Satya Nadella sits down with Trevor to discuss Microsoft’s shift to an “AI-first” era while keeping the company’s core identity as a platform for others to build. Nadella argues AI’s real challenge is delivering benefits broadly—education, health, productivity—while addressing job disruption as task-level automation and emphasizing reskilling (3 million people in Australia). The conversation highlights Copilot Studio agents, including Trevor’s family “butler” demo, plus real-world uses like Commonwealth Bank support and Cricket Australia’s ball-by-ball stats. Trevor then reveals the behind-the-scenes process to secure the exclusive interview and the technical hurdles of building the agent on short notice.
My Exclusive chat with Microsoft CEO Satya Nadella, here on the EFTM Podcast. If you haven't seen it, please enjoy!
Even if you have already seen it, jump to the end and I'll share with you some inside baseball on how it all came about.
Oh and the last minute drama just before the interview!
"In this case, you didn't sort of say, well, you know, Mike, this coordination task I have around my family, I want to basically build an agent agentic system that just does that so that I am not sending emails or text messages or keeping track mentally. And then the agent offloads all of that, right?"
Think of an agentic system like a helpful assistant that can actually do tasks for you. You tell it what you want, and it takes steps to make it happen instead of just talking about it.
An “agentic system” is an AI setup where software can take actions toward a goal, not just generate text. Instead of you manually coordinating tasks (like emails, texts, or reminders), the agent handles the workflow and updates you as needed.
"And then the agent offloads all of that, right? That ability. And that's what CBA is doing, right? For example, it's one, the customer directly is interacting with the bot they built. But more interestingly, their own agents are using agents in order to be able to do the escalation cases with higher quality, with less drudgery, and that type of improvement is what's fantastic."
Here, “agents” are like digital helpers that can do work and coordinate steps. Instead of you doing everything, they can handle parts of the process and make help requests smarter.
In this context, “agents” are AI systems that can perform tasks and coordinate with other systems or tools. The discussion highlights using agents to reduce repetitive work and improve the quality of escalations in customer support or banking.
"You also met some people from Cricket Australia who are using Microsoft Copilot to analyze stats in real time during cricket matches. You're a massive cricket fan, that must have been impressive."
Microsoft Copilot is an AI assistant from Microsoft. In this story, it helps analyze cricket data as the game happens so fans and analysts can see meaningful stats quickly.
Microsoft Copilot is a Microsoft AI assistant used to help people analyze information and generate insights. In the segment, it’s used by Cricket Australia to analyze cricket stats in real time during matches.
"and we're able to follow along ball by ball where the stats and it's just think about it, right? You know, it's not, you know, that ability to generate the most interesting stat relevant to a context of one, you know, ball."
“Ball by ball” means looking at the game one pitch at a time. Instead of only overall summaries, you can get stats that match what happened on each single delivery.
“Ball by ball” refers to tracking and analyzing each individual delivery in cricket. The segment emphasizes generating stats relevant to the context of a single ball, which is a data-driven way to make real-time insights more precise.
"And in fact, people describe this as the model overhang, right? I think the capabilities are there. The question now is to be able to harness them."
“Model overhang” is basically saying: AI is getting smarter, but people haven’t figured out how to use it well yet. The challenge is making it practical and helpful in real life.
“Model overhang” is the idea that AI model capabilities are advancing faster than organizations can effectively deploy and harness them. The speaker argues the frontier is less about raw capability and more about turning those capabilities into useful outcomes.
"I mean, I think the digital, you know, provenance and trust are going to be the most important aspects, right? It's not just about deep fakes, but everything, right?"
This is about being able to tell if something online is real or fake. As AI gets better at generating content, people will need ways to confirm where it came from and whether to trust it.
“Provenance and trust” refers to verifying where digital content comes from and whether it’s authentic. The segment connects this to concerns like deepfakes, emphasizing that trust signals will matter as AI-generated content becomes more common.
"For example, when you have lots and lots of agents, that's why I think even what we are doing with Agent 365 as a way to give organizations that core control plane to have the observability."
Agent 365 sounds like a tool or platform for managing AI assistants. The key idea is being able to see what the AI agents are doing so companies can keep things under control.
“Agent 365” is referenced as a way to give organizations a control layer for managing many AI agents. The speaker specifically mentions “observability,” meaning monitoring and understanding what the agents are doing so organizations can maintain control.
