Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Mind the Model: Episode 008: Adam Goodman
34 Plays19 days ago
Mind the Model is back, after our Australian Summer break.
This week we’re talking to Adam Goodman. Adam is the Director of AI in Advertising APAC for Microsoft Advertising, where he is responsible for helping launch new and future advertising experiences.
We caught up with Adam right before Christmas and had a wide ranging discussion.
If you’re after his prompting secret sauce it’s: Context, Outcome, Tone
Get in touch with Adam here:
https://www.linkedin.com/in/angoodman/
More info about Microsoft AI:
Recommended
Transcript
Final Recording of the Year
00:00:12
Speaker
Hello and welcome to Mind the Model. I'm Nathan Garrett and as always, I'm joined by my co-host Emily Crellin. Em, how's it going? It's good. How are you doing, Nate? Good. Kind of winding down for the end of the year. So I think this is our last recording, isn't it? This is our last recording of the year. So are we goingnna are we taking a little January so break? i think so. I think we should give ourselves a break and um kind of take some rest and ah stock up for next year. And yeah it should be good time. We have lots of um interesting guests queued up and a very special one today.
AI in Holiday Shopping
00:00:41
Speaker
and But before we get into that, Nate, I wanted to ask, have you bought all of your Christmas presents? I mean, like all is a strong word. I think I'm pretty close though. I mean, i might go out just in a frantic mad rush at the end and there maybe a few more stalking stuffers than I need. But yeah, for the most part, I think I'm there. I'm there. Nate, my tree was up ah early November.
00:01:05
Speaker
All of the ah presents have been wrapped underneath my tree. wow a free thanksgiving tree a pretty thanksgiving tree i know we were very eager beaver um this year but um i actually use ai for some of my christmas shopping believe it or not you were using um an agent to go out and buy things for you or what what would not that far i'm still i'm still a bit reticent about agentic shopping i know we've spoken about on the pod but it's the future it is apparently it's the future um No, so I, ah hopefully I don't spoil this if my in-laws happen to listen to the episode. um
00:01:44
Speaker
We do see, like, not a secret standout, what's it, White Elephant, like the Kris Kringle? Yeah, yeah. Yeah, so we do that um every year, and this year the limit $70. Those chunky gifts.
00:01:56
Speaker
Yeah. Chunky gift. Cause it's like, we don't want to get a gift for like each of the siblings and. Gotcha. So this is you guys being cheap. In other words, this is us just being cost conscious.
00:02:08
Speaker
So, um, I asked chat, Chippy T my favorite chatty Cathy. Yeah. i asked for a, a gift recommendation. that was aligned to a couple different categories and under that price range. And it came back with a few things. I'm like, o I'm not sure if this is what I want.
00:02:28
Speaker
yeah And then it recommended this really boutique design, homewares, cool, cool, cool store in Melbourne.
00:02:39
Speaker
okay And as I was like, oh, that's pretty cool. And i got the last, item that they had for sale i ended up submitting an inquiry through their website because i was like it says that it's for sale but it's not letting me buy it i'm emailing them back and forth yep cool i got it and i would show you but guess what it's already wrapped and it's at my in-laws house already because i'm so forward thinking yeah um it is the coolest like vase vase but it's a 70s ceramic so it's like pink and brown and cream Okay.
00:03:14
Speaker
It is my aim of the game to make sure it comes back home with me. Yeah. Because that's how I approach white elephant. I get a good gift and i want to bring it home with me. Yeah. Yeah. um And so my husband got a really nice candle. um I mean, I told him what to buy. was like, because if we get that one too, that one's great. um But I think my mother-in-law is in on the game too because she's like, what I bought, I want to take.
00:03:36
Speaker
So at the end of the day, it might just be that we're all going home with what we ended up bringing. That's fine. As long as, as long as the the thing is, as long as nobody buys anything really bad, because then you get stuck with it. I'd like, yeah. But with $70, you would, you would think that people would put some effort behind it. You would think. I don't know your family.
00:03:56
Speaker
And I don't know how, like my family sometimes will just be like, okay, so $70, we're going to just do something ridiculous. Like $70 worth of ah flour. No point at all.
Introducing Adam Goodman
00:04:08
Speaker
Well, ah speaking of vases and vases, I think that's a perfect segue into our ah guest today.
00:04:15
Speaker
I don't know how you're going to spin this one. Well, so Goodman is the director of AI and advertising for APAC at Microsoft. He spent 20 years in the industry with Microsoft, Adobe and working in agency size.
