the juice and the
knowledge, the backgroundknowledge, to explore new things
and to potentially build a newpractice.
But you're not so far alongYou're not a managing partner,
(18:06):
right.
You're not so far along where,if you decided to redirect
yourself to the field of AI law,law and regulation pertaining
to AI, that somehow at your firmthe financial services practice
would come toppling down.
So if you're a more juniorpartner at your law firm, I
would be seriously consideringbranching off and becoming
(18:30):
seeking to become a leader inthe legal and regulatory
landscape around artificialintelligence and to build that
practice within your firm.
On the flip side, if you're atechnologist, you need to figure
out how to have a betterrelationship with lawyers who
specialize in this realm.
Not that you have a badrelationship you probably have
(18:52):
no relationship.
But yeah, I think that thelegal, regulatory and ethics of
artificial intelligence needs tobe baked into the platforms
that you are deploying and theworkloads that you are building
from the get-go, not treatedlike oh my god, we have to go
defend ourselves if there's acatastrophe.
Mark Smith (19:12):
You know you wrote a post last week around
Trustworthy AI that youpublished on LinkedIn, which I
think is very poignant at thistime.
I also wrote a post because Ijust felt so motivated coming
off my flight back from the US,around the and it's something
that we had discussed in thework that you've been doing in
(19:33):
the UN and around human rightsimplications of AI and more than
ever, you know, under themoniker of ESG.
We've done very good onenvironmental piece, but the
social and governance piece iskind of like, oh yeah, we don't
really like.
How does that work?
Even in Microsoft's industryplatform suite, which they have
(19:57):
a whole sustainability ESB, it'sall on the S, but there's very
low referenceability on thesocial part of ESG or the
governance part.
And I just think that inhistory we stand at a point
where I'm not going to say thatcapitalism is broken, but we
need to re-look at it in thelens of massive advantage scale
(20:22):
that you get with tech and AI.
That has never been seen inhistory before, where two or
three you you know sam altmansaid last year, february last
year, that we are going to gointo a period of time where
we're going to see one person,assisted by ai, create a billion
dollar company.
Andrew Welch (20:40):
So the labor of one person, assisted by ai,
creating a billion dollarbusiness so I I went and I found
this line, this sentence that Iwrote in that post, which we
should do an episode maybe nextepisode where we really dig into
this post in more depth,because I know we're going to
have to end this here in amoment.
But what I wrote is we mustprevent AI's maladies graduating
(21:05):
from the realm of inaccuracy tosignificant financial calamity
and again from financialcalamity to the facilitation of,
for example, cybercrime or sextrafficking.
Yeah, and what I was trying toset up in that line, in that
thought, was that a lot of folks, when they think about, when an
individual thinks about whatcould go wrong with AI, we often
(21:28):
think, oh, I mean this data,this response that AI has
returned to me, is crap, right,like, I know that it's wrong or
I know that it's misleading.
So you graduate, like themaladies of AI graduates, from
simply being inaccurate or wrongto something that is going to
cause a significant financialcalamity to whoever owns it.
(21:49):
Right Like.
So this is the chatbot I thinkit was Air Canada, they don't
quote me on that right, thechatbot that taught someone how
to reprogram a car or not toreprogram a car, how to hijack a
car, right so?
And then they were found.
The company was found liable,right.
So you go from inaccuracy tofinancial calamity, something
that an organization would bevery deeply concerned with, but
(22:11):
then to things that we as asociety need to be deeply
concerned with, right, and theexamples that I gave were the
use of AI to facilitatecybercrime or sex trafficking,
and these are, I think, veryreal.
They are very real scenarios,they are very real possibilities
.
It's something that UnitedNations human rights is very
(22:35):
concerned about, but, yeah, Ithink that there is a huge human
rights dimension to what can gowrong with AI.
Mark Smith (22:48):
Now, just on that statement that andrew said, just
I did a fact check quickly.
It wasn't air canada that didthe reprogram of the car.
What air canada did that thecourts upheld is that they must
on a refund policy that theirchatbot had issued and they
tried to say it was ah, it was achatbot, it wasn't us now.
And the courts held no, youmust do it.
The actual crossover there withthe chatbot how to hijack a car
(23:11):
was another organization, itwasn't Air Canada.