"[1478.7s] And so basically I did go into the back end and built a lot of knowledge for it. ... [1487.3s] It did come back with answers. It did come back with suggestions and it was willing to contact my wife in that way."
Think of an AI agent like a digital assistant that can do tasks for you. Instead of only talking, it can take steps—like checking your schedule and helping you avoid conflicts.
An AI agent is software that can take actions toward a goal, not just answer questions. In this segment, the agent is described as checking a calendar for conflicts, making suggestions, and even contacting someone (the wife) based on the user’s intent.
"[1505.1s] You have no idea that it failed on the morning of and we had to do a bit of turn it back, turn it off, turn it back on again. [1511.3s] And in the end, it wasn't working on my Android phone, but I had a spare phone."
When something glitches, one of the fastest fixes is to restart it. It can clear out temporary problems so the feature works again.
This describes a classic troubleshooting step: rebooting a device or service to clear a temporary failure. In tech workflows, a quick restart can restore connectivity or reset a stuck process—especially when timing is critical.
"[1505.1s] You have no idea that it failed on the morning of and we had to do a bit of turn it back, turn it off, turn it back on again. ... [1511.3s] And in the end, it wasn't working on my Android phone, but I had a spare phone."
Android is a mobile operating system used by many phone brands. Here, the speaker notes the agent didn’t work on their Android phone, so they used a spare phone to get it working before the interview.
Concept
AI
"I've been really engaged with AI using Claude to do things because producer Rob put me on to Claude. And I've frankly found it to be far more engaging than chat GPT."
AI is software that can read and work with lots of information quickly, like text from podcasts and comments. People sometimes criticize it, especially when they think it uses too much energy or isn’t made locally.
AI is being used here as a tool to process and analyze large amounts of text (podcast transcripts and comments). The segment also touches on public backlash toward AI, including concerns about energy use and where computing hardware is manufactured.
"I've been really engaged with AI using Claude to do things because producer Rob put me on to Claude. And I've frankly found it to be far more engaging than chat GPT."
Claude is an AI tool that can help with writing and understanding text. They’re using it to make their podcast-related work easier and more interesting.
Claude is an AI assistant used to generate and analyze text. In the episode, they’re using it to help with tasks and to make the podcast workflow more engaging.
"I've been really engaged with AI using Claude to do things because producer Rob put me on to Claude. And I've frankly found it to be far more engaging than chat GPT."
ChatGPT is an AI chatbot that can answer questions and help write text. They’re comparing it to Claude and saying they prefer Claude for what they’re doing.
ChatGPT is another AI chatbot used for text generation and Q&A. The host compares it to Claude, saying Claude has been more engaging for their use case.
"And so we can now ask questions of the database about the sentiment of electric car owners versus buyers, brand sentiment. There's a bunch of things we can do to analyze."
They’re comparing what current EV owners think versus what people who might buy an EV think. The goal is to see what people like, what worries them, and what influences decisions.
They’re using podcast transcripts to compare how people who already own EVs feel versus people who are considering buying one. This kind of sentiment analysis helps explain what’s driving adoption (or resistance) beyond just sales numbers.
"And so we can now ask questions of the database about the sentiment of electric car owners versus buyers, brand sentiment. There's a bunch of things we can do to analyze."
Sentiment analysis means the computer reads lots of comments and figures out whether people sound happy, angry, or unsure. It’s a way to summarize opinions without reading everything manually.
Sentiment analysis is using AI/NLP to classify opinions in text—like whether comments are positive, negative, or neutral. In this case, it’s applied to podcast and video comments to quantify attitudes toward AI and EV-related topics.
"I understand there's a lot of negative press out there about data centers using power, data centers using energy and all this kind of rubbish."
Data centers are big buildings full of computers. The concern is that they use a lot of electricity, which can affect energy use and emissions.
They’re referring to the energy demand of data centers that run AI services and store/compute data. This is a common part of the debate around AI’s environmental footprint and energy use.
"the challenge with the broad success and potential of AI and the problem is actually capitalism and the stock market because there's such a requirement on companies to have more revenue,"
An AI strategy is a plan for how a company will use AI across products, operations, or customer service. The transcript frames it as requiring major investment and then producing measurable outcomes like cost savings and efficiency.
"That the intent of, let's say a big company implemented an AI strategy with a huge investment and found cost savings in it and decided to give their staff four day working weeks on the basis that they implemented AI into their work days to therefore make themselves"
This is the concept of working fewer days because AI helps you get the same work done faster. It’s more about how companies schedule work than about a specific technology.