00:04:30
Speaker
And he's responsible now for helping launch the new and future advertising experiences around ai in particular. But he's also English. And so it's a perfect person to answer the question, is it a vase or a vase? Because there are really very different you know points of view on how you approach that. So Adam, welcome to Mind the Model.
00:04:50
Speaker
Thank you very much. So first question, is it vase? Is it a vase? Where do you stand on that? It's difficult question. It's definitely a vase. Good to know. I'll keep it with the vase. How are you going to try to do AI in ads? that's That's going to be an interesting one.
00:05:05
Speaker
Well, I'm sorry. to To kick it off into more of an AI focus, given that that's the topic of our podcast, um Adam, we want to run you through a few of our quick fire questions. um And I'm pretty sure I know what you'll say, but what is your favorite AI model and why? And then follow up to that, what is the last thing that you used AI for? Well, I've got to say this and I'm going to say it's the MAI image one which we've just released. And how is it how is it different or how is better?
00:05:34
Speaker
Very, very photorealistic. okay um I've been playing around with it for photos for home, Christmas, things like that. Gotcha. But what have I just used it for?
00:05:45
Speaker
Very tough question because I literally use it all day, every day. um I use it in Teams and PowerPoint, you name it. But I also used it for this interview.
00:05:57
Speaker
I'm very curious how you've used it for this interview. I guess we'll wait to see. Also shopping, I would say. You mentioned about how you did that and your browsing for shopping.
00:06:08
Speaker
don't know if you've played around with Vision. Have you played? Owen, your favorite one's ChatDVT. You've already mentioned that. It is, yeah. At the end of this, hopefully going to change that.
00:06:21
Speaker
But if you haven't played around with vision, that's very cool. So it can see what you see. You can ask it questions as you kind of browse any website or buy anything online and just ask for advice. It's almost like having someone literally watching watching what you're buying, but also giving you advice of...
00:06:43
Speaker
whether it's the right fit or
Copilot and Microsoft 365 Integration
00:06:45
Speaker
whether it's the right tone or whether it's something that you might buy for your family or your wife. i'm Very, very cool. And this just sits alongside your, wid like in Windows? so It actually sits in Copilot. It sits in Edge as the browser as well. Okay. um So whether you're reading a document and you want to kind of ask questions on a document, it can see the document and you can just ask for advice from it My organization is all like Microsoft 365 products and whatnot, but I don't use Edge.
00:07:15
Speaker
I don't it as a browser. And I have this, it's not even an aversion. I just naturally always default to Google Chrome being my default browser. But now I'm going to give Edge a try and try this ah vision companion that you speak about.
00:07:30
Speaker
We've invested a lot of time into the kind of new AI browser. And obviously there's lots of browsers out there, right? we Speaking of AIs, do you have a particular prompting methodology or approach that you use or that you find that works really well for you?
00:07:44
Speaker
You know, be very specific on what you're trying to get out of it. So context, outcome. um as well as tone. Like yeah how do you want the response to kind of come back to you?
00:07:57
Speaker
My point point of view is that over time people have become less structured because AI so models have gotten better at interpreting what we want. Whereas we used to have to be really clear like I needed three bullet points, it needs to look like this and sound like that. And I think there's a bit more flexibility now because it's just better. What's your kind of approach to structuring it?
00:08:16
Speaker
I agree with Nate. Like there's I'm not as formatted, but it's almost like there's been a change of my behavior of how I approach prompting. Oftentimes it's very meta of like writing out and then asking the AI refine this prompt for me to get the most out of ah the the the ask, the request.
00:08:35
Speaker
That's my biggest unlock is the meta prompt of basically going, okay. And it's the lazy person's way of doing it, which is my default method of doing almost anything, which is basically to say, okay, um want I want to build this. i want to create this video with AI. Can you help me create a really great prompt to do this thing? And then I can review that and I can react to it. And I find that that's just a a faster way of getting to a really good place.
00:09:01
Speaker
We definitely see that a lot in the data is people just, want the ideal prompt copy paste. I think going going back to what I was saying with that that vision example, you're you're creating a video instead of actually having to prompt and refine and constantly have that kind of conversation.
00:09:20
Speaker
You can actually have vision, look at the video with you and just ask, well, how do I do that? And how do I remove the background or change the tone or change the voice? And and so that's where it's just iterative kind of a conversation. So often i realize like, how do I do this? And then literally the first step is like, just ask.
00:09:40
Speaker
And yeah i'm I'm no stranger to asking for help if I need help. So it it I do find that just um just whether it's a ah like a short prompt or very structured, well thought out master prompt.
