Designing products for AI agents

On this week's episode of Inside Outside Innovation, we talk about the rise of Identic AI, why you need to build for AI agents first, and how AI is bigger than SaaS ever was. Let's get started.

Inside Outside Innovation is the podcast to help innovation leaders navigate what's next. Each week we'll give you a front row seat into what it takes to grow and thrive in a world of hyper uncertainty and accelerating change. Join me, Brian Ardinger and Miles Zero's, Robyn Bolton. As we discuss the latest tools, tactics, and trends for creating innovations with impact, let's get started.

Podcast Transcript with Brian Ardinger and Robyn Bolton

AI Agents, Personal Concierge Tools, and the Future of Innovation

[00:00:30] Brian Ardinger: Welcome to another episode of Inside Outside Innovation. I'm your host, Brian Ardinger. And with me I have Robyn Bolton from Mile Zero. Robyn, welcome back. 

[00:00:50] Robyn Bolton: Thank you. Glad to be back. 

[00:00:51] Brian Ardinger: We are always on the hunt for new and innovative things. Every week we try to bring you some of the most interesting articles or things that we've come across in our world of innovation. We'll just jump right in.

The first article we wanna talk about today is called, With the Rise of Agents, we are entering the world of identic AI and this is an HBR interview with. Don Tapscott. This is actually a podcast that HBR put out, and Don talks about this movement of not only AI agents, but the fact that you're going to have AI agents that are identified specifically to you and your tastes almost like your virtual concierge in a variety of different topics.

And these agents will know everything about you as well as everything about what they need to do as an agent. And this world is going to fundamentally change the way we do business, et cetera. 

[00:01:39] Robyn Bolton: Yeah, this is an interesting one and it feels both very kind of sci-fi and very likely to happen tomorrow. I'm skeptical on the timeline, like I totally believe this will happen.

I don't really think it's going to happen in the next few years, especially because you know, yesterday I asked Claude to proofread something for me. I gave it a document, and it went off and proofread a totally different document from a different chat. So, if AI can't handle a straightforward request like that right now, I don't think it's anytime soon going to be understanding my judgment and my values and taking actions on my behalf. You know, could it happen one day? Sure. Why not? 

[00:02:19] Brian Ardinger: It will be interesting to see, I mean, we're seeing a lot of experiments out there with Clawbot and that people are jumping headfirst. I saw a Twitter post, there was an event in New York, I think yesterday where 2000 people who were doing things with their Clawbot got together and talked about what they were doing with their Clawbots.

Building for AI Agents First. Product Design, Trust, and What Comes Next

It was interesting from the standpoint of the amount of energy and excitement around it. But then on the flip side, a lot of the conversation was there wasn't still any real meat around it. It was nice to have testing, experimenting those tests and those are experiments will, you know, hopefully result in something, but I think we're not quite there yet.

But it is interesting to peer into the future. What's so exciting about the Clawbot scenarios and that is the fact that it really did give a vision of, oh, what happens if this could actually do this? And it opened up a whole new conversation pieces where it moved it beyond, oh, this is just a Google Chat bot kind of experience. And I think that's where that genie's not going back in the bottle. 

[00:03:14] Robyn Bolton: No, it's very exciting. It's still a ways off, probably years, not decades, but. 

[00:03:20] Brian Ardinger: Can only change, it's also hard to put everything that you own, you know, all your personality and all your quirks and everything into a bot so that it can do things for you when you don't trust the bot. So. 

[00:03:31] Robyn Bolton: Yeah, I just imagine trying to do that with a bot and being like, no, thank you. 

[00:03:34] Brian Ardinger: There are things I don't like.

[00:03:35] Robyn Bolton: You can keep your quirks. Yeah. 

[00:03:38] Brian Ardinger: Alright with the second article, Why you need to build your product for AI agents first. So tangentially similar to what we were talking about previously. This is Peter Yang wrote an article talking about since the structure of how you are building is changing. If indeed agents are going to do the bidding for you in a variety of these things, you no longer need to necessarily build for people going to your website or using a user interface because you are building for agents who are talking to other agents who are doing things.

So if you're in a new builder today, some of the things you should be looking at is how can you actually build for agent flow and how can you build so that the agents can work faster, and what might you strip away and what might you change? Based on that particular paradigm shift. 

[00:04:23] Robyn Bolton: Yeah. I mean, I am always one for, you know, simplicity, like getting to the root, getting really clear, really simple. And there's a certain amount of complexity that's required and things. It does get stripped out by AI as it, you know, goes through and kind of does the regression analysis or the prediction analysis and all of that.

Designing Products for AI Workflows, APIs, MCPs, and Human Experience

So I've really conflicted reading this one because I'm like, well, I don't know how to design for AI. I don't know what that means, especially because things are moving so fast and since another instance where Clawbot shows up as a big character in the story. But I was also like, what if I don't want to?

Like what if there's more nuance? What if there's more richness? What if there's things that will get lost if I design for AI? And that of course could sound like the death throes of the human. So yeah. So I was really conflicted. But I think it's, it makes a really interesting point and an important point that we've got to figure out.

[00:05:24] Brian Ardinger: If you strip away the beautiful UX that people have designed to make you feel the emotion around the product, and the agent never interacts with that, how does that change the product itself? Yeah. Or the experience that you're creating. The article also goes on to give you some of those skill sets of like how to think about this if you are building.

He has some talking about the APIs or the tools, the skills are the recipes, and then the MCPs are actually the kitchen and how it bundles it all together and how these particular components of AI. The way these AI tools are coming together, you can create environments and that such that you're building for agents to make it easier for those folks to do y...