Andrew Welch (23:14):
Okay, Cool, okay , thank you.
Thank you for that.
We have live fact-checking here.
That was actually done by ahuman, but I'm pretty sure AI
assisted.
Mark Smith (23:24):
I think that's all we've got time for.
We've got so much more toexplore.
But thank you for joining us.
If you remember, in the shownotes at the top you can leave
us a voicemail.
That means your voice.
We'll put it on here and ifyou've got questions you want
areas to ask to explore, drillinto, we're so keen to hear your
ideas and opinions on thesesubjects.
(23:45):
So with that, andrew, goodtalking to you.
Ciao for now.
Bye guys.
Thanks for tuning into theEcosystem Show.
We hope you found today'sdiscussion insightful and
thought-provoking, and maybe youhad a laugh or two.
Remember your feedback andchallenges help us all grow, so
don't hesitate to share yourperspective.
Stay connected with us for moreinnovative ideas and strategies
(24:08):
to enhance your software estate.
Until next time, keep pushingthe boundaries and creating
value.
See you on the next episode.
Welcome to the Ecosystem Show.
We're thrilled to have you withus here.
We challenge traditionalmindsets and explore innovative
approaches to maximizing thevalue of your software estate.
We don't expect you to agreewith everything.
Challenge us, share yourthoughts and let's grow together
.
Now let's dive in.
It's showtime.
Welcome back, welcome back.
(00:22):
Welcome back.
We're on air again, mpp Summitis behind us, the world has
changed dramatically for a lotof us and we're on the Ecosystem
Show.
Andrew, how are you?
Andrew Welch (00:38):
The world has always changed.
For you, mark, it's always likeevery couple of weeks that we
go without recording, recordingan episode and then, yeah, the
world has changed I'm in massiveabsorption mode at the moment
what was your highlight of?
This is a leading question, but, uh, what was your highlight
from mvp summit is, to theextent that you can share yeah,
hanging out with you, of course.
(00:59):
I mean the extent that you canshare.
Mark Smith (01:01):
Yeah.
Andrew Welch (01:04):
Hanging out with you.
Of course, I mean, that's noteven that.
So what was?
What's your second choice?
Mark Smith (01:10):
Hanging out with you , of course.
Andrew Welch (01:13):
Oh, great Okay.
Mark Smith (01:14):
Well, I think we can ?
Andrew Welch (01:15):
I think we can move on Okay.
Mark Smith (01:27):
So I suppose the highlight for me was the clarity
that I feel Microsoft issettling into an AI rhythm that
is not reactionary.
Andrew Welch (01:31):
I feel this.
Mark Smith (01:32):
You know they've gone through two years of
jamming copilot into everythingand getting their branding and
go to market kind of.
It's kind of definitely been aassemble the plane while it's
flying feel I remember, do youremember, mvp Summit two years
(01:52):
ago?
Andrew Welch (01:52):
And that was very much like that was the feel
.
Though MVP Summit two years ago, I will say best MVP Summit
I've been to, because that waswhen we learned about Fabric,
but anyway, yeah, sorry, Iinterrupted you.
Mark Smith (02:06):
Yeah, the stabilization and and I feel
that you know I don't know ifit's confirmation bias, but you
and I have been talking, all ofus have been talking on this
podcast for some time now, maybeeven a year the extreme
importance of purview in the aistory, the extreme importance of
fabric in the ai story and,yeah, which is all about data
(02:32):
governance, security I meandon't get me started on intra,
id etc.
Because that all plays a partin the story as well.
Co-pilot for security got a lotof push at, I feel it, at
summit.
I just think there's amicrosoft's vision in the space.
I just feel it's on par rightwith particularly the purview
(02:52):
and the the um fabric story,which enables then everything
else yeah, well, so so it youknow.
Andrew Welch (03:03):
Interesting that you that you mentioned that
because it is this week isFabCon, or the Fabric
Microsoft's worldwide Fabricconference, and and I couldn't
be there for family reasons, butI'm quite sad to have missed it
I would love to try to make itnext year.
One of the things that I thinkis emerging and this is a story
(03:27):
that Microsoft has been kind oftrying to tell for a couple of
years since Fabric was released.
But let me go back to lastsummer.