00:09:53
Speaker
But when you're when you're doing that, are you doing it within? ah you using CodePot in PowerPoint or are you going external using one of the others? Well, um unfortunately, he was working on Google Slides, not exactly PowerPoint. and he was trying ask Gemini. Yes, yes. the um The other guys. Yes. but And I ended up asking ChatTBT. But um it's it's funny. i was I was telling Nate earlier that i' I'm now on like the above the free version of Copilot.my organization. And That unlock from the free to the paid.
00:10:27
Speaker
Yeah. Oh my gosh. I am a co-pilot fan now. And I really am. And I'm not just saying that to you because a while back I was telling Nate. like She was not. i was not um I was not too keen on it. But as soon as I got to the paid version, I'm like, this is why it's so good. The kind of analysis it has helped me do in Excel. I use it in Excel.
00:10:46
Speaker
It has been a huge unlock. And um I'm not sure if this is... embarrassing to admit, or if I'm just using AI and, you know, need to be proud of the fact that I'm using it as an assistant, the auto rewrite functionality and outlook.
00:11:00
Speaker
Yeah. I write out an email and I'm like, hmm, this is going to some senior stakeholders. Highlight it, co-pilot, auto-rewrite. And then it writes it. I'm like, hmm, okay, yeah, that works. Replace. Send it off. um So yeah, I'm a big fan. Just a last question for you before we kick off to the more meaty questions. Adam, do you consider yourself to be an AI optimist, realist, or pessimist?
00:11:24
Speaker
I'm definitely the optimist in going, how do you get the most out of it in the context that you're in, like in the platform that you're in?
AI's Role in Advertising and Consumer Decision-Making
00:11:35
Speaker
um And we're heading in the right direction. Are we there yet? We're getting there, but I think the next kind of 12 months where we're headed is very exciting. Yeah, I used to consider it a special ability of mine to like spend an hour or two looking up how to do functions in you know for Google Sheets. you know I was like, oh, I can make it work. yeah But it would be like, yeah now I just ask an AI system to give me the answer and voila.
00:12:01
Speaker
if If you haven't tried it, one last plug, because you're you're talking about Excel, is in in Excel, go equals copilot. And it basically embeds copilot into the actual cell.
00:12:13
Speaker
You can't do that in Google Sheets, I'm afraid, Nathan. So you're going to have to load up Excel for this one. Okay. Yeah, check that one out. I'm looking at all these tricks. I bet you're kicking on the same thing. I've got a whole list of tricks.
00:12:26
Speaker
Speaking of Copilot in Excel and everything else, um I'm a little still confused as exactly what is Copilot and how it's different, I guess, from other models. Because... I kind of, I see it on my Windows PC, but then it's also embedded in things. And and i'm I'm not even 100% sure what it is. So maybe get your definition of what is Copilot.
00:12:49
Speaker
The best way to think about it. And Copilot is very different than the others. I will definitely say that. There's really, in my head, there's really two sides. There's the consumer side of Copilot.
00:13:02
Speaker
That is Copilot as an app. It's embedded into the likes of Bing. It's available on Edge. It's available on Windows. Anyone can use that completely for free.
00:13:13
Speaker
And we have, that's where actually we re rerun ads in that kind of consumer-facing version of Copilot. And then there's what a lot of people think Copilot is, which is it's into Teams, it's into Outlook, it's into Word.
00:13:29
Speaker
And so that is really the the business kind of commercial enterprise. That's what we call M365 Copilot. Okay. So there's the business context and then there's the consumer Do they more or less do the same thing or do you find that they actually are dramatically different tool sets? so So very, very similar, but the business one is grounded in your work context.
00:13:56
Speaker
okay It's privacy compliant. So essentially if you're a business and you load up and you have access to M365 Copilot or you're using it logged in with your work email address,
00:14:08
Speaker
You get a little green tick to say it's like basically safety and privacy endorsed by your corporation. This we hear all the time, which is everything that you're doing in there is not used to train the AI models.
00:14:23
Speaker
And so it's a huge benefit for businesses. Because the number one question I always get from agencies is how do they have ensured that the teams are, what they're inputting into these AIs is not then surfaced somewhere else. But very similar, very, very similar in terms of like output and structure and things like that.
00:14:47
Speaker
I suppose it just depends on the use cases as well. Like the consumer facing side is quite different from how you'd approach it from a business or enterprise wide side. Yeah, not as many secret Santa gifts, I'm guessing, in the business version.
00:14:59
Speaker
Maybe not a secret Santa, but and an Easter egg that a lot of people don't know is that you can access multiple models from Co-Pilot.