The idea is that AI can increase productivity so the same amount of work gets done in fewer days. In practice, it’s a workplace model change (hours and scheduling), not an AI feature by itself.
"So look, I am super proud of the Sachin Adela interview. I appreciate all the feedback I've had on it. And I think it was my great friend and Amazon Web Services PR person, Matthew,"
AWS is Amazon’s cloud service—basically servers and tools you can rent to run apps and AI. Many businesses use it instead of building their own computers.
Amazon Web Services (AWS) is a cloud-computing platform that many companies use to run software and AI workloads. In the segment, AWS is referenced via a PR contact, but it’s still a notable tech company tied to how AI is deployed.
"So he suggested that I needed to speak to Jensen, the CEO of Nvidia, and then the CEO of Amazon. And then we'll go to Google."
Nvidia makes the powerful computer chips that help AI “learn” and run fast. The host is saying they wanted to talk to Nvidia’s CEO.
Nvidia is a major semiconductor and AI-computing company, especially known for GPUs used in training and running AI models. The transcript mentions meeting Nvidia’s CEO as part of a list of big tech leaders.
"Bigger problem is we've got a new CEO at Apple later this year. So do I have to retick? I don't know. What's the story there?"
Apple is the company behind iPhones and other devices. In this segment, they’re talking about a CEO change, not car details.
Apple is referenced in the context of leadership changes, with the host noting a new CEO coming later in the year. While not automotive-specific here, Apple is a major consumer tech company that increasingly intersects with connected-car and in-vehicle ecosystems.
Select text to request an explanation
I was gobsmacked. I spoke to you on the Tuesday, Thursday afternoon, there's this lovely courier
man at me back door with a parcel for me.
All I know is, we've been 24 hours mate, you had it sorted after five weeks of hurl, raising hurl.
Mate, I can't thank you enough.
Join the conversation, head to eftm.com and click Ask Trev.
Hoping Australians would check questions for over 15 years, the EFT and podcast would travel on.
Real Australians, real questions, every week.
You can text Trev now, thanks to Vodafone on 0477 657 657.
Well, don't adjust your sets, do not check your calendar, and to produce a Rob, relax, this is not an accident.
I'm releasing a special edition of the EFT and podcast at the end of your week.
There'll be a new episode on Tuesday again with all your next tech questions and conversations.
But look, I had the opportunity this week to speak to Sachin Adela, the CEO of Microsoft.
I'm in China right now and I'm reflecting on that opportunity and where I placed that interview.
And I thought to myself, you know what, if you watched on the Today Show, you saw three minutes of it.
If you saw it on eftm, you might have glanced at the article and watched a little bit of the YouTube.
I just wonder how many people actually stopped and watched the whole 14 odd minutes of it.
So I thought to myself, for those that haven't, so let me be clear, this show is the Sachin Adela interview.
And then at the end, I might just talk about the process because I think it's quite interesting to people,
a little bit inside baseball perhaps, but I think you might be interested in it.
So if you've already watched the entire interview, the whole 14 minutes, that's what this is.
But maybe skip to the end and you'll hear a bit of the conversation about what actually happened and how it came about
because I think it's a little fascinating.
Excuse me, but yes, this is my opportunity to share with you a very special conversation,
especially because it's exclusive.
Only myself and Jared Lynch from the Australian, and I get on really well with Jared.
We were together the day of his interview with Sacha.
We're the only people in the Australian media to talk to Sachin Adela while he visited Australia.
And there's many reasons for that. I'm sure one of them is time.
The man only has so much time to give, but also you've just got to pick your battles when you're Microsoft.
I'll talk about that at the end of the show.
But a very unique opportunity to sit down and record this conversation.
This is the only recorded conversation with Sachin Adela.
The CEO of Microsoft in Australia.
We conducted it 10, 15 minutes after his keynote presentation at the Microsoft AI tour at the Tiktok Entertainment Center.
Once he was done, I'm sure he freshened up and came straight into a room and said good day.
Hi, Trevor. How are you? It's great to see you.
Great to meet you. Have a seat.
Welcome to Australia.
Thank you so much. It's been fun.
We've got a lot to talk about. You've made a lot of announcements. I want to talk about AI where it's at.
I've got some fun things to show you.
But just want to start by thinking about Microsoft at this position in time.
I was thinking about it during the week.
