October 22, 2025 • 36 mins
It’s been a year since AI agents burst onto the tech scene, driven largely by the vision and product roadmap of Salesforce founder Marc Benioff, who is betting the company’s future on the “agentic enterprise”.
Saleforce’s Agentforce platform, which allows customers to build and deploy agents in days or weeks, automating aspects of sales, customer service, marketing, and even software development, has racked up 12,000 customers in its first year.
But Benioff said last week in San Francisco that there’s a gap between the innovation AI companies are progressing, and the business sector’s ability to adopt AI at scale. The halls of the Moscone Centre in San Francisco were full last week of early adopters, companies that have deployed AI agents with promising results, from PepsiCo, to Lululemon.
Among them was One NZ, the telco that has spent years modernising its tech platforms and which uses Salesforce for its customer relationship management.
"Already across our AI investment, five times return," One NZ CEO Jason Paris told me on this week’s episode of The Business of Tech, describing how One NZ deployed an agentic tool to help prepaid customers migrate to better mobile plans.
“We've been deploying AI and all the variations of it, so robotic process automation or machine learning, for over 10 years. We've only been deploying agentic for last year, because no one even really knew about it 12 months ago,” he said.
“But yeah, five times ROI on our AI investment. So a lot of people are saying, is the value there? We can see the value. Absolutely. And we think this is just the beginning. We want more.”
Orchestrating multiple agents
The agent authenticates customers, presents personalised plan options, explains trade-offs, and completes the entire transaction autonomously via chatbot – all in a matter of minutes. Previously, a customer may have needed to make a phone call to the One NZ contact centre or visit a store to achieve the same outcome.
Paris added that the newly deployed Agentforce agent has delivered a 400% improvement in customer engagement compared to traditional digital journeys. "It took us five weeks to build and deploy," Paris explained, noting that the first agent One NZ built with Salesforce took just eight hours to create, though two additional weeks were needed for data integration and cleanup.
Sitting behind that simple chatbot designed to help prepaid mobile customers find a new plan, are actually seven AI agents working together.
"We're in this age now where this one agent can orchestrate multiple actions,” said Hamish Miles, Salesforce New Zealand’s managing director.
“In [One NZ’s case] it's like… a security check in for the trust layer. It's, have we got the right plan? Can I make some recommendations? We can do so much more as well."
Miles acknowledged that addressing Benioff's innovation-adoption gap requires overcoming hesitation. "I think there's probably a little bit of reluctance to make a start," he said, noting that companies often worry about data quality. His advice is to at least get going on creating AI agents.
"Make a start, because it's a low-risk entry. Is the agent going to be perfect straight away? No, it won't be. But will it learn? Yes, it will. Can we make corrections very quickly? Yes, it will. So start experimenting," he said.
Salesforce itself now runs more than 200 internal agents managing 550 different tasks. Mark Benioff says use of AI agents is saving the company $100 million annually, much of it in customer support costs, including reduced headcount.
Voice - the next frontier in AI agents
The next frontier for AI agents is voice. Paris described the upcoming product Agentforce Voice, which enables natural conversations with AI agents over the phone, as "a game changer" and "probably, of all the things that were discussed, personally, [what] I'm most excited about".
With up to 80% of customer service interactions in some industries still coming via phone, voice-enabled agents could fundamentally transform customer experience.
"If I can have a conversation where there's no latency, it's a humanised conversation that you can build rapport through transparency that you're talking to an agentic tool, it's probably an easier step for me to just talk to an AI agent in the same way I would talk to an agent in a call centre," Paris explained.
As New Zealand's economy continues to face headwinds, the ability to simultaneously reduce costs and increase productivity through agentic AI presents a compelling opportunity, but many businesses still need to get their data and IT infrastructure in place to deploy AI ag
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Welcome to the Business of Tech, powered by Two Degrees Business.
I'm Peter Griffin, coming to you from San Francisco, where
I've been hanging out with well around fifty thousand other
people in town for the annual dream Force conference put
on by software company Salesforce, and it's flamboyant founder Marc Benioff.
I've always liked getting to dream Force because Benioff, You've
(00:25):
got to hand it to him, puts on a great show.
He always gets big names to talk at his conference.
This year's highlights included Google CEO Sundaipachai, President Trump's aisar
and member of the PayPal mafia David Sachs, and an
unexpected one, Pete Bodhage Edge, the former mayor of South Bend, Indiana,
who ran in the Democratic presidential campaign in twenty nineteen
(00:50):
before dropping out and serving as President Biden's Transport Secretary.
Bootage Edge is still tipped as a future potential candidate
and his talk was really fascinating. I've included a clip
of it at the end of the podcast. Dream Force
was dominated once again by artificial intelligence. It's a year
since Salesforce debuted Agent Force and look Salesforce is a
(01:12):
tech company that is probably most enthusiastically embraced AI agents.
These are autonomous software programs that perceive their environment, make decisions,
and take actions to achieve goals, often with minimal human intervention.
Agents are particularly useful for automating aspects of customer service
(01:33):
and support, sales and marketing, which is why Salesforce has
gone all in on agents. It has twelve thousand Agent
Force customers after its first year. It's around eight percent
of the entire Salesforce customer base of companies, which a
number of analysts pointed out is relatively low uptake, but
(01:53):
as Benioff countered, it's a heck of a lot more
than any other tech company is managing at the moment. Still,
despite the AI boom, some would say bubble that we're in,
Salesforce has actually seen at share price retreat twenty three
percent this year. It claims it's using AI to reduce
one hundred million dollars in costs across the company each year.
(02:14):
It shared four thousand staff, many in customer support roles
by shifting to AI agents, and it told the share
market it expects to return to double digit growth out
to twenty thirty, when it expects to rack up sixty
billion dollars in revenue, up from around thirty eight billion
last year. That growth will largely come on the back
(02:35):
of businesses taking up AI agents. One thing that only
got a fleeting demo in the Dreamforce keynote, But one
of the things that Salesforce is piloting, which is really interesting,
is Agent Force Voice. It's like calling up a call
center where you might typically get a human on the
line or maybe a voice activated menu. Now you'll potentially
(02:57):
get an AI agent. You can talk to one that
can understand what you want. It might be changing your
broadband plan, returning some clothing that doesn't fit, reporting an
electricity outage, or buying a plane ticket. Given that up
to eighty percent of interactions with customer service in some
industries is still coming in via phone lines, that would
(03:19):
be huge. If it works as envisaged, that's a game changer.
More about Agent Force Voice coming up in the show.
At dream Force, I also got around a lot of
companies that are using agents to automate various processes. PepsiCo
is using agent Force to let its smaller merchants more
easily check on the status of their orders just using
(03:40):
a simple chat based agent. Lululemon, the active where company,
is offering customers the chance to use an agent to
customize a wardrobe through an app based experience and then
go on to purchase what they've been looking at. Pearson's,
the education provider, is letting students use agents to change
courses and by textbooks. I saw a really cool agentic
(04:03):
service from Chicago Medical University to give patients one chat
based interface to order prescriptions, change appointments, get their medical
records delivered to them. That would remove a heck of
a lot of hassle dealing with this sort of gap
we have between patient portals which we all sort of
have in New Zealand now, and the mirrid range of
healthcare professionals who actually deliver frontline services. It can be
(04:26):
frustrating actually going from one to the other. Good to
see some New Zealand companies one in Zed, Fisher and
Pikeland Zero among those being showcased at Dreamforce for Innovative
Uses of AI agents. So you're hearing the piece from
Jason Paris, who I caught up on a show floor
at Dreamforce where one in Zed had a display and
demo area. But first I wanted to invite back onto
(04:47):
the show Hamish Miles, the managing director of Salesforce New Zealand,
to get his take on what's transpired since Agent Force
unleashed the AI agents almost exactly a year ago. Here's
Hamish Miles from s Salesforce. Okay, so Hamish, welcome back
(05:10):
to the business of Tech.
Speaker 2 (05:11):
How are you doing.
Speaker 3 (05:11):
Yeah, awesome, It's been an epic few days. I'm absolutely
pumped and looking forward to what comes next.
Speaker 1 (05:18):
Yeah, we're on the roof of the Muscone Center in
San Francisco. Dream Force, the biggest conference in San Francisco,
the biggest tech conference, has been going all week and
interesting what you say is that the highlights obviously Agent
Force three sixty is the centerpiece of this and to me,
it seems like we're a year into this agentic ai
(05:40):
movement that Salesforce really led you back in October twenty
twenty four. It's sort of moved from being a product
to a platform. This is the systematizing of agentic ai
into everything that Salesforce does.
Speaker 3 (05:54):
For me, it's been an absolutely fascinating few days and
you can see the progression of the journey I've been
GenTech AI and you can see some of the great
outcomes our customers are starting to produce. I think Agent
four three sixty is a natural next step for us.
It makes a lot of sense to our customers. What
excites me the most is the agentic enterprise. Where are
we going to go next? And then the combination of
(06:16):
humans and AI working together things an exciting time for
the industry and we're just starting right We're just scratching
the surface.
Speaker 1 (06:22):
Twelve thousand customers, I think half of them are paying
customers are using agent Voice. Now, what's been the adoption
like in New Zealand.