00:15:10
Speaker
okay this is This is something that a lot of people don't know is that you've got the Microsoft AI models feeding in, but you've also got partners like OpenAI, where you access OpenAI's models in there.
00:15:22
Speaker
And we recently made the announcement around Anthropic, so you can access Claude models within that too. okay And so that's a big difference for us versus probably some of the others.
00:15:33
Speaker
I love that because Anthropic and Clotter is so well known for like more of the writing styles, but then, um you know, perhaps there's a different use case for using that open AI model. um I know that when Nate read out your bio, you do a lot of ai and advertising. um And without giving away too much to the special sauce, what are the biggest strategic bets that Microsoft is making in that AI advertising space right now?
00:16:00
Speaker
Well, going back like two, two and a half years, we've we've had ads running in Copilot about two and a half years now, which some people still brands and agencies still aren't aware of. um We haven't ramped that up to 100% of volume yet because the huge um positioning from microsoft ai is we don't want an environment a highly engaged environment to just be flooded with ads it's not it's very different than say traditional search right three keywords you get and ads everywhere right that's that's the bread and butter from search engines bing has that google etc etc when you move into like a conversational interface the the flow is different
00:16:47
Speaker
Your organic context is up the top. And we've spent a huge amount of time making sure that if we are going to run ads in that environment, that they are highly relevant and personalized based on the conversation.
00:17:02
Speaker
So text, feed, multimedia ads, they've been running for about two and a half years. So sorry, can I just ask then, so what kind of, how would, what what would a contextual ad look like then? So am I i searching for movie cinema, ah you know, and it would show me like, you know, tickets for the latest Hoyts or something like that? Kind kind of, yeah. So you get your organic results at the top and then down the bottom, we've got something called ad voice.
00:17:24
Speaker
It's a a visual differentiation between the organic result and the ad. But in this case, a cinema could sponsor, hey, you're looking for a certain film as an example, you know, or you're trying to watch Stranger Things as an example, you might have Netflix going, oh, you want to sign up to Stranger Things because we've actually got season five coming out in in a week's time.
00:17:45
Speaker
So it's very relevant to the conversation, but you don't just flood it and have it everywhere. Only one ad but would would a appear. And so sorry, can i ask one more? I'm um'm yeah super intrigued. So from a targeting perspective, then um if you flip it from that's the consumer sees from an advertising agency perspective, like how are they buying that? Are they buying every instance of Millie Bobby Brown or are they buying like a, you know,
00:18:10
Speaker
how do How do they ensure that they're targeting correctly in an AI world, which is going to be fuzzy in its base nature? sure Yeah. so So today you actually run those same same campaigns.
00:18:22
Speaker
So if the same campaigns are applied in into Bing or some of the other search engines that we power, Yahoo, et cetera, then essentially if it's a relevant ad, then it would appear into Copilot.
00:18:33
Speaker
Okay. So you can't buy Copilot and not Bing or Bing not Copilot. Right. so And it's not like you can like skew and buy impressions just in Copilot because that initial prompt has to surface the request. I see. I see. Correct. that That's as of right now. Whether we change that in the future, we'll have to wait and see. I do appreciate that mindful approach because I think and consumers are quite worried about the flood of ads that yeah could potentially come. and And, you know, it's only, we keep talking about it on the pod, like it's only a matter of time before we get ads in chat to you. And we've seen kind of the whispers of um like advertising references and they're like development kits. And we we know it's coming. It's just, we want it to not, um,
00:19:21
Speaker
like feel relevant to the conversation and that's why context is so important. and So I do really appreciate the approach that Microsoft is taking in that space as well. That's the other thing is, um you know, the others are likely to follow suit.
00:19:35
Speaker
um But it's the brands that are testing on us right now and kind of getting the learnings. I think they're setting themselves up for some potential success going forward. So yeah. it's just It's so interesting you say that, I think, um you know especially the larger agencies that have like a lot of data and proprietary technology at their disposal,
00:19:56
Speaker
They'll be at ah at such an advantage. But then you think about like the small indies that don't necessarily have that same um investment into like their own proprietary data and tech stack, but they would have access to a variety of different models. And it' one of my kind of future bets for 2026 is perhaps the rise of indies because they are starting to see what learnings they can unlock from you know multiple models and what they can do a bit differently without necessarily requiring that heavy investment into their own tech stack.
00:20:33
Speaker
The indies have always done a fantastic job on agility, winning new business, but also very quick to lean in on new pilots, beaters, the second that they come out.
00:20:46
Speaker
this How can they show something new to the client then in some cases, um some of the larger agencies aren't kind of tapping into.