We went from Command Prompt MS-DOS to Windows 3.11.
This feels like a similar transition from being a software operating system company to being AI first.
Is that a safe assessment?
I always think about Microsoft's core identity as a platform and a partner company.
I looked back and the anecdote I was sharing at the keynote is that the first time I was in Sydney doing some demos
was trying to talk about visual basic applications inside of Excel that others can use to build applications.
That is at the core.
Even though it is a shift, a massive shift, very different.
It may feel like the internet or, as you said, the move from DOS to Windows and maybe bigger even, quite frankly.
But the core thread line for me is how have we enabled others to build more ambitious things than anything that we have built?
That, I think, is the defining line of Microsoft.
It was interesting listening to your keynote earlier today.
There was people behind me literally saying, that is cool.
Those kind of things under their breath.
That is a crowd of people who work in software and they are developing tools.
For the average person in Australia, let alone around the world, is AI something that they should be worried about or concerned about?
I would say the most important thing for us collectively, both in the tech industry but in the broader society,
is to make sure that the benefits of AI are really delivered to everybody.
Because if we don't, then, of course, it is not going to have the social permission.
What does that mean?
It means you want to be really good users of water and energy to create AI
that then is helping with education outcomes, health outcomes, public sector is becoming more efficient.
The companies that are using AI are becoming more productive.
They are driving economic growth that then is making its way to people's lives.
Because that is the way permission is earned.
I feel, at this point, the narrative around AI is such, which I am glad,
where we are talking about both the benefits and the unintended consequences versus waiting.
And that debate is a good debate to have versus just indexing on any one of them.
Because just to say AI is all about the fear of AI, that won't help because it is inevitable that this is a secular change.
The question is how do we drive that change such that the benefits really are getting through to everybody?
It feels like a secular change and also a societal change.
Which also brings up one of the other concerns, I think, in society, which is jobs.
One of the most common questions I get when interviewed about it is, you know,
but it is going to take jobs or cost jobs.
How do you answer that to people in the general public about where jobs are going to go with AI?
Yeah, I mean, I think the one thing is, you know, to start with, I think recognizing even right now the state of AI
and quite frankly even for the foreseeable future is more about what I will call task level automation inside of jobs.
Even yesterday I was talking to some local Australian startup doing a fantastic job of building a scribe tool for doctors.
And one of the things he was describing to me was how much it is benefiting the physicians
by reducing their paperwork burden and helping them spend more time with patients.
So that's the type of task that they did that quote-unquote AI does so that they can do the real job, which is take care of patients.
And that's happening even in office work, right, when I think about even my own job, right?
You know, was I meant to mostly triage email or can I do things where I can focus my attention
on being able to really get the deep insights, pass better judgment on things that I need to do so and so on.
So I think that that's one aspect of it. But at the same time, I'm clear-eyed that that's why I think even the skills announcement we had,
which is about how do we really make sure that as even new job opportunities get created, maybe in different forms,
we have the skills required to be able to take those jobs.
And you've made a big commitment here in Australia.
Absolutely, that's right. We have now committed to training more than 3 million people in Australia,
along with obviously all the other investments we're making around the data centers.
So big investment in skills for Australians, and that's coming from Microsoft.
Is there a reason why Australia gets that investment? Is that an endorsement of our people?
Australia is a very important market for us.
And whenever I think about it, again, I go back to, you know, as a multinational company,
how are we going to have a license to operate in Australia?
It's primarily by making sure that we invest in this country, not only our own people, but also our data centers.
And we're doing things like skilling and things that we're doing in cybersecurity with the signals directorate.
These are all important investments.
Ultimately, to create local surplus, right, when that start up, when, you know,
Australian Post can use this in order to be able to, you know, have accessibility be core to them
and their co-pilot adoption or CBA uses it for their customer applications or what have you.
And these are the things ultimately that will drive our business.
And did you get a sense you obviously met with the Prime Minister to make some of those announcements?
Is the Australian government ready for AI? Are you helping them along the way or are they coming to you for help?
First of all, I think in Australia, I would say the government has taken really a progressive approach
to sort of making sure that they're thinking about what is the positives of AI,
what are some of the things that they need to get ahead in terms of whether it's safety or whether it's around skills
and what have you, and it's fantastic to see their adoption and thoughtfulness when it comes to AI.