Speaker 3 (06:31):
You know, I think we're at the front edge as well.
I've got some great customers doing some great things. I mean,
you've heard from Jason Paris at one and said, that's
an agentic journey, that prepay agent does multiple things all
in one journey, all in a few minutes for Shre
and Pike or on the path with field service agents
and are scaling the premium product with Premium Service zero
(06:51):
are going to do some exciting things very shortly.
Speaker 1 (06:53):
And it seems to me, like Jason said, the customer
facing ones, which typically will look like a chatbot or
maybe a voice call. So but he said, you know
that prepaid product shifting sort of agent that they've created.
If you on a mobile plan and you want to
move to another one, and this will do it all
automatically and suggest the right plan for you. But he said,
there's actually seven agents in the background of that, So
(07:14):
all the handoffs and querrying data from one ZED system,
there's actually a whole bunch of agents doing that in
the background.
Speaker 3 (07:22):
Yeah, So they were sort of looking at the first
stage of orchestration of agents in that sense. So the
first sort of a augentic agents we spun up. They
were sort of sort of single task orientated. Salesforce has
more than two hundred and two hundred and ten.
Speaker 2 (07:34):
Agents live at the moment.
Speaker 3 (07:36):
Internally, I think there are five hundred and fifty sort
of tasks that it can carry on. We're in this
age now where this one agent, like the one New
Zealand or one assystem I think it's called one assistant,
can orchestrate multiple actions. And in their case it's like
fraud check. It's a security checking for the trust layer.
It's have we got the right plan? Can I make
(07:57):
you some recommendations? And it's absolutely fascinating and we can
do so much more as well.
Speaker 1 (08:03):
Before I go any further with Hamish, I thought it'd
be good to bring in Jason Parris from One end Zed,
who I caught up with at Agent for City in
the depths of the Moscone Center where all of these
demos of agents are being shown. So he gave me
a really good rundown on what exactly one end Zed
was doing with these AI agents to help roll out
(08:23):
a new service for its prepaid customers. So tell us
about how you're using agentic.
Speaker 4 (08:29):
Enterprise, Jason, So, Peter, this is something that's just gone
into production. We've been working with Salesforce and deploying their
agent Force technology for over a.
Speaker 5 (08:39):
Year now and this is the latest agent.
Speaker 4 (08:42):
And So from a tallocommunications perspective, one of the most
challenging journeys that a customer can go through is migration.
So when you're moving from a plan that you probably
are quite happy with to a new plan that we
think you should be interested in, it's quite often a
complex and clunky process or a customer to go through.
Speaker 5 (09:01):
And it's also a bad.
Speaker 4 (09:02):
One for a talco because you can get that reconsideration moment,
and you want the journey to be so seamless and
effortless that you love the Talco even more at the
end of it. And so we've deployed an agentic agent
in our business, which is helping hundreds of thousands of
our customers move to.
Speaker 5 (09:24):
A better prepaid plan for them.
Speaker 4 (09:26):
We've started with our most difficult customers, tens of thousands
of those and four hundred percent improvement and engagement on
the agentic journey compared to the journeys that we would
traditionally have used through normal retail or call center or
even our online website presence.
Speaker 1 (09:47):
So is the customer interacting with the agent directly, yes,
by a chatbody or something.
Speaker 4 (09:53):
Yeah, So via the agent itself, right, it'll authenticate you
to know who you are. It'll know so the plan
that you're on currently, but also the next best plan
for you. It will present a carousel of options that
you can interact with, and the agent will talk to
you about the pros and cons of the plan compared
(10:14):
to what you're currently on, and then we'll ask you
to confirm and then once you're happy with.
Speaker 5 (10:18):
It, it'll tell you that you're done.
Speaker 4 (10:21):
Right and await, it'll have a record for you and
for us of what we've agreed should anything change, and
then also the agent will ask if it can help
you with some other frequently asked questions of someone who
is a mobile customer that might want to know. So
that part of it we're going to continue to build
out and approve on.
Speaker 5 (10:41):
But the upgrade path for a customer.
Speaker 4 (10:47):
Seamless, and it took us five weeks to build and
deploy it.
Speaker 1 (10:51):
Right so, and doesn't require a lot of coding or
anything like that. It's all drag and drop, build your
own templated agent building.
Speaker 4 (10:58):
Yeah, so some configuration, but it's mainly around data cleaner,
making sure the data is accurate, integration with our existing
it the Salesforce Agent Force product set are mainly taken
and deployed.
Speaker 5 (11:14):
The time to deploy it probably is build.
Speaker 4 (11:16):
It's making sure you understand what the journey is and
the integration with your own rest of your technology and
the data cleanup. Like the very first agent we built
with Salesforce took us eight hours and then two weeks
to deploy again because of the data and the integration
with another to other technology.
Speaker 1 (11:34):
Right So, those tens and thousands of problematic customers there,
ones that are may be about to churn off one
into the you're going to go somewhere else, So that
the agent is looking at all the data that sits
in Salesforce about that customer, going okay, I know what
you need. Here are some suggestions.
Speaker 5 (11:50):
Yeah, well that's the Salesforce Marketing Cloud.
Speaker 4 (11:52):
So an another tool within the ecosystem that helps us
look at our data set, query customers into different SEGM
and cohorts and put exactly that brief in and then
it will give us that subset. And so you know
that when you are trialing an agentic tool with your
most difficult customers, if it works for them, it's going
(12:13):
to work for everyone. So you always start with the
naliest ones first, normally to give you confidence that when
you do decide to scale it from tens of thousands
to hundreds of thousands to millions of customers, you've got
that clients.
Speaker 1 (12:25):
So like, if I'm looking at changing plans, I'll go
to the one in ZED website and I'll see all
the panels here. If the different are the broadband or
mobile plans that you've got, so in the near future
you're more likely to just go straight to an agent
and say this.
Speaker 2 (12:38):
Is what I need.
Speaker 1 (12:39):
What have you got for me?
Speaker 4 (12:40):
Yeah, you will, and we'll be transparent about it as well,
by the way, so it's not as if you'll think
you're talking to a human.
Speaker 5 (12:46):
You'll know that you're talking to an agentic tool.
Speaker 4 (12:50):
Increasingly, because they're doing such a good job, we think
that our customers will go thank God, because I don't
want to talk to those use of body humans. I'd
like to talk to more efficient and so that's exactly
how that will work. And then what we will do
is we'll keep the human conversations to the probably more
important ones. So if I really have a complex technical
(13:11):
issue that I need.
Speaker 5 (13:12):
Some help with, or I want to talk to someone.
Speaker 4 (13:14):
About upgrading my handset because it's going to cost me
three thousand dollars to do that, or maybe unfortunately I'm
calling because there's been someone who's l or under financial
stress or has passed away in the family, those are
the moments where you really want to have a conversation
with a person versus a tool. So we're deploying the
(13:36):
technology or the agentic technology on the areas where really
humans shouldn't be doing that work, but it's just by
default that we're having to throw a lot of people
at it. Now we can put those people onto more
important stuff and let the agents do the hard yards
for us.
Speaker 1 (13:54):
Yeah, so that the agent will complete the transaction. So
if you decide yes, I do to move on to
a different plan, that will do that automatically.
Speaker 5 (14:02):
Yeah, that's great.
Speaker 1 (14:03):
And then so everything in the back end in terms
of your record of what plan they're on, that's all
updated automatically by the agent.
Speaker 4 (14:10):
Yeah, correctly, So into our CRM Salesforce tool at the
same time. So you've got the ecosystem of the traditional
customer relationship management, which we're moving.
Speaker 5 (14:21):
On to Salesforce.
Speaker 4 (14:22):
So all of our prepaid customers, all our products, all
our plans are now on the Salesforce CRM, and then
you apply the agentic player on top of it, so the.
Speaker 5 (14:31):
Intelligence layer on top of it.
Speaker 4 (14:32):
Yeah, where you've got a whole bunch of agents performing
a whole bunch of tasks and one of them is
a plan upgrade or a customer migration. So end to ends,
the ecosystem has a record and that the.
Speaker 5 (14:45):
Customers clear about.
Speaker 4 (14:46):
The only thing is there's disco movement, dancing and music,
and that doesn't happen when you're on our website.
Speaker 1 (14:52):
Right right, Yeah, Well, interestingly, you know they have been
talking here about Agent Force Voice, which they're piloting get
a lot of coal traffic. Still people want to call
up and then you know, if it's overloaded, you might
get a callback option or something like that. What do
you think about the option of being able to talk
to an agent's and you know how hard it is
(15:14):
to get voice recognition and stuff really accurate.
Speaker 5 (15:17):
Do you see promise in that it's a game changer.
Speaker 4 (15:20):
It's probably of all the things that were discussed yesterday
that personally I'm most excited about. I genuinely believe that
voice to agentic tools is the game changer because you
think about, I don't know, an eighty year old who
might be really kind of resistant or nervous about typing
(15:40):
something into an AI which is perceived to be a bot,
Whereas actually, if I could have a conversation where there's
no latency, that it's a humanized conversation that you can
build rapport through transparency that you're talking to an agentic tool,
it's probably an easier step for needed just to talk
(16:00):
to an AI agent in the same way I would
talk to an agent and a call center compared to
being forced down a text or social media type conversation.
