AI as a Thinking Partner: ADHD, Agents, and Human-Centered AI with Adam Federman

Lauren: Mm-hmm. Derek: Okay, so get this. Today we are officially in AI as your thinking body mode. Speaker 3: Oh man, finally an episode where the AI is not your boss, it's your brain's chaotic roommate. Derek: Exactly. Welcome back to Tech Insider Weekly. I'm here with Derek, and this is a special interview episode. Speaker 3: Yeah, today we've got Adam Federman joining us and he has receipts on using AI as a thinking partner, not just a glorified to-do list. best. Derek: So we'll talk ADHD and generative AI, how a noisy brain can actually click with these tools instead of getting steamrolled by them. Speaker 3: Right. And then we'll zoom out into the enterprise world where Derek from procurement meets AI and suddenly the human is the new bottleneck. Derek: Wait for it, because Adam has lived that 50 internal agents only a handful survive life and the stories are wild. Old. Speaker 3: Plus, we'll dig into what adoption cycles really look like when your pilot doesn't just go to production, it has to survive office politics. Derek: Oh, you're going to love this part because it turns all those glossy AI case studies into like actual human drama. Speaker 3: Wow. Alright, enough teaser trailer, let's get into it. Derek: Welcome back to Tech Insider Weekly with our special guest, Adam Federman. Speaker 3: Segment one, ADHD and generative AI as a As a thinking partner starts now. Speaker 4: Hi, Adam. Thanks for joining me today. We'll spend about 10 minutes discussing AI as a thinking partner, not a productivity tool, then wrap up with final thoughts. Does that sound good before we begin? Speaker 5: Yeah, that sounds great. Speaker 4: Wonderful. So let me start here. You've talked openly about having ADHD and spending years feeling like your ideas were moving faster than you could get them out. When Gen AI finally clicked for you personally, what did that moment actually feel like? Speaker 6: What did that moment feel like when Gen AI clicked for me? Well, I would tell you that it's not like it just clicked. There were stages along the way. One of the things that I was able to do was, I'd say it's kind of like brain vomit, where I can just dump everything out of my brain into AI on a continuous stream of thought. And the way my brain works with ADHD is that some of these... Topics that I start touching on have a lot of detail into them and I can go all the way into that detail but I can come all the way back and start talking about the larger subject again but it takes time and for a human it's a lot to listen to but for an AI it actually is very easy to listen to and in fact it provides enriched context on things that I'm doing so it actually knows how to help shape my ideas with me and ask the questions that need to be asked. Yes, that was a real aha moment for me. Speaker 4: That's a powerful way to describe it, the AI becoming a real thinking partner that can actually handle the way your mind naturally works. Given your background in both enterprise sales at CDW and machine learning at Remark Systems, how do those two very different worlds shape the way you approach deploying AI in large organizations today? Speaker 6: That's a great question. One of the things I would say is I've seen the infrastructure, the enterprise infrastructure application of technology, and I've seen the endpoint application of technology in both my time at CDW and Remark Systems, and I see it here at Accenture as well. And what I've noticed is there are certain things that... um a user needs to understand and believe in before it can trust something like a search engine right if you were around when the search engines were brand new and you put in a search request you started to put in like full sentences at first and your results would really stink but then Google had the predictive predictive text that would come up and when that predictive text would come up you started to realize That less was more. If you could get the keywords, it could sort it and it would almost predict what you were saying based on other searches people had done. Fast forward now to AI and a user is very used to using that Google shorthand way of searching or asking for assistance. But really, with AI, it's more context is more valuable. And so now I'm seeing how we revert back to almost the beginning of search engines where you need to put as much as you can so the AI can truly help you. And so when you ask what my background has taught me, that there's these cycles that happen in AI, in technology, and AI is showing them once again where... We have data, networking, and processing power. And those three elements are kind of the loop that kind of drive or slow down or become the bottleneck for the next wave of technology. And I've seen this throughout my career in every place I've worked. And now for the first time, the human is actually the bottleneck in some ways, because we need to better understand now that the data can flow freely and the processing power is pretty phenomenal and the networking speed is as fast as we need right now. It's no longer about bandwidth. It's no longer about storage capacity or recalling of information. It's no longer about the ability to process large amounts of data. Now it is about the human understanding how to leverage those things in plain English. We don't even have to know how to program. And so all this history I've had across all these different areas has really just kind of come to another cycle and I can see all these, I can predict what's happening and we need to get people on board before we can get entire companies on board, if that makes sense. Speaker 4: What separates the ones that actually survived from the 45 that didn't? Speaker 6: Oh, on my AI agents, they were built early. And what I'm, I think anyone building anything in AI right now knows that when you build something and technology continues to get enhanced on the backside. So what you're building can actually Lauren: become more or less useful depending on how the models are maturing so When I looked across the 50 different agents I made, they were really only made in a way that I knew how to use them. I couldn't really give them to someone else to use. And then I also realized some of them were just me