I was working with a client,working with an organization
that is they're a big enterprisefinancial services organization
.
They are the extreme ofmulti-cloud.
(03:50):
I mean they've got everythingin there.
They have a data platformthey've built with AWS, they use
Microsoft 365.
They have Copilot for M365.
They dabble in power platform,they have ServiceNow, they use
SAP.
I mean these guys have they'rebuilding AI workloads in the
(04:13):
Google Cloud.
So this thing is thisorganization is.
You know that's a decision youcan make, right.
But the problem that they werereally wrestling with is that,
wow, we've gone, we've backedinto this multi-cloud world.
Yeah, how do we bring all ofthis together?
How do we bring all thistogether for use with AI?
(04:34):
So the one thing that I knew Icouldn't do, the recommendation
I couldn't make, and for goodreason, right, I couldn't say
well, let's rebuild your dataplatform, right, they've just
spent the last several yearsspending millions of dollars to
build, to platform their dataplatform in AWS.
So that was sort of aconstraint that I had
architecturally speaking.
(04:54):
So I pitched to them this ideaof using OneLake, which is a
part of Fabric, and I called itthe AI landing zone, which is a
part of Fabric, and I called itthe AI landing zone, nice.
And my thought was that whatwe're going to do with OneLake,
primarily, wherever we could,through shortcuts, and OneLake,
obviously or not obviously, butif you don't know, shortcuts
(05:16):
data not just from Microsofttechnologies like Dataverse or
Azure SQL, but it shortcuts datafrom Amazon S3 or from Google
Cloud, know from from Googlecloud.
So my concept was let's use anand Oracle, right, so?
So let's use off the top of myhead, do you know, I don't know
if there's a shortcut fromOracle off the top of my head.
Mark Smith (05:38):
Well, the mirroring technology came out of GA just
recently.
There's that option and I don'tknow if you've seen that the
mirroring option for a lot ofthese workloads AWS et cetera as
well is that it's free.
Yeah, have you seen that, whichis kind of like it breaks down
that barrier of no.
I've already done this in aws,already done this somewhere else
(05:58):
, so I could take advantage ofthis free method of still
bringing that data in for yourai landing zone.
Andrew Welch (06:06):
I like it yep, yep and and I I just think that
this idea of using one lake asthe sort of the final point of
data consolidation and thetechnologies that Microsoft has
developed from a cutting andfrom a mirroring perspective,
treating OneLake as that finalpoint, right, and thinking of it
(06:27):
as the AI landing zone for yourdata, or maybe the data landing
zone for AI, however you wantto couch it, or maybe the data
landing zone for AI, however youwant to couch it, it's becoming
more and more clear that thatis Microsoft's vision and that
that is a complete vision, or anincreasingly complete vision.
The world and the technology ismoving so quickly today that
(06:49):
calling anything complete is alittle tough.
But yeah, so I don't know howwe got on this riff, but it is
Fabcom this week.
It'll be really interesting tosee what comes out of it.
But I just think that the otherthing that I think makes Fabric
and OneLake specifically sounique is that there is no
competitor to it.
Yeah, there's lakes and there'sdata platforms, but this is one
(07:11):
where, for all of its faultsand to me, the only faults I
have with Fabric are that it isstill a work in progress, right,
it's like Power Platform five,six years ago.
Right, but for all of itsfaults.
It has no competitor, it has nopeer, and I think that that's
as someone who has been justobsessed with data platform tech
, like when AI really hit thescene two and a half years ago
(07:35):
and we we had chat, gpt, thatthat entered the market, which I
cannot believe.
It's been two and a half yearsnow.
I became a data platform geekfar more than I became an ai
geek.
Mark Smith (07:45):
So, um, interesting, yeah, I'm very excited about it
do you know, and, by the way,I'm going to couch what I'm
going to say is this was not anannouncement, nobody even
alluded to it, but one of mystrengths is be as is has been
able to connect futuristic dotsin the space and draw
conclusions about what I see.
(08:06):
Yeah, my observation was thisis that I think business
applications is going away atmicrosoft in the next five years
as we know it now.
I think that the dataversecomponent naturally will go
across to fabric right and andand, and that will become part
of their suite.