And I think it's going to be way more efficient,
like even the way that I use my own tools.
Speaker 5 (16:14):
Now, if I'm talking to.
Speaker 4 (16:16):
Chat GPT, I'm not texting you speak of having a
conversation with it. Yeah, And I think even that's what
ob an Ai has seen over the last three years
of when the first versions of chat GPT, it was
all people were using it like a Google search query.
Speaker 5 (16:31):
Now it's a conversation. And I think that conversation can
move from a text.
Speaker 4 (16:36):
Based one to a voice based one with lower latency.
Speaker 1 (16:40):
It's going to be a game change dealing which salesforce
say they've been able to cope with background noise, distracted callers,
QWI accent, all that sort of stuff. If we can
nail all of that.
Speaker 5 (16:51):
Yeah. And also I think talking about the ki.
Speaker 4 (16:54):
Accent, the ability for you to talk in different languages
and understand each other at the same time.
Speaker 5 (17:00):
How cool is that?
Speaker 4 (17:01):
So you think the possibilities are endless, And I think
that the voice upgrades is going to be a really
big feature.
Speaker 5 (17:09):
And it's one of the things I've been talking to.
Speaker 4 (17:11):
Salesforce about over here is how can New Zealand be
a world first in Talco for the service.
Speaker 5 (17:17):
So you look around here and you look at the brands.
Speaker 4 (17:20):
I am really proud that one New Zealand is the
only Talco here. But also the way that we get
the attention of a company like Salesforces by moving a pace.
And so it's our job to say, serves Salesforce has
got a cool feature, let's experiment.
Speaker 5 (17:36):
With it and roll it out in New Zealand.
Speaker 1 (17:38):
First, where do you see the other real areas of
promise for agents across one in z Well?
Speaker 5 (17:44):
I think actually sales is probably the next next area.
We've been focusing a lot on service.
Speaker 4 (17:52):
But again kind of sales is in the DNA of Salesforce, right,
It's in the name, and so everything from being able
to pre qualify leads before our salespeople talk to the
customer to be able to capture the information I've got
the customer wants, the size of the business, the complexity,
and then also do some sales coaching for our teams
(18:15):
next best action that they might want to take all
of those things as a tool that can sit beside
your salesperson to do a lot of the heavy lifting
before they actually need to talk to the customer, but
also sales coaching to get higher conversion as part of
that conversation.
Speaker 5 (18:32):
That's something the next big wave.
Speaker 4 (18:34):
So we've already using salesforce within marketing, we're using salesforce
within service and that great journeys and now using I
think salesforce and from a sales perspective, whether it's any segment,
consumer or small business or large enterprise. That's the next wave.
And it's proven technology, which.
Speaker 5 (18:52):
Is why we're so excited to roll without.
Speaker 1 (18:55):
Yeah, and you're seeing the return on investment already.
Speaker 5 (18:58):
Across our AI investment five times return wow. Yeah.
Speaker 4 (19:01):
And so we've been deploying AI and all the variations
of it, so robotic process automation and machine learning for
over ten years.
Speaker 5 (19:09):
We've only been deploying agentic for last year.
Speaker 4 (19:12):
Because no one even really knew about it twelve months
twelve months ago. But yeah, five times ROI on our
AI investment. So we a lot of people are saying
is the value there? We can see the value absolutely,
and we think this is just the beginning.
Speaker 5 (19:29):
We want more.
Speaker 1 (19:37):
So there are these early adopters and one inserts a
real example of that. Jason said, for every dollar he's
investing in AI, he's already getting a five dollar return
on that investment. So that's incredible. So where is that
coming from. That's obviously efficiencies and is it people. It's
people not having to do all of that work and
(19:57):
you can deploy them to somewhere else.
Speaker 3 (19:59):
That agent on twenty four to seven and when it launched,
it launched at four o'clock in the morning, and it
had nine conversations by the time the core center had
come alive and nine engagements a week later. We're talking
about four hundred percent improvement and conversion of plans, you know,
getting people off old prepaid plans onto the new plans
that they want to promote, and a four times improvement
rate on the traditional rate. I know that sounds ironic,
(20:20):
traditional journey digital journeys that they have by the web.
I think they've still got the same head count the
people doing it. So it's augmenting humans to make them
more efficient and it can be open twenty four seven,
which as a consumer, that's a great thing.
Speaker 1 (20:33):
What do we need to do in new Zealand to
address this issue that Mike Benioff raised, you know, in
his keynote on the first day, this gap between incredible
innovation that is moving at pace and this sort of
lag and adoption. The companies are just struggling a little
bit to keep up. Is it purely an education issue
or do they not have the confidence and frankly the
(20:55):
money at the moment and tight times to say we're
going to invest in this.
Speaker 2 (20:59):
So I'll park the.
Speaker 3 (21:00):
Training sort of side, because I think skills is really important.
Speaker 2 (21:03):
We need more skills in the market.
Speaker 3 (21:05):
But I think there's probably a little bit of reluctance
to make a start, and I think people think about data,
but data doesn't have to be perfect. You've probably got
three areas of data goal bronze, silver and gold and
your account information. Well that's goal. We want to wrap
that one. Now, that's true, but marketing messages, et cetera.
Data doesn't have to be perfect and it never will be.
And you've worked out on your journey. But I think
(21:26):
most importantly you can start, make a start because it's
a low risk entry. Is the agent going to be
perfect straight away? No, it won't be, But will it
learn yes, it will. Can we make corrections very quickly, yes,
I will. So start experimenting, do something internal first before
you go external, and you'll incremnially keep on improving. And
I think for New Zelling customers we need to really
(21:47):
keep moving already.
Speaker 2 (21:48):
Look with this customer, we're seeing.
Speaker 3 (21:50):
Great productivity gains, like great productivity leagain, so it's cost effective.
It's also seeing revenue increase and productivity. And I think
it's a pretty strong message to share a New Zealand.
Speaker 1 (22:01):
So you know there's Data three sixty which is data Cloud,
so that's all been agentized and all that, so it's
all there. If you want to put your data on Salesforce,
you can do all of this stuff enable it in
in a very simple way. A few of the other
sort of innovative things that I think have come out
of this week is the move into voice agent Force
voice so if you can power a voice interaction with
(22:24):
a customer with an agent's and it'd be a good experience.
Have you seen that demo.
Speaker 3 (22:30):
Or have you used We've seen it demode and we're
looking into the first few New Zealand customers to get
on and start trialing it right and that's happening shortly so,
but I won't talk about it today, but it's going
to be really impactful because we want to meet the
customer where they are in the journey, and quite often
that's voice. It's a voice conversation, could be a text,
(22:50):
and it could be a digital channel, and it could
be an agent, but more often than not, a still voice.
It's I think seventy eighty percent for some businesses that
are coming in and protectingly for telcos.
Speaker 2 (23:00):
Thin think about having a digital twin.
Speaker 3 (23:01):
If you're in the contact center, my name's Maddie, and
I create a digital twin that's on twenty four to seven.
I can have a personal relationship with you, Peter, twenty
four seven. I don't even have to be at work
and to answer some of your questions, and so you
actually feel like you will have a personalized service with
your digital assistant who has actually got a human behind it.
That I think for a lot of people will be
(23:22):
very very good experience and quite comforting to have that
consistent service because you know when you ring up, you're
a different person every time.
Speaker 2 (23:28):
Now I want to speak to Maddie.
Speaker 1 (23:30):
Yeah, it's going to be interesting to see it at scale,
particularly with the Kiwi accent.
Speaker 3 (23:34):
How Yeah, So that is a really because that's when
you get into you know, people talk about large language
models and contextual stuff, so that accent, that stuff is
going to be really important. But also picking up other
languages as well. You know in New Zealand we just
you know have TODAYO and we have all those languages
from all the people come from other parts of the world. Yeah,
(23:55):
pacifica language in Latin America, Asia.
Speaker 2 (23:58):
Yeah, it's going to be interesting.
Speaker 1 (24:00):
That is going to be really interesting. A couple other
things move into it service management. You know, that's obviously
was a bit of a gap there, but as Bennyov said,
you know from it query support, queries, organizing field staff, technicians,
all of that sort of thing that's an area Service
now is really good at. But it is going to
be possible now on the Salesforce platform.
Speaker 3 (24:21):
Too, you know, because we've always been sales service and
marketing very externally focusing. It's not that much different from
our service operations and you know, raising tickets and distributing,
so it's quite a natural adjacency for us to enter
and it's going to give customers that singular platform opportunity
if I raise a field service ticket or I raise
a ticket which turns into a field service action. You know,
(24:42):
that could be a tech person or it could be
a contractor. And as a consumer, I don't really care.
I just want that service, and I think that we
lean naturally into that, so you know, it's a great
option for our customers.
Speaker 1 (24:53):
Any use cases you saw as you walked around Agent
Force City downstairs that you thought, Wow, that's its pretty cool.