00:20:57
Speaker
So what we've seen is, it's actually the small businesses, the independents that are using, whether it's in this case, co-pilot as a product,
00:21:10
Speaker
But more importantly, we're seeing them use Copilot in platform. So we've we announced an integration of Copilot into the Microsoft Ads platform as an example.
00:21:22
Speaker
It's basically like having ah an analyst, an account manager in a box sitting there, you know, what was last week's performance? What was the last three months performance? How do I compare that year on year?
00:21:35
Speaker
And you're getting all of these kind of quick answers, quick snippet answers that then you can relay that back to your client really, really quickly. And to your point earlier, like things that used to take weeks of analysis. Exactly. Time-saving results.
00:21:48
Speaker
The Indies for me are the ones that are looking at that going, oh, I can save an hour here and I can use this and save two hours there. And when you accumulate that, that's how they're efficient.
00:22:00
Speaker
The large agencies there, there are a couple. not Not all, but there are a couple that are really using this pretty well. And that's where they're building agents. you know Everyone's talking about agents.
00:22:11
Speaker
To that point, I mean, if I'm putting my agency hat on, does that mean I'm going to be having to talk to the Microsoft agent and the Google agent and the Meta agent? and Or is there a world where someone like Microsoft or another partner pulls all that stuff together?
00:22:28
Speaker
Well, I think you youre you're actually talking about two two things, right? is For me, there's the... There's the agents that save you time. We've already done that actually. So Copilot Studio, if you've ever used that, you can kind of create your individual agents. That's one way, right? And there's enterprise ones, but also third party developers that have built these agents that allow any of us to go and kind of go, yeah, I want to use that.
00:22:53
Speaker
So Microsoft Marketplace is one of those. You can go there in leverage and leverage any of the others. For the ads side, what we're really excited about is the brand agents.
00:23:04
Speaker
In the beginning of the new year, we're gonna make um a great announcement. It's public already, but we'll make ah ah a great announcement. This is not a scoop? We're not getting the scoop here? This is a scoop, actually.
00:23:15
Speaker
The challenge with businesses is one, how do they show up in AI? But two, the future of the website. We get that a lot. But what we built is a platform that allows a brand to leverage conversational AI on their website,
00:23:34
Speaker
And in the future, what we think are that these brand agents can then be invited into the conversation of your platform of choice.
00:23:45
Speaker
So we can, in theory, let's say if Nike created a brand agent, I go to the Nike website today, i have to navigate through the site in order to find a pair of Jordans, pair of trainers, whatever I want.
00:23:57
Speaker
Going forward, if I open up Copilot, I might say, I'm looking for some Jordans. I don't want to spend any more than a hundred bucks. they might go, ah, do you want to invite Nike to the conversation?
00:24:10
Speaker
Nike come into the conversation and it knows my context and I can have a full blown conversation with that brand. we we've We've had a pilot running in APAC or actually globally for the last couple of months. It is fascinating. And it's something that came up on a previous episode with Gayle Roy. And she's talking about the discoverability of the brand is just such a hot topic right now. So I've,
00:24:34
Speaker
And, you know, Nate and I have discussed at length about how um brands really need, like, it's always been like performance, branding versus performance. And i think there'll just be i ah resurgence of the importance and the need for where a brand actually shows up in the branding itself. Yeah. And we're we're playing around with a few different formats.
00:24:56
Speaker
You know, there's there's a couple of um key invite clients that we kind of bring in that work directly with our engineering. Customer listening is really important to us. is It's not about us coming up with ideas and throwing know like spaghetti on the wall and hope something sticks.
00:25:14
Speaker
I think that's that's hugely the focus for the ai Microsoft AI business is what are are brands and businesses and agencies looking for?
00:25:27
Speaker
And then how do we build new experiences that might help address some of those challenges? You know, agentic commerce, retail is one massive area, but that's not to say all our eggs are in one our own one basket. The brand element is an amazing challenge.
00:25:46
Speaker
You know, what does branding and visual branding look like in a conversational era? we're We're playing around with what that looks like at the moment. So more to come on that next year too.
00:25:57
Speaker
Well, i'm I'm curious from your side of things, because you did touch on at the start how there's you know two versions of Copilot effectively, the consumer side and then the enterprise side. Have there been any learnings for from Copilot on the consumer facing side that Microsoft has come to understand more and perhaps some surprising elements that AI has really revealed about consumer behavior?
00:26:21
Speaker
The funnel is kind of completely shrunk, that we all used to think, you know, awareness conversion, et cetera, et cetera. We've seen that marketing funnel hundreds of times in our career. On so many sites. Right.