And also the human capital in Australia, right, which is, as I said, is whether it's in these large scale multinational companies
that are operating or the startups that are getting born here and not only building for Australia but for the world,
the high ambition here. This is the first time I'd say in the world, and you see it in Australia,
that the rate of diffusion is such that there's no sort of, I would say, time lag between what may be happening
on the west coast of the United States or the east coast of China and what's happening in Australia.
Now, I wanted to take a minute. You spent some time yesterday with a lot of Aussie businesses looking at how they're using AI.
And I think the average Aussie, let alone average person in the world, thinks of AI as asking a question,
getting a more thoughtful answer. It's very basic from a couple of years ago.
But there's a thing called agents. You know all this, but you had the Commonwealth Bank talking about how they're using an agent in a live chat.
So what I did, I took Copilot Studio and I tried to create an agent for my family.
So I've got three kids, my wife and I, and I thought about my busy life, my busy diary, and we have a shared family calendar.
So I wanted to demonstrate it to you if I could.
Cool.
How could you build it?
Well, I use Copilot Studio and I integrated the Outlook calendars.
And so we've called it Charles, our family butler. So I'll give it a quick example.
What's my schedule look like on Monday?
And so basically it's now going to go and check my personal schedule and see how busy my day is.
But it's also because I've engaged it to always check the family calendar because I want to know what I'm going to be able to do and how I'm going to help the family.
So it's looking at my calendar.
I think people need to understand that it's not just reading a calendar.
It's trying to understand the schematics of the day.
And so now it's given me a whole output of my day telling me what I've got on.
But by default, it's telling me about the clashes and problems that I might have.
So considerations and even making suggestions.
Would you like me to send Amanda, my wife, a quick note about the netball pickup clash?
But here it gets even better.
So what I can also do is say, move my planning meeting to earlier in the afternoon and let my wife know that I can pick Victoria up for netball.
So basically now it's going to change my calendar and it's going to send my wife an email letting her know that I've now changed my calendar.
I can pick up Victoria from netball.
So I know it's a really simple example.
No, it's a fantastic demo.
I wish I'd gotten you into my keynote and demo.
Because I think what you just showed is like the pride, first of all, the pride you have, the sense of empowerment you have that you built this application.
Just like in the past, you may have built an Excel spreadsheet or a Word document.
You just literally created a piece of software.
And then there's an agent.
There's an email to my wife, which to be honest, probably not appropriate.
I would normally send a text maybe, but just letting you know that I've adjusted my meeting so I can pick Victoria up from school on Monday and take it to netball.
So it's a great, I feel like is that the next phase for AI to see real world usage of it?
Correct. And I think you captured it well, which is what is the output that you care about?
In this case, you didn't sort of say, well, you know, Mike, this coordination task I have around my family, I want to basically build an agent agentic system that just does that so that I am not sending emails or text messages or keeping track mentally.
And then the agent offloads all of that, right?
That ability.
And that's what CBA is doing, right?
For example, it's one, the customer directly is interacting with the bot they built.
But more interestingly, their own agents are using agents in order to be able to do the escalation cases with higher quality, with less drudgery, and that type of improvement is what's fantastic.
And critically, Conbank also talked about how it then enabled the human staff to have more, I guess, insightful information by the time they got to the staff for support.
Exactly. And I think that, you know, the complexity of some of our products and others inside of an organization are such that sometimes when even you're in the front lines and customer service or as a banker even, you want to be able to have this agentic sort of support that's right at your fingertips, right?
In fact, that is the way, you know, we used to have this thing called information at your fingertips and now going forward, it's about agents at your fingertips that are really helping you be more productive.
You also met some people from Cricket Australia who are using Microsoft Copilot to analyze stats in real time during cricket matches. You're a massive cricket fan, that must have been impressive.
You've got a cricket pitch at Microsoft, head off to Seattle.
We do, we do, and it's well used. In fact, I get to overlook it. But yeah, I mean, what Cricket Australia has done is fantastic.
I mean, first of all, cricket is a very sort of stats rich sport and they've been able to take what's been a treasure trove of data that they have had since 1897 or something like that and parlay that into now this Cricket Australia app and they have the stats guru functionality
and we're able to follow along ball by ball where the stats and it's just think about it, right? You know, it's not, you know, that ability to generate the most interesting stat relevant to a context of one, you know, ball.
That's just extraordinary.
If we rewind five years, I don't think anyone in the general public saw AI coming. You maybe did and other people in business, but I don't know what the next three to four years looks like. Do you have a crystal ball of how much more is going to change for average Australians around the use of AI or what's next?