I then see copilot studio,which I feel is the last thing
(08:29):
standing as look what we haveinside microsoft now, and Apps
is going to go to M365.
And therefore, you know, itseems to be that's the only you
know.
The big thing that is oneveryone's radar is the value
from Biz Apps, is the CopilotStudio story, which is that, and
(08:51):
the AI mix.
So I'm just wondering.
I feel in my other observationsacross Microsoft that M365, a
lot of modern work folks alreadythought that Copilot Studio was
theirs.
They've already for some timethought that Power Automate was
right Because of the SharePointyou know history and stuff like
that.
They thought Power Apps wasbecause it lit up in SharePoint
(09:14):
with that little purple buttonthat said let's turn this into
an app.
And so I just see that there's.
You know, as microsoft goesthrough a reinventing itself as
an ai company, which it clearlyis that you know you've seen
mustafa come into the businessand it is the ceo of um ai
inside microsoft.
The whole re-engineering partof the whole engineering org
(09:37):
massively got reassigned to himin that move coming into
Microsoft.
And I'm just wondering ifthere's this, as they
rationalize their business,going forward and AI really
blends everything that I thinktheir organizational structural
models might start to change.
And hence for me you know, Isee my career, you know,
(10:00):
although it's been a powerplatform for 21, 22 years now,
it's really starting to leanmuch more to the data and AI
kind of story and the work we'redoing.
Andrew Welch (10:11):
That's a really fascinating excuse me,
fascinating question, and let'sdefinitely so.
I have some thoughts that I'llshare, but let's definitely
explore that on a future episodewhen we have more of the posse
here, because I'd love to knowwhat Anna, some combination of
Anna, chris and Will think aboutthat.
(10:32):
But know, listen, I haveconversations with I have
conversations with you know acombination.
Sometimes it's with a you knowkind of a C-level decision maker
in a Microsoft customer andthey're trying to, they're
considering which technology toinvest in.
Sometimes it's with a Microsoftpartner that is considering
(10:55):
which lines of business toinvest them in, to pursue.
Sometimes it's actually withinvestors who are looking at
buying a Microsoft partner.
Potentially right, andinvariably we talk about biz
apps.
Business applications has beenan odd fit for Microsoft for a
(11:18):
long time.
Right, like, remember,microsoft is a platform company,
so we sort of saw this comingand going right.
Like sometimes, when Microsoftwas doing really well and really
wanted to expand beyond itsborders, you'd get a lot of
interest in Dynamics.
Many of those Dynamicsapplications were actually
acquired by Microsoft.
They were not originallydeveloped in-house because
(11:40):
Microsoft felt they needed acompetitor to SAP or to Oracle
or to Salesforce or to whatever.
So BizApps have always been areally interesting fit within
Microsoft.
First of all, I think that noneof the technology within what
we think of as businessapplications, dynamics and Power
Platform.
I don't think that technologyis going away.
I do think that Dynamics isgoing to be seriously, seriously
(12:06):
rethought and not re-platformedis not the right way to put it
but the whole concept of SaaSbusiness applications are going
to be very, very different infive years.
We've talked about that on otherepisodes.
You do raise an interestingpoint, though, about our
platform.
So I want to be clear in caseanyone from Microsoft is
(12:28):
listening to this.
I'm not advocating adismembering of that group or of
that team or of that technology.
But if you do think about it,power BI already has a foot in
both worlds right, it's brandedas Power Platform, but for all
intents and purposes it's ownedin Microsoft as part of Fabric
right, as part of the dataplatform team, dataverse.
(12:48):
I think that it's becomingincreasingly difficult to
implement Fabric or Dataverseindependent of one another right
, so it makes a ton of sense forDataverse to really be the
application, the transactionaldata orchestrator of the future
for Microsoft.
And then you know, for a longtime I've said that the other
(13:08):
Power Platform components, powerApps specifically, really are
just a collection of Azuremicroservices with a marketing
label wrapped around it.
Yes, yes.
So there's a very I mean this isa really interesting
(13:30):
hypothetical going away as as aiand ai's ability to do a lot of
the things that citizendevelopers previously turned to
low code for.
Yeah, but certainly doingmaking some moves there would
really help microsoft in thisfoot into world's issue that
(13:50):
it's had for a long time.
Is power platform for seriousapplication development, which
we have long said that it istime?