kind of automating very simple tasks by themselves, where really those simple tasks could have become features of other agents. And so it's not necessarily that they failed. It was more of the... If I have these as individual agents, I have to then remember when to use these 50 agents and that in itself could become cumbersome. So the ability to know what skills you have, put them in the right agents and then either do a agent skill or a a separate skill altogether for your AI was really helpful and so when I say only five made it it's really five core agents that do a lot of different things with different personas and different perspectives and that anyone could use and it's also secure so that if it scales it's not going to drift and it's going to force users to think and it forces users to ideate with them it's not a It's not a do this for me, here you go type of agent. Those are more or less like printers, like you're just asking it to complete a task and it does it. These are more thought provoking, collaborative and ideating agents that can eventually build something but force you to think through the process first. Derek: That's a really sharp distinction. Moving from 50 isolated task automators to five collaborative thinking partners that actually force you to stay in the process. You mentioned forcing users to think and ideate rather than just execute. In your experience, when you put one of those five agents in front of a team that's used to the do this for me model, what actually changes in how they work? Lauren: Oh, man. So I think the biggest thing is... People think that agents are going to take over their job or AI will take over their work. And if you collaborate with an agent, you'll see that they won't take over. They don't do they don't have the artistry that we have as humans to know the to know the full. career trajectories someone had and the perspectives they've seen everyone's perspective is unique and that's one thing that AI can't get it it really becomes a mishmash of perspectives that it's then using algorithms right to figure out the highest probability answer based off of mathematic models and the biggest distinction I think when you get people to actually think with it along the way is you force cognition to stay on the user but you're removing the cognition what's the word I'm looking for you're removing the cognition kind of burden as well on the things that are repetitive that this person has done forever but necessarily can't explain it it's almost like the tacit knowledge part of it and so that's where things become interesting and that's where again bringing forcing the user to collaborate with the AI can really take everyone's artistry to a new level Derek: That's a beautiful way to frame it. You're not replacing the artistry. You're freeing up the mental space for it. Thank you so much for sharing your insights today. It's been a pleasure. Lauren: Thank you. This was really cool. And can't wait to talk again another time. Derek: Thank you so much, Adam. This has been fantastic. I really appreciate you taking the time to dive deep into how you're thinking about AI as a true thinking partner, not just a productivity tool. Your perspective on building for human capability rather than replacement is something our audience really needs to hear. Thanks for being here on Tech Insider Weekly, and we'll definitely have you back soon. Speaker 3: That was such a great conversation with Adam. Speaker 4: Right. I feel like he completely reframed the way I think about AI. Like we always hear about productivity this, efficiency that, and he just kind of flipped the whole script. Speaker 3: Exactly. And I think the thing that hit me hardest was when he talked about his ADHD and how generative AI became this thinking partner for him. Not a crutch, not a shortcut. shortcut, but something that actually helps him externalize and organize his thoughts in a way that finally clicked. Speaker 4: Yeah, that was really personal and honest. And I think a lot of people listening probably felt seen by that because there's a stigma around needing help thinking things through and you just normalize that in such a genuine way. Speaker 3: Totally. And then it connects so naturally to what he said about team adoption, forcing people to actually. we ideate and not just execute. That's a harder sell inside organizations, right? People want the easy button. Speaker 4: Oh, always. But I love this point that if you skip the thinking phase, you lose the value. The AI just becomes another tool you half use and eventually ignore. Speaker 3: And that tracks with what he shared about AI agent consolidation. He started with like 45 agents and whittled it down to the ones that actually made him think differently. differently, that kind of discipline is rare. Speaker 4: Seriously, most people just keep piling on new tools. The fact that he was willing to cut what wasn't serving the deeper goal says a lot about how intentional he is. Speaker 3: Reflectively, his enterprise background at CDW and Remark Systems clearly shape that. He knows what adoption actually looks like at scale, and he's not romanticizing it. Speaker 4: Real talk from someone who's lived it. That's what made this one so good. Speaker 3: Couldn't agree more. All right, we'll wrap things up right after this. Stick around for our outro. Okay, so get this. We started with ADHD brain chaos, and we somehow landed on AI as this calm, slightly nerdy thinking partner instead of a bossy productivity cop. Speaker 4: Dude, yes. The big thing for me was your line about humans being the new bottleneck. That hit hard. Speaker 3: Right! If there's one takeaway, it is this. AI gets powerful when it thinks with you, not for you. Speaker 4: Wait for it-that also explains why half those internal agents you saw just fizzled out. The ones that stuck actually Speaker 7: - Lauren: fully collaborated with people instead of screaming tasks at them. Derek: Wow. Lauren: Exactly. Derek: Okay, okay, okay. Quick thing. If this episode gave you ideas, subscribe, drop a review, and send it to that one friend who lives in their notes app. Lauren: And if there's a founder or topic we should hit next, tag us and pitch it. Derek: Thanks for hanging with us. Lauren: New episodes every Wednesday. Derek: See you next week.