00:26:34
Speaker
Every industry event is somewhere someone's got the the the the legacy funnel. We do it too, you know. But our our version is different because
Enhancing Business Productivity with AI
00:26:45
Speaker
consumers not just are aware, but they make decisions literally now in seconds.
00:26:50
Speaker
what used to take in some cases days and weeks, you're getting all of this information curated very quickly and making decisions. I think the stat was um the increases in conversion within 30 minutes since Copilot launched was up 70%.
00:27:08
Speaker
Wow. see When you talk about the funnel shrinking, it's like, yeah you it's not the stage, it's like straight through. Exactly. You're making it, you're understanding what's available, deciding what you're going to buy and buying within that 30 minute timeframe, 70% quicker than legacy while you're using like a search engine.
00:27:33
Speaker
um That decision making is probably the big surprise. it's It's fascinating to hear just how quick that funnel is. And um i don't know, i still I still think that yes, whilst in environments like Copilot, it is so important to you know be there contextually and you know give them the best information. I do think there's still gonna be a resurgence of like the import importance of branding. like we still We're such visual creatures. We still need to see out of home placements and um like beautiful TV ads that spark emotion.
00:28:07
Speaker
It's the middle of the funnel that's disappearing, right? You're going to have that brand building an awareness piece. And then once I'm ready to move, you move. I mean, yes, in some in some cases. But I think, you know, going back to CTV and display and that's where that there's a product called Copilot Discover.
00:28:30
Speaker
MSN, the old em ah MSN has had a massive resurgence. It's still that today, what a lot of people don't know is it's still the number one desktop news site in the world. What? What? Yep, still today. you see?
00:28:45
Speaker
and And the reason being is because it's a very clever approach that you have MSN as you always used to. The number of people that navigate from a browser like Chrome over to MSN to get their news and personalized news content.
00:29:01
Speaker
that's That's one option. The second is you put Edge on 1.4 billion Windows devices around the world. And when they open up that edge browser, the default is what is the legacy MSN.
00:29:16
Speaker
yeah Or if you click on a new tab, you might not have the MSN logo, but you've got a skinned version of MSN. And what Microsoft have done is something called Copilot Discover.
00:29:28
Speaker
So this is, it's a blend between the copilot experience and a very hyper-personalized news curated feed.
00:29:41
Speaker
That is where you're going to have display and video. We do it today. i mean, is when you think about how you navigate the web, whether you're going to a chat interface and having those conversations, but going back to, or as you navigate any other browser or a new tab, you're still going to have the opportunity to kind of hit remarketing and video and display content throughout that. And that's those brand catching opportunities that you can do that too.
00:30:08
Speaker
Don't know about everyone else doing something similar, but that's our approach. We've talked a lot about the kind of the funnel and the branding and the future of all of it. And and I think that's going to be one of the most interesting aspects of how AI rules out. And I don't think anybody has a ah solid, you know, if if you told me you had the answer right now, I would say you were almost certainly lying, but it's going to be an interesting way that the whole industry changes over time.
00:30:36
Speaker
um Speaking of the industry changing, I know that you're um on the IAB ah Council, AI Council, along with Kellen Coetzee, one of our our first guests, in fact.
00:30:49
Speaker
um And I'd love to know a little bit more about what the AI Council does. and I know you guys just released a new report. Is there anything you can tell us about that? so the group really is trying to put our best foot forward and help the industry from quick tips and tricks. you know We've released a prompt guide, know different prompt structures for us, different than Gemini and ChatGVT, et cetera, et cetera, which is great because some people just don't know where to start.
00:31:20
Speaker
I think that's a massive, massive thing for us is how do we help every marketer, everyone in the industry get a starting point for those that haven't jumped in two feet yet?
00:31:33
Speaker
I think the other side of it is things like measurement, guardrails, governance, um the importance of a structured approach to privacy.
00:31:46
Speaker
You know, it's massive in the industry. Obviously, it's hugely important for every single brand. um And so we're building research, tips, tricks, tools um as kind of destinations that anyone can kind of go to and and leverage.
00:32:01
Speaker
And we know that even since since that those reports, other markets, US, UK have looked at the IAB, Australia, AIB team are gone. Fantastic work. Thanks very much. Can we kind of apply it universally in other markets, which is great.
00:32:15
Speaker
It's is's fascinating that you're really at the forefront of um AI for you know ah learning from the council, what they can share, and then also like what you see personally from your work in Microsoft.
Evaluating AI: Overhyped vs Underrated
00:32:29
Speaker
but Taking a step back, if you look at overall the AI industry, um what do you think is the most overhyped aspect of it? And conversely, what do you believe is the most like underrated aspect that people just aren't talking about enough?