Speaker 3 (25:00):
I've enjoyed, and excuse the pun here, watching the journey
of Heathrow Airport. You know, I've been on planes and
thought I've missed a gift opportunity for family member that
might have had a birthday or something like that, a
moment in their life that I so the opportunity to
interact with their agent. Also know, I need a gift
and I need to go to Judy Free. Can you
(25:22):
have that ready for me? And I'm running late already
and having that shopping consumer experience at the airport making
life really easy. And also like once say no, I
booked publishing data, pushing data to meet your flight's on time.
It's going to be at this gate. Oh, by the way,
we've changed the gate. Just making life easy at an airport,
because airports can be stressful. Yeah, I just think that's
a wonderful story.
Speaker 1 (25:43):
Yeah, and many the what I've seen, like Chicago Medical University,
a huge health provided as four billion dollars in revenue,
and like us in New Zealand, they're using patient portals.
You can you can get all of that, but every
time you so you need to talk to your doctor
or a nurse, it's sometimes it's by a text message,
(26:03):
sometimes it's a phone call, sometimes it's an email. It's
just all scattered. Bringing that all into one interface where
you can query it through chat. They just think it's
going to cut their call volumes by seventy percent.
Speaker 3 (26:16):
I can play that foward to that digital twins. So
the person that's with you on that journey. Yeah, because
sometimes when people go through these procedures, Yeah, very stressful,
lots of information, someone coaching you through. Yes, your appointments
here in two days, you're going to be this. This
is what they're going to be talking about. This is
what was talked about in the session. Yeah, So it's
I think I said it. I think we're I'm not
even sure we're scratching the surface yet. I mean, I
(26:37):
think the opportunity is going to be huge and when
we sit here in ten years time, we'll look back
and go, Wow, what a start. Yeah.
Speaker 1 (26:43):
And you know, we touched on the capability issue and
the training issue, and I think that is a bit
of a gap, particularly in New Zealand. I mean there
is this perception it's all sort of vibe coding. You
don't need that much expertise to do it, but you
sort of do to do it properly and to put
in the right governance around it, and just have the
confidence across the business for people to go I want
(27:04):
to go in this journey. And what I've learned this week,
it's got to be from the leadership of the company
saying we are going to be in like Jason Paris
and said, we're going to be in AI centric talco
and the first one in New Zealand. To do it,
you sort of have to have that approach. What do
we need to do to get there?
Speaker 3 (27:18):
In New Zealand we have a skills gap and the
tech sector and I think some support around that would
be amazing. Lo we have all these programs online and
probably the entire industry feels the same way. So some
assistance in getting more people into the tech secer, and
we know it transforms people's lives, right, they can go
from minimum wage to double that quite quickly.
Speaker 2 (27:37):
So that's one aspect.
Speaker 3 (27:38):
And I think to your point around what do New
Zealand companies need to do, You're absolutely right. Leadership needs
to come from the top. And if you look at
any successful transformation project over the last ten to twenty
years in technology, the one core ingredient that I had
that it got right was governance, and leadership came from
the top. If you look at all the ones that
have missed and delayed, and you know there is a
(27:58):
graveyard of transformation projects that have missed their deadlines, they
have a common theme that the leadership wasn't all on board.
Speaker 2 (28:05):
I think as we move into.
Speaker 3 (28:06):
The surgetic enterprise, it'll break down the silos of how
we work. You know, relationships will become absolutely critical. What
does the agent do, which is really task and function
and high processes and orientated to what do the humans
do then, which is relationships, empathy, context and that constant
training and oversight. So governance leadership really important and.
Speaker 1 (28:29):
We just need to start I mean it's still pretty
soft in New Zealand's economically hoping we'll turn our coin
interest in what you're seeing, I mean that rich data
coming from your customers as to how they're doing, and
there seems to be an opportunity here to cut cost,
to use salesforce and agents to actually take some costs
(28:52):
out of your business, which is what they need to
do at the moment.
Speaker 3 (28:54):
Everyone's quite cost conscious, but they're also productivity oritented. So
how do you do both? And I think the opportunity
is that you know we can. You can reduce cost,
you can increase your productivity, and typically that means improving
your top line.
Speaker 2 (29:08):
I think you're right.
Speaker 3 (29:09):
I think you know there's no secret that the economy
is sort of bouncing around a wee bit at the moment.
We're hoping to see it continue to improve. But the
customers that we've got that are embracing the agentic journey
are seeing some pretty good results early in the piece,
which is encouraging. And I would love to see more
of New Zealand. I think we New Zealand can lift
if we embrace this agentic journey. Yeah, because you will
get cost out and you will get productivity up and
(29:30):
you probably get revenue up.
Speaker 1 (29:31):
It's good to see the governments just put seventy million
into literally asking for AI centric projects through the new
Advanced Technology Institute, which is the first big investment in
AI at a government levels. That so, I guess is compelling.
Speaker 2 (29:45):
Yeah, and I.
Speaker 3 (29:46):
Think we need to continue that push for training, right,
training and skills in the industry because that will help
New Zealand lift that will benefit all of our businesses,
central government, the vendor, community, partners and you know, the
community in the end.
Speaker 2 (30:01):
Yeah.
Speaker 1 (30:02):
Just finally, the other thing that struck me this week
is around Slack.
Speaker 2 (30:06):
Oh yeah, this is cool.
Speaker 1 (30:07):
Yeah, and Slack sort of being like I've used it
not in the salesforce domain, but you know, as a
lot of media companies use slack great for messaging, but
it's become more than that, hasn't. It's become the front
end really of sales. That's quite what it was originally
found for. But right now it's the central console for
people to go and get their work.
Speaker 2 (30:27):
Now.
Speaker 3 (30:28):
I made a quote earlier two before, you know, two
hundred and thirteen agents, five hundred and fifty tasks. I
got that through Slack because Jason was asking me a question.
He goes, how many agents you've got? I went, so
I went to slack pot, which I'm now using more
than Gemini to ask questions about internally, how many agents
we have.
Speaker 2 (30:43):
What are they doing?
Speaker 3 (30:43):
So we're surfacing all that Argentic layer through Slack. Now
that won't be just us, you know, we've got partnerships
with work Day and Slack will become that console for
where you go to get that information.
Speaker 1 (30:55):
Is New Zealand a big, big market for Slack?
Speaker 2 (30:58):
Are there a lot of customers there? Yeah, we have
some great customers.
Speaker 3 (31:00):
I think Zero is a huge user of Slack, huge
user and there've been a reference for us for a
while McCloud cranes. Right, So then the other end of
the scale in the ESMB space an organization out there
lifting and shifting parts of New Zealand around with cranes.
So it's something that can fact any price that we
have Ryman homes and you know when they change rooms around,
(31:21):
they use that Slack as their workflow when they start
updating their villages. Yeah, yeah, it's cool. So that has
come a long way. Yeah, Okay, well good luck.
Speaker 1 (31:31):
It's going to be really interesting to see in it
twelve months time, where the heck we're at Given what's
changed in the last year, it's.
Speaker 2 (31:39):
In the next twelve months.
Speaker 3 (31:40):
What I'd love to see is the multiple agents and
the orchestration of work, and then the human and agent
collaboration going on, and those journeys as they spread across
the enterprise. It's just going to go deeper and wider.
I think the numbers that we're seeing early now are fascinating,
just fascinating. So what is it going to be in
twelve months. I think we're going to be in for
some really great surprises.
Speaker 1 (32:01):
Well, let's revisit then and see where we're at.
Speaker 2 (32:04):
It has to come back.
Speaker 3 (32:05):
This is my fourth one on the bounce, and I think,
far away the best one I've been giant leaps every
time it's in.
Speaker 2 (32:12):
Yeah.
Speaker 3 (32:12):
Yeah, also really satisfying to see our customers going so well,
like it's just.
Speaker 2 (32:16):
A real us kiwis. It's a really proud moment.
Speaker 1 (32:19):
Right, Okay, Hey, well thanks for coming back on the
Business of Tech and we'll see any years time.
Speaker 2 (32:24):
Yeah, thanks Peter, thank you.
Speaker 1 (32:33):
So my head was spending after a week immersed in
the agentic enterprise, I definitely felt I got a lot
more of a handle on the tangible things AI agents
can now achieve. It's plenty of hype at Dreamforce. It's
basically a sales organization at height, but this agentic wave
is coming for businesses Salesforce, customer or not. Every organization
(32:55):
needs to figure out what that means for them. And
when we do have AI centric companies that are automating
more and more of the work many people do, it
is going to have serious consequences for the workforce, which
I don't think we fully appreciate yet. There's only so
much augmenting of tasks that will be done before agents
really do eat into certain roles. It's just a matter
(33:18):
of when that happens and how quickly. So I thought
i'd finish off the podcast with this particularly insightful comment
from Pete butodhage Edge, who rounded out Dreamforce with a
really nice discussion about AI, about keeping our purpose in
mind when we develop AI, what are we really doing
all of this activity for and who does it serve?