Yeah, I think the speed with which some of these new capability jumps are happening are pretty stunning and that to your point was what, you know, when I first read even the Scaling Laws paper, I felt like, wow, if this happens, this would be really interesting.
And it's turned out that it's held out line. But the more important aspect I think we're reaching now is the models. And in fact, people describe this as the model overhang, right? I think the capabilities are there.
The question now is to be able to harness them. Like, that's why I'm more impressed with your demo than any model capability, right? Because at the end of the day, did we all use it, do things that are useful for outcomes we care about individually and as organizations throughout society?
And that I think is more going to be the defining thing about what's the real frontier of this technology.
How quickly, finally, how will people know what's real and what's not in the future? I feel like that's a real concern for people, especially with what's happening on the Internet.
100%. I mean, I think the digital, you know, the provenance and trust are going to be the most important aspects, right? It's not just about deep fakes, but everything, right?
For example, when you have lots and lots of agents, that's why I think even what we are doing with Agent 365 as a way to give organizations that core control plane to have the observability.
And this is the issue with any new technology, right? As the rapid adoption comes on the unintended consequences. And this time around, we have to really get ahead, right?
Whether it is about identity, provenance, security and governance policies and so that way that we can be secure about it.
Can Australia be an AI powerhouse?
I mean, I think that we think about AI is more an accelerant to the comparative advantage of countries and companies that already have that going, right?
So in some sense, AI is not some abstract new thing that sits outside of what is the rest of the economy, right?
Australia today is a powerhouse and many fronts and those fronts will get more amplified because of their use of AI.
And of course, when it comes to even just generating the startups that go on to capture large swaths of market share in different categories can come out of Australia as much as any other place.
I've watched you bounce from place to place over the last two days. You're a very busy man. You're the boss of one of the biggest companies in the world.
Do you get a chance to just look around Sydney at any point while you're here?
That is one of my dreams is to just come here when I'm looking around. But look, it's always a pleasure to be able to come and visit and quite frankly meet with people.
It's energizing to me to be able to see the demo that you did or the demos from the Commonwealth Bank or the Cricket Australia and it's always fun.
Thanks for your time.
Thanks so much, Trevor.
Cheers.
Thank you.
Cheers.
This is the EFTM podcast.
So look, it is a fascinating thing that you probably wouldn't have any insight in and I understand that.
But just from an inside baseball perspective, that was 15 minutes of my life that took three months and that is absolutely no lie.
I think I had a conversation with Microsoft earlier in the year and you reflect back on these lunches or coffee catch ups or phone calls you have with these companies.
Like I reflect back on a coffee slash pineapple juice conversation I had at the Hilton in Sydney with Fiona Martin at Apple in 2008 maybe.
And we would then catch up regularly at that same coffee shop which was in the kind of lobby area of the building where Apple at that time was head office in Australia.
But I look at those conversations and I realize that a lot of it was just lovely friendly conversation catching up about what was happening but also at the same time there's a little level of vetting going on.
And I don't think there's any harm in that at all.
These companies are trillion dollar companies that can't just let any old Joe Blow get access to their executives in conversation.
So in the Microsoft sense, I reflect on and I wonder whether the conversations we had led to this.
But essentially it was before the Formula One Grand Prix that, so that's early March, that I ended up having a conversation with Microsoft's PR team about the idea that we might get the chance to speak to such a person.
So obviously at that time they knew he was coming to Australia.
Maybe that wasn't public yet but they knew he was coming here for this AI event.
And this is not a man that does a lot of interviews.
And so the conversation was along the lines of how could we do it differently?
What would you do to pitch yourself to such an Adela?
And so first and foremost I literally had to send a little rundown of who I am and what I do and that would probably be sent to such an Adela's personal PR team.
And they would look at the person and the profile and some of my work and he seems like the right person.
But then it was what is this conversation?
Because I got the sense they didn't just want to sit down and have a chat.
So we talked, myself and the Microsoft PR team, Dan did an amazing job guiding me with this.
But the idea was how do we engage such an Adela in a conversation about AI, at an event about AI, in a way that would stand out from a regular interview.
And that's when we had the idea of building an AI agent.
As you heard me, I showcased this to him.
Now I've got to tell you, I don't know how to build an AI agent.
But Microsoft set me up with an account on their co-pilot show.
I actually went to co-pilot and I just typed, I want to do this.