00:32:46
Speaker
I think that the challenge when new models are launched or new products are launched, there's an initial surge in usage. how do you use this new capability that's going to save you time as opposed to, yeah I can create some great imagery or create a quick video and it's great for a meme or something over here. But if I don't have it applied in my day to day, how is it really going to benefit me today, but also my role going forward?
00:33:16
Speaker
In a meeting recently, someone, um you know, shared, oh, I've been using co-pilot and it has literally given me two hours back a month that i previously had to do all this, you know, um, individual analysis on this project. So I think just quantifying those wins and sharing it, it will elevate us and, to your point earlier, a rising tide lifts all boats, yeah will become better marketers and better advertisers with the assistance that is AI, not the fear that it's just going to take our jobs.
00:33:48
Speaker
and And I think what you're talking on there is is an amazing kind of tie-in The reason why it's copilot is it's meant to be an AI companion.
00:33:59
Speaker
Like it's meant to work alongside you. Right. And the the reason why Microsoft put copilot into PowerPoint and Excel and teams, et cetera, et cetera, is in the context of where you're spending your time.
00:34:11
Speaker
How do you get the most out of it? There is a, an amazing. Piece called copilot insights. If you haven't looked at it, that for me is where you're heading with how do you quantify where the time is saved and you can get this report. It's amazing to see the time that you've taken using Copilot in say Excel across your entire team.
00:34:38
Speaker
How many hours would that have saved? And in PowerPoint. And when you look when you apply that across analysts that live and breathe in data and analytics all the time, versus the creative teams that are building pitch decks and imagery and video, when you apply it across, I think that's where you want to see how do you save time across the context of where you are in the flow.
00:35:03
Speaker
um That is what people are looking for. I think people are absolutely desperate for the insight to say, How do you, end is if this is a cost, what's my savings on the back of it?
00:35:16
Speaker
You are giving us so many good. No, I feel like I'm crammed in throwing all of these words in. so No, it's um it's it's fascinating. And to be honest, I do think that the the winner, the long-term winners...
00:35:32
Speaker
will be Google and Microsoft because Google you know is well integrated into a consumer's behavior. um And then Microsoft has just truly captured the enterprise market that, you know ah going back to my personal experience, when I went from the free trial and I had to you know write the business case and submit it and um I got the approval for the paid version, it has just been transformative. Now, just that that green tick of approval in the top corner of Copilot saying, like, it is endorsed by the business. Yes, use this. um i'm I'm keen to hear from you, and I know it's not necessarily your domain within Microsoft, but for companies that, you know, do are already on the Microsoft suite, um what do you see to be the biggest friction point into scaling AI usage effectively?
Challenges and Future of AI Adoption
00:36:24
Speaker
I've definitely seen from my own experience, I can't talk to to to everyone, but sometimes, I won't name which agencies, but sometimes they sign up to incredible technology that might be the decision is made in Europe or the decisions made in the US.
00:36:45
Speaker
And sometimes that never trickles its way down into Australia, into the corner. So they've had access to these things, but they just don't know. that is That is our challenge is to get them aware of what they've got and then how do we help them use it to the best of their ability.
00:37:06
Speaker
So that that sounds like a ah combination of education and culture. Education, culture, but also, you know, you've got leaders and I'll call her out. Rose Herseg is absolutely incredible.
00:37:19
Speaker
But Rose very quickly, years ago, stood on stage and said, hey team, go and use all of the tools and see how to use it. So from the top down, as we tried to emphasize a culture of test, learn, see what you can do.
00:37:36
Speaker
Some agencies are trying to restrict it. It's like only certain job functions get to access certain things.
00:37:44
Speaker
Or the IT t team have access to this, but the marketing team don't. And you're like, well, the marketing team are the ones that would use it more than the IT team. And really benefit from it, yeah. Right. So it it has to come down to culture.
00:37:59
Speaker
It has to come down to education. And just trying to, I mean, we call it like trying to democratize it, is if you put the tools into people's hands, you're going to find use cases and people using it in ways that you never thought originally.
00:38:13
Speaker
Adam, thank you so much. and I think we've had you for over an hour. It's been an amazing chat. Just to sort of wrap things up, um
00:38:23
Speaker
what's one skill that you would tell people to master now to be relevant alongside AI going forward? I now do these six, seven minute, very quick podcasts about a certain topic on my way to work.
00:38:37
Speaker
And on the consumer facing version of Copilot, you can open it and go, create me a podcast about agentic commerce. I did this literally for a panel that I was presenting in in Melbourne a couple of months ago.