Speaker 6 (33:39):
Anyway, here's Pete footage Edge. I imagine a machine that
could do more and more things picture like the replicator
from Star Trek, like you can basically make everything, and
the more sophisticated the machine gets, the fewer people you
need to work in until one day all you really
need is just one guy to push a button and
the machine can do everything else. And where that leads
(34:03):
you is the wages that that worker can command, according
to economic theory, fall towards zero. And all that actually
matters is who owns the machine. If we come out
of the thought experiment, what is the machine? It's the
intellectual property, it's the software, it's the physical plant, the compute,
all of those things that add up into these firms
(34:24):
right that are doing this.
Speaker 7 (34:25):
And I think there is a question of how some
of that ownership gets shared in a broader way, some
kind of dividend that might come to the American people,
for example, which I think is fair game, given that we,
the American taxpayer, kind of sort of invented the Internet
and that these models trained on our data. Right, So
there's an economic conversation that we need to have that
(34:47):
respects the dignity of the role that humans played in
creating this technology.
Speaker 1 (34:53):
That's it for the Business of Tech Powered by two Greece.
Thanks so much to Hamish Miles for having me at
Dreamforce again and coming back on the podcast. Thanks to
Jason Parris from one in Z we're giving us those
insights into how one in d ed is rolling out
a Genticai to really good effect. Show notes her at Businessdesk,
dot co, dot NZED. You'll find them in the podcast section,
(35:16):
and of course you can stream the podcast from iHeartRadio
or in your favorite podcast app. Next week, got a
really nice episode for you. Sir Peter Beck comes back
on the show, reflecting on twenty years of rocket Lab. Yes,
the anniversary is just around the corner, twenty years of
one of our most famous successful tech companies. I've just
(35:36):
written a book that has three hundred beautiful pages from
the history off rocket Lab that's going out next week
as well, and so great to have Sir Peter to
reflect on that as he prepares for the launch off Neutron.
Tune in next week for another episode of the Business
(35:56):
of Tech. I'll catch you then,
Welcome to the Business of Tech, powered by Two Degrees Business.
I'm Peter Griffin, coming to you from San Francisco, where
I've been hanging out with well around fifty thousand other
people in town for the annual dream Force conference put
on by software company Salesforce, and it's flamboyant founder Marc Benioff.
I've always liked getting to dream Force because Benioff, You've
(00:25):
got to hand it to him, puts on a great show.
He always gets big names to talk at his conference.
This year's highlights included Google CEO Sundaipachai, President Trump's aisar
and member of the PayPal mafia David Sachs, and an
unexpected one, Pete Bodhage Edge, the former mayor of South Bend, Indiana,
who ran in the Democratic presidential campaign in twenty nineteen
(00:50):
before dropping out and serving as President Biden's Transport Secretary.
Bootage Edge is still tipped as a future potential candidate
and his talk was really fascinating. I've included a clip
of it at the end of the podcast. Dream Force
was dominated once again by artificial intelligence. It's a year
since Salesforce debuted Agent Force and look Salesforce is a
(01:12):
tech company that is probably most enthusiastically embraced AI agents.
These are autonomous software programs that perceive their environment, make decisions,
and take actions to achieve goals, often with minimal human intervention.
Agents are particularly useful for automating aspects of customer service
(01:33):
and support, sales and marketing, which is why Salesforce has
gone all in on agents. It has twelve thousand Agent
Force customers after its first year. It's around eight percent
of the entire Salesforce customer base of companies, which a
number of analysts pointed out is relatively low uptake, but
(01:53):
as Benioff countered, it's a heck of a lot more
than any other tech company is managing at the moment. Still,
despite the AI boom, some would say bubble that we're in,
Salesforce has actually seen at share price retreat twenty three
percent this year. It claims it's using AI to reduce
one hundred million dollars in costs across the company each year.
(02:14):
It shared four thousand staff, many in customer support roles
by shifting to AI agents, and it told the share
market it expects to return to double digit growth out
to twenty thirty, when it expects to rack up sixty
billion dollars in revenue, up from around thirty eight billion
last year. That growth will largely come on the back
(02:35):
of businesses taking up AI agents. One thing that only
got a fleeting demo in the Dreamforce keynote, But one
of the things that Salesforce is piloting, which is really interesting,
is Agent Force Voice. It's like calling up a call
center where you might typically get a human on the
line or maybe a voice activated menu. Now you'll potentially
(02:57):
get an AI agent. You can talk to one that
can understand what you want. It might be changing your
broadband plan, returning some clothing that doesn't fit, reporting an
electricity outage, or buying a plane ticket. Given that up
to eighty percent of interactions with customer service in some
industries is still coming in via phone lines, that would
(03:19):
be huge. If it works as envisaged, that's a game changer.
More about Agent Force Voice coming up in the show.
At dream Force, I also got around a lot of
companies that are using agents to automate various processes. PepsiCo
is using agent Force to let its smaller merchants more
easily check on the status of their orders just using
(03:40):
a simple chat based agent. Lululemon, the active where company,
is offering customers the chance to use an agent to
customize a wardrobe through an app based experience and then
go on to purchase what they've been looking at. Pearson's,
the education provider, is letting students use agents to change
courses and by textbooks. I saw a really cool agentic
(04:03):
service from Chicago Medical University to give patients one chat
based interface to order prescriptions, change appointments, get their medical
records delivered to them. That would remove a heck of
a lot of hassle dealing with this sort of gap
we have between patient portals which we all sort of
have in New Zealand now, and the mirrid range of
healthcare professionals who actually deliver frontline services. It can be
(04:26):
frustrating actually going from one to the other. Good to
see some New Zealand companies one in Zed, Fisher and
Pikeland Zero among those being showcased at Dreamforce for Innovative
Uses of AI agents. So you're hearing the piece from
Jason Paris, who I caught up on a show floor
at Dreamforce where one in Zed had a display and
demo area. But first I wanted to invite back onto
(04:47):
the show Hamish Miles, the managing director of Salesforce New Zealand,
to get his take on what's transpired since Agent Force
unleashed the AI agents almost exactly a year ago. Here's
Hamish Miles from s Salesforce. Okay, so Hamish, welcome back
(05:10):
to the business of Tech.
Speaker 2 (05:11):
How are you doing.
Speaker 3 (05:11):
Yeah, awesome, It's been an epic few days. I'm absolutely
pumped and looking forward to what comes next.
Speaker 1 (05:18):
Yeah, we're on the roof of the Muscone Center in
San Francisco. Dream Force, the biggest conference in San Francisco,
the biggest tech conference, has been going all week and
interesting what you say is that the highlights obviously Agent
Force three sixty is the centerpiece of this and to me,
it seems like we're a year into this agentic ai
(05:40):
movement that Salesforce really led you back in October twenty
twenty four. It's sort of moved from being a product
to a platform. This is the systematizing of agentic ai
into everything that Salesforce does.
Speaker 3 (05:54):
For me, it's been an absolutely fascinating few days and
you can see the progression of the journey I've been
GenTech AI and you can see some of the great
outcomes our customers are starting to produce. I think Agent
four three sixty is a natural next step for us.
It makes a lot of sense to our customers. What
excites me the most is the agentic enterprise. Where are
we going to go next? And then the combination of
(06:16):
humans and AI working together things an exciting time for
the industry and we're just starting right We're just scratching
the surface.
Speaker 1 (06:22):
Twelve thousand customers, I think half of them are paying
customers are using agent Voice. Now, what's been the adoption
like in New Zealand.
Speaker 3 (06:31):
You know, I think we're at the front edge as well.
I've got some great customers doing some great things. I mean,
you've heard from Jason Paris at one and said, that's
an agentic journey, that prepay agent does multiple things all
in one journey, all in a few minutes for Shre
and Pike or on the path with field service agents
and are scaling the premium product with Premium Service zero
(06:51):
are going to do some exciting things very shortly.
Speaker 1 (06:53):
And it seems to me, like Jason said, the customer
facing ones, which typically will look like a chatbot or
maybe a voice call. So but he said, you know
that prepaid product shifting sort of agent that they've created.
If you on a mobile plan and you want to
move to another one, and this will do it all
automatically and suggest the right plan for you. But he said,
there's actually seven agents in the background of that, So
(07:14):
all the handoffs and querrying data from one ZED system,
there's actually a whole bunch of agents doing that in
the background.
Speaker 3 (07:22):
Yeah, So they were sort of looking at the first
stage of orchestration of agents in that sense. So the
first sort of a augentic agents we spun up. They
were sort of sort of single task orientated. Salesforce has
more than two hundred and two hundred and ten.
Speaker 2 (07:34):
Agents live at the moment.
Speaker 3 (07:36):
Internally, I think there are five hundred and fifty sort
of tasks that it can carry on. We're in this
age now where this one agent, like the one New
Zealand or one assystem I think it's called one assistant,
can orchestrate multiple actions. And in their case it's like
fraud check. It's a security checking for the trust layer.
It's have we got the right plan? Can I make
(07:57):
you some recommendations? And it's absolutely fascinating and we can
do so much more as well.
Speaker 1 (08:03):
Before I go any further with Hamish, I thought it'd
be good to bring in Jason Parris from One end Zed,
who I caught up with at Agent for City in
the depths of the Moscone Center where all of these
demos of agents are being shown. So he gave me
a really good rundown on what exactly one end Zed
was doing with these AI agents to help roll out
(08:23):
a new service for its prepaid customers. So tell us
about how you're using agentic.