Is power platform for seriousapplication development, which
we have long said that it is, oris power platform a citizen
developer tool?
Right now, the answer is both,but microsoft has long had a
marketing problem there yeah, I.
Mark Smith (14:05):
I think that like well, one year away now from
power platform being 10 yearsold, right 2016, when it came
out, I know the nomenclaturecame out of power platform in
2019, but the tooling was it wascalled.
I know the nomenclature cameout of Power Platform in 2019,
but the tooling was there.
Andrew Welch (14:17):
It was called the BAP Mark, the BAP, the
Business Applications Platform,bap.
Yeah.
Mark Smith (14:22):
And I just think that, you know, will Microsoft
365 become a platform, maybe Inthe time ahead?
It's just a positioning.
I don't think any of the tech'sgoing to go away.
I think it's like it's becomingvery mature now.
It's a proven thing, you know.
It has massive scale, massiveum companies that use it at
(14:47):
phenomenal scale, and so it'sit's better down now and it's
not something we're trying tosell as the new thing.
And I do believe, though,within five years, the concept
of menus, of lists and thingsand the way we display data will
just not be a thing anymore,because you'll be able to get
directly you know, with AI andthings like that to what you
(15:08):
need, without having to navigateand go oh, it's in this table,
in this location, in this area,you know.
Andrew Welch (15:16):
As the hamburger once gave way to the ellipse,
now the ellipses shall give wayto nothing to the microphone to
the microphone yeah, it'sexciting times.
Mark Smith (15:31):
I had some good side discussions around red teaming
with red teaming folks insidemicrosoft, which I think is is
probably an under investigatedarea for a lot of companies.
And you know I spoke at theCanadian Power Platform Summit
and did really my first forayinto presenting on AI, as you
(15:52):
know, publicly, after two yearsof consumption in the space, and
we talked about trustworthy AIand really I referenced a lot
the white paper you know thatyou wrote.
You released in Vegas last yearand I just see once again that
is now really becoming into amature model in that
(16:12):
organizations need to understandthat it's not just about
building a use case or aworkload right, it's so much
more Rigor needs to be putaround this.
And another talk track thatcame out a lot was, for the
first time in softwaredevelopment, the need for legal
(16:35):
risk mitigation has never beenhigher or has never been like I.
I can think of all the projectsI've been involved in the last
21 years.
At no point did we go listen,we need to get a lawyer involved
in what we're doing Never Yetin AI.
I feel that is definitely aconversation that's happening
more.
Andrew Welch (16:56):
No, totally.
And I think that therelationship between lawyers and
technologists is really goingto be changed significantly,
really going to be changedsignificantly in the coming
years.
And, funny enough, I know Iknow a guy who is a lawyer, went
(17:19):
to law school lawyer by tradewho actually then went and
learned to write code and is nowa software engineer and he's a
fascinating guy, right, likefrom that perspective.
But one piece I have my advicefor lawyers, particularly if you
are, you know you might be apartner by now, right?
(17:41):
You know, I think probably thesweet spot are partners or very
senior associates, counsel, thattype of thing, that type of
area of time in your careerwhere you have the juice and the
knowledge, the backgroundknowledge, to explore new things
and to potentially build a newpractice.
But you're not so far alongYou're not a managing partner,
(18:06):
right.
You're not so far along where,if you decided to redirect
yourself to the field of AI law,law and regulation pertaining
to AI, that somehow at your firmthe financial services practice
would come toppling down.
So if you're a more juniorpartner at your law firm, I
would be seriously consideringbranching off and becoming
(18:30):
seeking to become a leader inthe legal and regulatory
landscape around artificialintelligence and to build that
practice within your firm.
On the flip side, if you're atechnologist, you need to figure
out how to have a betterrelationship with lawyers who
specialize in this realm.
Not that you have a badrelationship you probably have
(18:52):
no relationship.
But yeah, I think that thelegal, regulatory and ethics of
artificial intelligence needs tobe baked into the platforms
that you are deploying and theworkloads that you are building
from the get-go, not treatedlike oh my god, we have to go
defend ourselves if there's acatastrophe.
Mark Smith (19:12):
You know you wrote a post last week around
Trustworthy AI that youpublished on LinkedIn, which I
think is very poignant at thistime.