00:38:48
Speaker
Is what's the future of agentic commerce in the next six months and what's Microsoft's perspective, what's OpenAI's perspective, et cetera, et cetera. And very quickly, you're getting these six minute learnings in a vision in a video or or a audio format.
00:39:05
Speaker
that I learn by listening, but that's my my strength. So I think the skillset is what features or capabilities that these have, how do you quickly learn and use micro learning as that ability?
00:39:23
Speaker
That's how people are gonna catch up to what what's happening. What you're saying is you know people need to continue their um their learning journey, but also understand how you learn best and really lean onto AI to fuel that.
00:39:36
Speaker
it's Exactly. if you had If you'd done that two, three years ago, how long would it have taken you to get up to speed with some of those things? Now you kind of we all take it kind of for granted for those that lean in.
00:39:48
Speaker
is You can go and learn something as quickly and easily as just asking it to go and create you a five five minute video of a certain topic and you've got a video now that that's where it's headed. So those micro learnings pieces is certainly what I'm trying to use. And um as a starting point, my advice to everyone would be play around with the different ones, see what you can create, learn, see going back to how do you learn and what can you use to help you do it in a quickly and easily format. Last question for you, Adam, if you can fill in the blank. In three years time, ai will be blank for marketers.
00:40:27
Speaker
I think the one thing, and I'm going to get back to the opportunistic piece is, and whether this is a tie-in or whether I'm going completely off tangent here, have you ever seen the film Pay It Forward?
00:40:39
Speaker
i think so. It's about a little boy that starts i doing something good for someone, not expecting anything in return, but all the hit that little boy asks for is that that person pays it forward to someone else.
00:40:53
Speaker
okay not What I'm trying to say here is, as an industry, collectively, we're all using these AI models, capabilities in our own way.
00:41:05
Speaker
How do you help the next person next to you leverage or learn that something that you're doing And how do they then take it forward and pass it on to someone else? So I want to put it forward to everyone and say, I hope we're in a market and an industry that is absolutely thriving from an Australia perspective and showing the rest of the world that we're an industry in the region that far outpaces others by helping each other kind of improve and use it that way.
00:41:34
Speaker
I think we should rename the podcast to the paid AI Pay a Forward podcast. so Yeah. Adam, it has been just such such a pleasure speaking with you. There was definitely some aha moments for Nate and And really appreciate your time this morning. Thank you very much.
00:41:49
Speaker
Have a good one. Nate, what did you think about that chat with Adam? I really enjoyed it. Having the perspective of somebody who lives and breathes and sits in a company that is creating foundational AI content and models, I thought was really helpful. It was a different perspective than we've had before. And I learned a lot.
00:42:09
Speaker
He has a very interesting perspective of what Microsoft is doing with Copilot and all of the you know cool bells and whistles and you know the nice plugs that he was putting in, but i'm also just what agencies are doing quite well. um i really appreciated our little short chat on indies and how they are just gonna smud. I really do think 2026 is the year of indies. Well, I know a bunch of people in in the industry who seem to say the same thing, although you've just seen what the disappearance of Cramerama and a couple of big acquisitions recently. Yeah, with Chimera going. um
00:42:45
Speaker
and But, you know, that's the that's the thing. like that's the I think the natural evolution is like a business grows, grows, grows, and then eventually gets acquired. like it's Yeah. Yeah. it it's the It's the way to make money in advertising. if Create an agency and sell it back. Yeah, exactly. um One of my key takeaways as as well was um to start using Edge as a browser. um Didn't even you know consider that. And now when he was talking about all of you know, the vision and and the type of um Microsoft tools and Copilot enabled, and you know, activations. i I'll be starting that and after we hit stop on this recording. Okay, you're going to edge it away. um
00:43:29
Speaker
You know, really appreciate your time this morning, Nate. um Merry Christmas. It's been the heck of a year doing this with you. And to all of our listeners and viewers of the podcast, we would love if you could like, share, and subscribe. if you have any questions on AI or the show, feel free to email them through to mindthemodelpod at gmail.com.
Podcast Conclusion
00:43:50
Speaker
And remember, as always, the intelligence might be artificial, but the wins are real. See you next time. This episode was brought to you by Testchimp. Testchimp is a web developer's best friend.
00:44:04
Speaker
Testchimp uses deep runtime context to automate your testing workflows, generate scenarios, write scripts, and fix bugs all in one place. Stop wasting time and start shipping quality code faster.
00:44:18
Speaker
Bugs suck, but finding bugs doesn't have to with Testchimp.