Speaker 4 (08:29):
Enterprise, Jason, So, Peter, this is something that's just gone
into production. We've been working with Salesforce and deploying their
agent Force technology for over a.
Speaker 5 (08:39):
Year now and this is the latest agent.
Speaker 4 (08:42):
And So from a tallocommunications perspective, one of the most
challenging journeys that a customer can go through is migration.
So when you're moving from a plan that you probably
are quite happy with to a new plan that we
think you should be interested in, it's quite often a
complex and clunky process or a customer to go through.
Speaker 5 (09:01):
And it's also a bad.
Speaker 4 (09:02):
One for a talco because you can get that reconsideration moment,
and you want the journey to be so seamless and
effortless that you love the Talco even more at the
end of it. And so we've deployed an agentic agent
in our business, which is helping hundreds of thousands of
our customers move to.
Speaker 5 (09:24):
A better prepaid plan for them.
Speaker 4 (09:26):
We've started with our most difficult customers, tens of thousands
of those and four hundred percent improvement and engagement on
the agentic journey compared to the journeys that we would
traditionally have used through normal retail or call center or
even our online website presence.
Speaker 1 (09:47):
So is the customer interacting with the agent directly, yes,
by a chatbody or something.
Speaker 4 (09:53):
Yeah, So via the agent itself, right, it'll authenticate you
to know who you are. It'll know so the plan
that you're on currently, but also the next best plan
for you. It will present a carousel of options that
you can interact with, and the agent will talk to
you about the pros and cons of the plan compared
(10:14):
to what you're currently on, and then we'll ask you
to confirm and then once you're happy with.
Speaker 5 (10:18):
It, it'll tell you that you're done.
Speaker 4 (10:21):
Right and await, it'll have a record for you and
for us of what we've agreed should anything change, and
then also the agent will ask if it can help
you with some other frequently asked questions of someone who
is a mobile customer that might want to know. So
that part of it we're going to continue to build
out and approve on.
Speaker 5 (10:41):
But the upgrade path for a customer.
Speaker 4 (10:47):
Seamless, and it took us five weeks to build and
deploy it.
Speaker 1 (10:51):
Right so, and doesn't require a lot of coding or
anything like that. It's all drag and drop, build your
own templated agent building.
Speaker 4 (10:58):
Yeah, so some configuration, but it's mainly around data cleaner,
making sure the data is accurate, integration with our existing
it the Salesforce Agent Force product set are mainly taken
and deployed.
Speaker 5 (11:14):
The time to deploy it probably is build.
Speaker 4 (11:16):
It's making sure you understand what the journey is and
the integration with your own rest of your technology and
the data cleanup. Like the very first agent we built
with Salesforce took us eight hours and then two weeks
to deploy again because of the data and the integration
with another to other technology.
Speaker 1 (11:34):
Right So, those tens and thousands of problematic customers there,
ones that are may be about to churn off one
into the you're going to go somewhere else, So that
the agent is looking at all the data that sits
in Salesforce about that customer, going okay, I know what
you need. Here are some suggestions.
Speaker 5 (11:50):
Yeah, well that's the Salesforce Marketing Cloud.
Speaker 4 (11:52):
So an another tool within the ecosystem that helps us
look at our data set, query customers into different SEGM
and cohorts and put exactly that brief in and then
it will give us that subset. And so you know
that when you are trialing an agentic tool with your
most difficult customers, if it works for them, it's going
(12:13):
to work for everyone. So you always start with the
naliest ones first, normally to give you confidence that when
you do decide to scale it from tens of thousands
to hundreds of thousands to millions of customers, you've got
that clients.
Speaker 1 (12:25):
So like, if I'm looking at changing plans, I'll go
to the one in ZED website and I'll see all
the panels here. If the different are the broadband or
mobile plans that you've got, so in the near future
you're more likely to just go straight to an agent
and say this.
Speaker 2 (12:38):
Is what I need.
Speaker 1 (12:39):
What have you got for me?
Speaker 4 (12:40):
Yeah, you will, and we'll be transparent about it as well,
by the way, so it's not as if you'll think
you're talking to a human.
Speaker 5 (12:46):
You'll know that you're talking to an agentic tool.
Speaker 4 (12:50):
Increasingly, because they're doing such a good job, we think
that our customers will go thank God, because I don't
want to talk to those use of body humans. I'd
like to talk to more efficient and so that's exactly
how that will work. And then what we will do
is we'll keep the human conversations to the probably more
important ones. So if I really have a complex technical
(13:11):
issue that I need.
Speaker 5 (13:12):
Some help with, or I want to talk to someone.
Speaker 4 (13:14):
About upgrading my handset because it's going to cost me
three thousand dollars to do that, or maybe unfortunately I'm
calling because there's been someone who's l or under financial
stress or has passed away in the family, those are
the moments where you really want to have a conversation
with a person versus a tool. So we're deploying the
(13:36):
technology or the agentic technology on the areas where really
humans shouldn't be doing that work, but it's just by
default that we're having to throw a lot of people
at it. Now we can put those people onto more
important stuff and let the agents do the hard yards
for us.
Speaker 1 (13:54):
Yeah, so that the agent will complete the transaction. So
if you decide yes, I do to move on to
a different plan, that will do that automatically.
Speaker 5 (14:02):
Yeah, that's great.
Speaker 1 (14:03):
And then so everything in the back end in terms
of your record of what plan they're on, that's all
updated automatically by the agent.
Speaker 4 (14:10):
Yeah, correctly, So into our CRM Salesforce tool at the
same time. So you've got the ecosystem of the traditional
customer relationship management, which we're moving.
Speaker 5 (14:21):
On to Salesforce.
Speaker 4 (14:22):
So all of our prepaid customers, all our products, all
our plans are now on the Salesforce CRM, and then
you apply the agentic player on top of it, so the.
Speaker 5 (14:31):
Intelligence layer on top of it.
Speaker 4 (14:32):
Yeah, where you've got a whole bunch of agents performing
a whole bunch of tasks and one of them is
a plan upgrade or a customer migration. So end to ends,
the ecosystem has a record and that the.
Speaker 5 (14:45):
Customers clear about.
Speaker 4 (14:46):
The only thing is there's disco movement, dancing and music,
and that doesn't happen when you're on our website.
Speaker 1 (14:52):
Right right, Yeah, Well, interestingly, you know they have been
talking here about Agent Force Voice, which they're piloting get
a lot of coal traffic. Still people want to call
up and then you know, if it's overloaded, you might
get a callback option or something like that. What do
you think about the option of being able to talk
to an agent's and you know how hard it is
(15:14):
to get voice recognition and stuff really accurate.
Speaker 5 (15:17):
Do you see promise in that it's a game changer.
Speaker 4 (15:20):
It's probably of all the things that were discussed yesterday
that personally I'm most excited about. I genuinely believe that
voice to agentic tools is the game changer because you
think about, I don't know, an eighty year old who
might be really kind of resistant or nervous about typing
(15:40):
something into an AI which is perceived to be a bot,
Whereas actually, if I could have a conversation where there's
no latency, that it's a humanized conversation that you can
build rapport through transparency that you're talking to an agentic tool,
it's probably an easier step for needed just to talk
(16:00):
to an AI agent in the same way I would
talk to an agent and a call center compared to
being forced down a text or social media type conversation.
And I think it's going to be way more efficient,
like even the way that I use my own tools.
Speaker 5 (16:14):
Now, if I'm talking to.
Speaker 4 (16:16):
Chat GPT, I'm not texting you speak of having a
conversation with it. Yeah, And I think even that's what
ob an Ai has seen over the last three years
of when the first versions of chat GPT, it was
all people were using it like a Google search query.
Speaker 5 (16:31):
Now it's a conversation. And I think that conversation can
move from a text.
Speaker 4 (16:36):
Based one to a voice based one with lower latency.
Speaker 1 (16:40):
It's going to be a game change dealing which salesforce
say they've been able to cope with background noise, distracted callers,
QWI accent, all that sort of stuff. If we can
nail all of that.
Speaker 5 (16:51):
Yeah. And also I think talking about the ki.
Speaker 4 (16:54):
Accent, the ability for you to talk in different languages
and understand each other at the same time.
Speaker 5 (17:00):
How cool is that?
Speaker 4 (17:01):
So you think the possibilities are endless, And I think
that the voice upgrades is going to be a really
big feature.
Speaker 5 (17:09):
And it's one of the things I've been talking to.
Speaker 4 (17:11):
Salesforce about over here is how can New Zealand be
a world first in Talco for the service.
Speaker 5 (17:17):
So you look around here and you look at the brands.
Speaker 4 (17:20):
I am really proud that one New Zealand is the
only Talco here. But also the way that we get
the attention of a company like Salesforces by moving a pace.
And so it's our job to say, serves Salesforce has
got a cool feature, let's experiment.
Speaker 5 (17:36):
With it and roll it out in New Zealand.