I also wrote a post because Ijust felt so motivated coming
off my flight back from the US,around the and it's something
that we had discussed in thework that you've been doing in
(19:33):
the UN and around human rightsimplications of AI and more than
ever, you know, under themoniker of ESG.
We've done very good onenvironmental piece, but the
social and governance piece iskind of like, oh yeah, we don't
really like.
How does that work?
Even in Microsoft's industryplatform suite, which they have
(19:57):
a whole sustainability ESB, it'sall on the S, but there's very
low referenceability on thesocial part of ESG or the
governance part.
And I just think that inhistory we stand at a point
where I'm not going to say thatcapitalism is broken, but we
need to re-look at it in thelens of massive advantage scale
(20:22):
that you get with tech and AI.
That has never been seen inhistory before, where two or
three you you know sam altmansaid last year, february last
year, that we are going to gointo a period of time where
we're going to see one person,assisted by ai, create a billion
dollar company.
Andrew Welch (20:40):
So the labor of one person, assisted by ai,
creating a billion dollarbusiness so I I went and I found
this line, this sentence that Iwrote in that post, which we
should do an episode maybe nextepisode where we really dig into
this post in more depth,because I know we're going to
have to end this here in amoment.
But what I wrote is we mustprevent AI's maladies graduating
(21:05):
from the realm of inaccuracy tosignificant financial calamity
and again from financialcalamity to the facilitation of,
for example, cybercrime or sextrafficking.
Yeah, and what I was trying toset up in that line, in that
thought, was that a lot of folks, when they think about, when an
individual thinks about whatcould go wrong with AI, we often
(21:28):
think, oh, I mean this data,this response that AI has
returned to me, is crap, right,like, I know that it's wrong or
I know that it's misleading.
So you graduate, like themaladies of AI graduates, from
simply being inaccurate or wrongto something that is going to
cause a significant financialcalamity to whoever owns it.
(21:49):
Right Like.
So this is the chatbot I thinkit was Air Canada, they don't
quote me on that right, thechatbot that taught someone how
to reprogram a car or not toreprogram a car, how to hijack a
car, right so?
And then they were found.
The company was found liable,right.
So you go from inaccuracy tofinancial calamity, something
that an organization would bevery deeply concerned with, but
(22:11):
then to things that we as asociety need to be deeply
concerned with, right, and theexamples that I gave were the
use of AI to facilitatecybercrime or sex trafficking,
and these are, I think, veryreal.
They are very real scenarios,they are very real possibilities
.
It's something that UnitedNations human rights is very
(22:35):
concerned about, but, yeah, Ithink that there is a huge human
rights dimension to what can gowrong with AI.
Mark Smith (22:48):
Now, just on that statement that andrew said, just
I did a fact check quickly.
It wasn't air canada that didthe reprogram of the car.
What air canada did that thecourts upheld is that they must
on a refund policy that theirchatbot had issued and they
tried to say it was ah, it was achatbot, it wasn't us now.
And the courts held no, youmust do it.
The actual crossover there withthe chatbot how to hijack a car
(23:11):
was another organization, itwasn't Air Canada.
Andrew Welch (23:14):
Okay, Cool, okay , thank you.
Thank you for that.
We have live fact-checking here.
That was actually done by ahuman, but I'm pretty sure AI
assisted.
Mark Smith (23:24):
I think that's all we've got time for.
We've got so much more toexplore.
But thank you for joining us.
If you remember, in the shownotes at the top you can leave
us a voicemail.
That means your voice.
We'll put it on here and ifyou've got questions you want
areas to ask to explore, drillinto, we're so keen to hear your
ideas and opinions on thesesubjects.
(23:45):
So with that, andrew, goodtalking to you.
Ciao for now.
Bye guys.
Thanks for tuning into theEcosystem Show.
We hope you found today'sdiscussion insightful and
thought-provoking, and maybe youhad a laugh or two.
Remember your feedback andchallenges help us all grow, so
don't hesitate to share yourperspective.
Stay connected with us for moreinnovative ideas and strategies
(24:08):
to enhance your software estate.
Until next time, keep pushingthe boundaries and creating
value.
See you on the next episode.
Popular Podcasts
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.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!