And the idea very early on was I want it to be family-based.
Because for me it was one of two things.
It was either self-employed small businessmen trying to make life easier.
And as you've heard me talk on this show and with Stephen a lot about the troubles or the challenges of being a self-employed and a small business.
Just time.
I just didn't think that conversation would resonate with a today's show audience, for example.
Because obviously these conversations, when you're having them with a brand like Microsoft about getting an exclusive interview like this.
If I said, oh great, I'll put him on the FGM podcast, they'd be like, that's great, but what else?
So I'm sure they'll be stoked that I'm putting this chat on the podcast.
But let's be honest, television was where we wanted that to land.
And so that's where we landed on the idea of something for the family.
Now I had grand plans.
This was going to be like an app that the kids could have and it would manage their chores and give them points.
And it would nag them if they hadn't successfully completed tasks.
And it was like the concept was amazing.
And I asked co-pilot how to build this and it just basically said, you've got to go do this.
And it said I needed co-pilot studio.
So Microsoft gave me co-pilot studio, which is a much more advanced enterprise level thing.
And I started with wanting to just link my calendars.
We have a Google calendar.
I have a Google calendar.
Off we go.
And it was all working, but not perfectly.
So there was a lot of, I don't know how many, but I had maybe three or four team schools,
so video conference schools with a couple of product experts at Microsoft.
And a massive, massive shout out to Jack and Eliza, who was so patient with me.
I'm not an idiot when it comes to tech, but I'm certainly not at their level.
So they were super patient.
And it's funny because I said to Microsoft afterwards, I said, you know what,
when you set up that first teams core where I would work with Jack and Eliza,
I just figured they'd take over and build it for me.
But they never did.
They never did.
They never remotely controlled my computer.
The only thing would be to screen share so that they could see what I was doing
and they could talk me through different things.
But in the end, most of the time, I genuinely and proudly would come to them
into the next meeting with a solution that I'd created or founded.
So one of the solutions was, you know what, I'm just going to export my calendar out of Google
and put it in Outlook because obviously staying in the Microsoft environment is going to be easier.
So we did that.
And we had a lot of little challenges.
I did learn how to make like a web app so that you could do fun things like chores,
but it was becoming too close to time.
And I was time poor as I regularly am.
I didn't have a week to just dedicate to this.
I was dedicating, you know, two hours on a Monday and one hour on a Thursday
to try and build this thing.
And so we kind of got to the point of, hey guys, minimal viable product, magic of television.
So look, hand on heart, my family have never seen that chat agent.
No, we haven't implemented it in the family.
I didn't get to film the kids using this AI agent or anything like that.
But what I did genuinely do myself was build an agent that could see my calendar and my family calendar.
And what's really interesting is behind the scenes, you build an agent with skills or intent.
And so what's fascinating is I could say to it, what's my diary like a Monday?
It would say, here's your diary.
But then I'd have to say, well, does anything conflict with the long family calendar?
And so that's a second question.
But then I went, hang on a minute, I'm always asking that same question.
So why don't I build into the skill, into this agent's brain that whenever I ask about my calendar,
hey, at the same time, check the family calendar and let me know about conflicts and make suggestions.
And then I came up with the idea that I would, you know, okay, send an email to this email address my wife.
And then I went, well, hang on a minute, tell it, teach it your wife's address.
And so basically I did go into the back end and built a lot of knowledge for it.
So that when I asked those questions, it did come back with answers.
It did come back with suggestions and it was willing to contact my wife in that way.
It's amazing.
And I could genuinely build that out, excuse me, build that out far more.
And it could be really interesting and exciting.
So I was super relieved that it worked.
You have no idea that it failed on the morning of and we had to do a bit of turn it back, turn it off, turn it back on again.
And in the end, it wasn't working on my Android phone, but I had a spare phone.
So I brought that out and I used it on that instead.
So little things that go wrong hours before a, you know, very, very big interview and we got it working.
And if you, if you've only listened to this, if this podcast is your only exposure to that interview,
I'd love you to go to the AFTM YouTube channel and look up that interview because when we were having a conversation about AI and jobs and yada, yada, yada, yada,
Sachin Adela was sitting back and relaxed and chatting.
Fine.
But when I brought up the agent, he lent forward to the point where my cameraman had to move because he blocked the shot of me.
The cameraman had to move to reposition because Sachin Adela was lent in and engaged fully in this concept of an agent.
It was so good.