Speaker 1 (17:38):
First, where do you see the other real areas of
promise for agents across one in z Well?
Speaker 5 (17:44):
I think actually sales is probably the next next area.
We've been focusing a lot on service.
Speaker 4 (17:52):
But again kind of sales is in the DNA of Salesforce, right,
It's in the name, and so everything from being able
to pre qualify leads before our salespeople talk to the
customer to be able to capture the information I've got
the customer wants, the size of the business, the complexity,
and then also do some sales coaching for our teams
(18:15):
next best action that they might want to take all
of those things as a tool that can sit beside
your salesperson to do a lot of the heavy lifting
before they actually need to talk to the customer, but
also sales coaching to get higher conversion as part of
that conversation.
Speaker 5 (18:32):
That's something the next big wave.
Speaker 4 (18:34):
So we've already using salesforce within marketing, we're using salesforce
within service and that great journeys and now using I
think salesforce and from a sales perspective, whether it's any segment,
consumer or small business or large enterprise. That's the next wave.
And it's proven technology, which.
Speaker 5 (18:52):
Is why we're so excited to roll without.
Speaker 1 (18:55):
Yeah, and you're seeing the return on investment already.
Speaker 5 (18:58):
Across our AI investment five times return wow. Yeah.
Speaker 4 (19:01):
And so we've been deploying AI and all the variations
of it, so robotic process automation and machine learning for
over ten years.
Speaker 5 (19:09):
We've only been deploying agentic for last year.
Speaker 4 (19:12):
Because no one even really knew about it twelve months
twelve months ago. But yeah, five times ROI on our
AI investment. So we a lot of people are saying
is the value there? We can see the value absolutely,
and we think this is just the beginning.
Speaker 5 (19:29):
We want more.
Speaker 1 (19:37):
So there are these early adopters and one inserts a
real example of that. Jason said, for every dollar he's
investing in AI, he's already getting a five dollar return
on that investment. So that's incredible. So where is that
coming from. That's obviously efficiencies and is it people. It's
people not having to do all of that work and
(19:57):
you can deploy them to somewhere else.
Speaker 3 (19:59):
That agent on twenty four to seven and when it launched,
it launched at four o'clock in the morning, and it
had nine conversations by the time the core center had
come alive and nine engagements a week later. We're talking
about four hundred percent improvement and conversion of plans, you know,
getting people off old prepaid plans onto the new plans
that they want to promote, and a four times improvement
rate on the traditional rate. I know that sounds ironic,
(20:20):
traditional journey digital journeys that they have by the web.
I think they've still got the same head count the
people doing it. So it's augmenting humans to make them
more efficient and it can be open twenty four seven,
which as a consumer, that's a great thing.
Speaker 1 (20:33):
What do we need to do in new Zealand to
address this issue that Mike Benioff raised, you know, in
his keynote on the first day, this gap between incredible
innovation that is moving at pace and this sort of
lag and adoption. The companies are just struggling a little
bit to keep up. Is it purely an education issue
or do they not have the confidence and frankly the
(20:55):
money at the moment and tight times to say we're
going to invest in this.
Speaker 2 (20:59):
So I'll park the.
Speaker 3 (21:00):
Training sort of side, because I think skills is really important.
Speaker 2 (21:03):
We need more skills in the market.
Speaker 3 (21:05):
But I think there's probably a little bit of reluctance
to make a start, and I think people think about data,
but data doesn't have to be perfect. You've probably got
three areas of data goal bronze, silver and gold and
your account information. Well that's goal. We want to wrap
that one. Now, that's true, but marketing messages, et cetera.
Data doesn't have to be perfect and it never will be.
And you've worked out on your journey. But I think
(21:26):
most importantly you can start, make a start because it's
a low risk entry. Is the agent going to be
perfect straight away? No, it won't be, But will it
learn yes, it will. Can we make corrections very quickly, yes,
I will. So start experimenting, do something internal first before
you go external, and you'll incremnially keep on improving. And
I think for New Zelling customers we need to really
(21:47):
keep moving already.
Speaker 2 (21:48):
Look with this customer, we're seeing.
Speaker 3 (21:50):
Great productivity gains, like great productivity leagain, so it's cost effective.
It's also seeing revenue increase and productivity. And I think
it's a pretty strong message to share a New Zealand.
Speaker 1 (22:01):
So you know there's Data three sixty which is data Cloud,
so that's all been agentized and all that, so it's
all there. If you want to put your data on Salesforce,
you can do all of this stuff enable it in
in a very simple way. A few of the other
sort of innovative things that I think have come out
of this week is the move into voice agent Force
voice so if you can power a voice interaction with
(22:24):
a customer with an agent's and it'd be a good experience.
Have you seen that demo.
Speaker 3 (22:30):
Or have you used We've seen it demode and we're
looking into the first few New Zealand customers to get
on and start trialing it right and that's happening shortly so,
but I won't talk about it today, but it's going
to be really impactful because we want to meet the
customer where they are in the journey, and quite often
that's voice. It's a voice conversation, could be a text,
(22:50):
and it could be a digital channel, and it could
be an agent, but more often than not, a still voice.
It's I think seventy eighty percent for some businesses that
are coming in and protectingly for telcos.
Speaker 2 (23:00):
Thin think about having a digital twin.
Speaker 3 (23:01):
If you're in the contact center, my name's Maddie, and
I create a digital twin that's on twenty four to seven.
I can have a personal relationship with you, Peter, twenty
four seven. I don't even have to be at work
and to answer some of your questions, and so you
actually feel like you will have a personalized service with
your digital assistant who has actually got a human behind it.
That I think for a lot of people will be
(23:22):
very very good experience and quite comforting to have that
consistent service because you know when you ring up, you're
a different person every time.
Speaker 2 (23:28):
Now I want to speak to Maddie.
Speaker 1 (23:30):
Yeah, it's going to be interesting to see it at scale,
particularly with the Kiwi accent.
Speaker 3 (23:34):
How Yeah, So that is a really because that's when
you get into you know, people talk about large language
models and contextual stuff, so that accent, that stuff is
going to be really important. But also picking up other
languages as well. You know in New Zealand we just
you know have TODAYO and we have all those languages
from all the people come from other parts of the world. Yeah,
(23:55):
pacifica language in Latin America, Asia.
Speaker 2 (23:58):
Yeah, it's going to be interesting.
Speaker 1 (24:00):
That is going to be really interesting. A couple other
things move into it service management. You know, that's obviously
was a bit of a gap there, but as Bennyov said,
you know from it query support, queries, organizing field staff, technicians,
all of that sort of thing that's an area Service
now is really good at. But it is going to
be possible now on the Salesforce platform.
Speaker 3 (24:21):
Too, you know, because we've always been sales service and
marketing very externally focusing. It's not that much different from
our service operations and you know, raising tickets and distributing,
so it's quite a natural adjacency for us to enter
and it's going to give customers that singular platform opportunity
if I raise a field service ticket or I raise
a ticket which turns into a field service action. You know,
(24:42):
that could be a tech person or it could be
a contractor. And as a consumer, I don't really care.
I just want that service, and I think that we
lean naturally into that, so you know, it's a great
option for our customers.
Speaker 1 (24:53):
Any use cases you saw as you walked around Agent
Force City downstairs that you thought, Wow, that's its pretty cool.
Speaker 3 (25:00):
I've enjoyed, and excuse the pun here, watching the journey
of Heathrow Airport. You know, I've been on planes and
thought I've missed a gift opportunity for family member that
might have had a birthday or something like that, a
moment in their life that I so the opportunity to
interact with their agent. Also know, I need a gift
and I need to go to Judy Free. Can you
(25:22):
have that ready for me? And I'm running late already
and having that shopping consumer experience at the airport making
life really easy. And also like once say no, I
booked publishing data, pushing data to meet your flight's on time.
It's going to be at this gate. Oh, by the way,
we've changed the gate. Just making life easy at an airport,
because airports can be stressful. Yeah, I just think that's
a wonderful story.
Speaker 1 (25:43):
Yeah, and many the what I've seen, like Chicago Medical University,
a huge health provided as four billion dollars in revenue,
and like us in New Zealand, they're using patient portals.
You can you can get all of that, but every
time you so you need to talk to your doctor
or a nurse, it's sometimes it's by a text message,
(26:03):
sometimes it's a phone call, sometimes it's an email. It's
just all scattered. Bringing that all into one interface where
you can query it through chat. They just think it's
going to cut their call volumes by seventy percent.
Speaker 3 (26:16):
I can play that foward to that digital twins. So
the person that's with you on that journey. Yeah, because
sometimes when people go through these procedures, Yeah, very stressful,
lots of information, someone coaching you through. Yes, your appointments
here in two days, you're going to be this. This
is what they're going to be talking about. This is
what was talked about in the session. Yeah, So it's
I think I said it. I think we're I'm not
even sure we're scratching the surface yet. I mean, I
(26:37):
think the opportunity is going to be huge and when
we sit here in ten years time, we'll look back
and go, Wow, what a start. Yeah.
Speaker 1 (26:43):
And you know, we touched on the capability issue and
the training issue, and I think that is a bit
of a gap, particularly in New Zealand. I mean there
is this perception it's all sort of vibe coding. You
don't need that much expertise to do it, but you
sort of do to do it properly and to put
in the right governance around it, and just have the
confidence across the business for people to go I want
(27:04):
to go in this journey. And what I've learned this week,
it's got to be from the leadership of the company
saying we are going to be in like Jason Paris
and said, we're going to be in AI centric talco
and the first one in New Zealand. To do it,
you sort of have to have that approach. What do
we need to do to get there?
Speaker 3 (27:18):
In New Zealand we have a skills gap and the
tech sector and I think some support around that would
be amazing. Lo we have all these programs online and
probably the entire industry feels the same way. So some
assistance in getting more people into the tech secer, and
we know it transforms people's lives, right, they can go
from minimum wage to double that quite quickly.
Speaker 2 (27:37):
So that's one aspect.
Speaker 3 (27:38):
And I think to your point around what do New
Zealand companies need to do, You're absolutely right. Leadership needs
to come from the top. And if you look at
any successful transformation project over the last ten to twenty
years in technology, the one core ingredient that I had
that it got right was governance, and leadership came from
the top. If you look at all the ones that
have missed and delayed, and you know there is a
(27:58):
graveyard of transformation projects that have missed their deadlines, they
have a common theme that the leadership wasn't all on board.
Speaker 2 (28:05):
I think as we move into.
Speaker 3 (28:06):
The surgetic enterprise, it'll break down the silos of how
we work. You know, relationships will become absolutely critical. What
does the agent do, which is really task and function
and high processes and orientated to what do the humans
do then, which is relationships, empathy, context and that constant
training and oversight. So governance leadership really important and.
Speaker 1 (28:29):
We just need to start I mean it's still pretty
soft in New Zealand's economically hoping we'll turn our coin
interest in what you're seeing, I mean that rich data
coming from your customers as to how they're doing, and
there seems to be an opportunity here to cut cost,
to use salesforce and agents to actually take some costs
(28:52):
out of your business, which is what they need to
do at the moment.
Speaker 3 (28:54):
Everyone's quite cost conscious, but they're also productivity oritented. So
how do you do both? And I think the opportunity
is that you know we can. You can reduce cost,
you can increase your productivity, and typically that means improving
your top line.
Speaker 2 (29:08):
I think you're right.
Speaker 3 (29:09):
I think you know there's no secret that the economy
is sort of bouncing around a wee bit at the moment.
We're hoping to see it continue to improve. But the
customers that we've got that are embracing the agentic journey
are seeing some pretty good results early in the piece,
which is encouraging. And I would love to see more
of New Zealand. I think we New Zealand can lift
if we embrace this agentic journey. Yeah, because you will
get cost out and you will get productivity up and
(29:30):
you probably get revenue up.
Speaker 1 (29:31):
It's good to see the governments just put seventy million
into literally asking for AI centric projects through the new
Advanced Technology Institute, which is the first big investment in
AI at a government levels. That so, I guess is compelling.
Speaker 2 (29:45):
Yeah, and I.
Speaker 3 (29:46):
Think we need to continue that push for training, right,
training and skills in the industry because that will help
New Zealand lift that will benefit all of our businesses,
central government, the vendor, community, partners and you know, the
community in the end.
Speaker 2 (30:01):
Yeah.
Speaker 1 (30:02):
Just finally, the other thing that struck me this week
is around Slack.
Speaker 2 (30:06):
Oh yeah, this is cool.
Speaker 1 (30:07):
Yeah, and Slack sort of being like I've used it
not in the salesforce domain, but you know, as a
lot of media companies use slack great for messaging, but
it's become more than that, hasn't. It's become the front
end really of sales. That's quite what it was originally
found for. But right now it's the central console for
people to go and get their work.
Speaker 2 (30:27):
Now.
Speaker 3 (30:28):
I made a quote earlier two before, you know, two
hundred and thirteen agents, five hundred and fifty tasks. I
got that through Slack because Jason was asking me a question.
He goes, how many agents you've got? I went, so
I went to slack pot, which I'm now using more
than Gemini to ask questions about internally, how many agents
we have.
Speaker 2 (30:43):
What are they doing?
Speaker 3 (30:43):
So we're surfacing all that Argentic layer through Slack. Now
that won't be just us, you know, we've got partnerships
with work Day and Slack will become that console for
where you go to get that information.
Speaker 1 (30:55):
Is New Zealand a big, big market for Slack?
Speaker 2 (30:58):
Are there a lot of customers there? Yeah, we have
some great customers.
Speaker 3 (31:00):
I think Zero is a huge user of Slack, huge
user and there've been a reference for us for a
while McCloud cranes. Right, So then the other end of
the scale in the ESMB space an organization out there
lifting and shifting parts of New Zealand around with cranes.
So it's something that can fact any price that we
have Ryman homes and you know when they change rooms around,
(31:21):
they use that Slack as their workflow when they start
updating their villages. Yeah, yeah, it's cool. So that has
come a long way. Yeah, Okay, well good luck.
Speaker 1 (31:31):
It's going to be really interesting to see in it
twelve months time, where the heck we're at Given what's
changed in the last year, it's.
Speaker 2 (31:39):
In the next twelve months.
Speaker 3 (31:40):
What I'd love to see is the multiple agents and
the orchestration of work, and then the human and agent
collaboration going on, and those journeys as they spread across
the enterprise. It's just going to go deeper and wider.
I think the numbers that we're seeing early now are fascinating,
just fascinating. So what is it going to be in
twelve months. I think we're going to be in for
some really great surprises.
Speaker 1 (32:01):
Well, let's revisit then and see where we're at.
Speaker 2 (32:04):
It has to come back.
Speaker 3 (32:05):
This is my fourth one on the bounce, and I think,
far away the best one I've been giant leaps every
time it's in.
Speaker 2 (32:12):
Yeah.
Speaker 3 (32:12):
Yeah, also really satisfying to see our customers going so well,
like it's just.
Speaker 2 (32:16):
A real us kiwis. It's a really proud moment.
Speaker 1 (32:19):
Right, Okay, Hey, well thanks for coming back on the
Business of Tech and we'll see any years time.
Speaker 2 (32:24):
Yeah, thanks Peter, thank you.
Speaker 1 (32:33):
So my head was spending after a week immersed in
the agentic enterprise, I definitely felt I got a lot
more of a handle on the tangible things AI agents
can now achieve. It's plenty of hype at Dreamforce. It's
basically a sales organization at height, but this agentic wave
is coming for businesses Salesforce, customer or not. Every organization
(32:55):
needs to figure out what that means for them. And
when we do have AI centric companies that are automating
more and more of the work many people do, it
is going to have serious consequences for the workforce, which
I don't think we fully appreciate yet. There's only so
much augmenting of tasks that will be done before agents
really do eat into certain roles. It's just a matter
(33:18):
of when that happens and how quickly. So I thought
i'd finish off the podcast with this particularly insightful comment
from Pete butodhage Edge, who rounded out Dreamforce with a
really nice discussion about AI, about keeping our purpose in
mind when we develop AI, what are we really doing
all of this activity for and who does it serve?
Speaker 6 (33:39):
Anyway, here's Pete footage Edge. I imagine a machine that
could do more and more things picture like the replicator
from Star Trek, like you can basically make everything, and
the more sophisticated the machine gets, the fewer people you
need to work in until one day all you really
need is just one guy to push a button and
the machine can do everything else. And where that leads
(34:03):
you is the wages that that worker can command, according
to economic theory, fall towards zero. And all that actually
matters is who owns the machine. If we come out
of the thought experiment, what is the machine? It's the
intellectual property, it's the software, it's the physical plant, the compute,
all of those things that add up into these firms
(34:24):
right that are doing this.
Speaker 7 (34:25):
And I think there is a question of how some
of that ownership gets shared in a broader way, some
kind of dividend that might come to the American people,
for example, which I think is fair game, given that we,
the American taxpayer, kind of sort of invented the Internet
and that these models trained on our data. Right, So
there's an economic conversation that we need to have that
(34:47):
respects the dignity of the role that humans played in
creating this technology.
Speaker 1 (34:53):
That's it for the Business of Tech Powered by two Greece.
Thanks so much to Hamish Miles for having me at
Dreamforce again and coming back on the podcast. Thanks to
Jason Parris from one in Z we're giving us those
insights into how one in d ed is rolling out
a Genticai to really good effect. Show notes her at Businessdesk,
dot co, dot NZED. You'll find them in the podcast section,
(35:16):
and of course you can stream the podcast from iHeartRadio
or in your favorite podcast app. Next week, got a
really nice episode for you. Sir Peter Beck comes back
on the show, reflecting on twenty years of rocket Lab. Yes,
the anniversary is just around the corner, twenty years of
one of our most famous successful tech companies. I've just
(35:36):
written a book that has three hundred beautiful pages from
the history off rocket Lab that's going out next week
as well, and so great to have Sir Peter to
reflect on that as he prepares for the launch off Neutron.
Tune in next week for another episode of the Business
(35:56):
of Tech. I'll catch you then,
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
CrimeLess: Hillbilly Heist
It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.