And he was genuinely into it.
And to the point where he mentioned it several times afterwards, it was brilliant.
And look, I'm not saying for a second that interviewing Sachin Adela was better or worse or different to interviewing Tim Cook.
It was just a different set of circumstances.
Interviewing Tim Cook from Apple was purposeful in the sense that it was about Apple Vision Pro launching in Australia.
So that was what I was talking about.
And it felt no different.
He was, he was not any less lovely, personable and welcoming of me.
The only difference was I didn't get a photo with Tim Cook.
That was not part of the plan.
Whereas Microsoft had a photographer ready to snap photos.
So it was a brilliant experience for me to learn about these things, but also then for me to hear from Sachin Adela that what I did was quite purposeful in the sense that I built something.
And you would have heard me talking with Steven maybe on TubeLex or Intek if you've been listening over the last few weeks.
I've been really engaged with AI using Claude to do things because producer Rob put me on to Claude.
And I've frankly found it to be far more engaging than chat GPT.
And we've now built a database of every word I've ever spoken in a podcast.
Like I'm now thinking, okay, I'm going to download all my old 2GB radio segments that Chipped for Brains show with Brian Wilsch here.
And put all that in the cloud as well because it's basically everything I've ever said.
And it's not a directly useful thing today, but I wonder what it will do for me.
And what I found for example with the EV Podcast is I've been able to ingest all of that.
And so we can now ask questions of the database about the sentiment of electric car owners versus buyers, brand sentiment.
There's a bunch of things we can do to analyze.
And it's fascinating reading the comments of the, especially the Nine News video of the Sachin Adela interview because there's so much hate for AI.
And I understand there's a lot of negative press out there about data centers using power, data centers using energy and all this kind of rubbish.
And also about like the $25 billion investment being not actually in Australia.
It's really, you know, $20 billion that is just going to Taiwan and other places where they make the semiconductors and computers and it's coming back to Australia.
Fine, whatever. That doesn't really, I don't struggle with that morally.
And I do know that jobs are going to change. Some jobs will go.
But I also think that there will be jobs created and the balance will be there.
I said to Stig Roblock, my best mate and cameraman who follows me around the world and he's here with me in China.
I said something like, and this is very deep, but the challenge with the broad success and potential of AI and the problem is actually capitalism and the stock market because there's such a requirement on companies to have more revenue,
to have more users make more money, deliver more profits to shareholders, right?
That the intent of, let's say a big company implemented an AI strategy with a huge investment and found cost savings in it and decided to give their staff four day working weeks on the basis that they implemented AI into their work days to therefore make themselves
more efficient but deliver the same outputs. Imagine how society could change and be better with AI.
So I look at my own, and as a person who does control their own destiny, I'm still at the point where I go, if I could make a day back on myself, I'd want to make more money out of it, right?
But if you're a small business that is grinding away behind a computer five days a week and you work out that you could do four and a half or four and get the same work done because you implement some AI, your life's going to be better.
And I don't think we're talking about that enough. So no, coal mining and Woolworth's work is a different story. But think about the number of small businesses there are out there and the efficiency that could be gained.
That's super exciting. So look, I am super proud of the Sachin Adela interview. I appreciate all the feedback I've had on it. And I think it was my great friend and Amazon Web Services PR person, Matthew,
Wuthry Char, as I know him, sent me a text with the biggest companies based on market capitalization, and they are Apple, Microsoft, Nvidia, Amazon and Google.
So he suggested that I needed to speak to Jensen, the CEO of Nvidia, and then the CEO of Amazon. And then we'll go to Google.
I'm ticking off the CEO list. Bigger problem is we've got a new CEO at Apple later this year. So do I have to retick? I don't know. What's the story there?
It's like pay phone tag. You've got to retag it if someone else grabs you. Grab your phone booth, by the way, hyper fun, and still enjoying it. So thank you to everyone who sent me feedback on that as well.
Anyway, thank you for listening. I hope you enjoyed my conversation with Sachin Adela, as well as a little unpacking of it here on the EFTM podcast.
Request an explanation for:
Request an Explanation
Heard something you'd like explained? We'll add it to this episode.
Sign in to request explanations for terms you heard.
Want to learn more?
Browse our glossary for plain-English explanations of automotive terms, jargon, and concepts.
See something that's not quite right? Our annotations are AI-generated and can sometimes miss the mark.
Click the flag icon on any annotation to suggest a correction.
