Full transparency: This post was developed collaboratively with AI - specifically Claude - as a research and writing partner. The perspectives, positions, and voice are mine.
After hosting more than 120 conversations in the Elevate Your AIQ ® studio about AI, work, and human potential, I thought I would recap some of the highlights and insights we’ve heard so far this year.
I originally started the show because I believed these conversations needed to happen out loud, in public, with practitioners, researchers, educators, and futurists who could provide us with the cognitive diversity and lived experiences that help us put it all in perspective. What I didn’t fully anticipate was how much - or how quickly - the conversations themselves would evolve.
In the first 50 episodes, the dominant energy was anticipation. In episodes 51–100, it was urgency. In 2026 so far, something has shifted again. The guests who joined me in the first (almost) half of 2026 weren’t speculating about what AI would do to organizations. They were reporting back from deployments already underway, decisions already made, and consequences already visible.
The future of work is a strange concept, especially when technology and change are happening so fast it can feel like the future is now. What does it mean to stay human through one of the most disorienting transitions any of us has ever navigated?
Here’s a summation of what I heard.
If I had to name the single most underestimated factor in AI adoption across every industry I’ve covered this year, it would be trust — not AI’s trustworthiness in a technical sense, but the broader web of trust that has to hold for any of this to work. Trust between organizations and their data. Trust between employers and employees. Trust between systems and the humans who are supposed to rely on them. Trust between candidates and the hiring processes that are increasingly mediated by AI on both sides of the table.
Vijay Swami of Draup put it plainly: transparency is about giving people the right to know, even when they don’t ask. His team built traceability and explainability into their talent intelligence platform by design, not as a retrofit — because they understood that trust, once lost, is nearly impossible to rebuild in enterprise relationships.
Adam Gordon at Poetry made a similar argument from a different direction: the “trust barometer” in organizations is dropping, and when that happens, candidate experience and employer brand collapse together. His response was to build guardrails so tight that the platform literally cannot hallucinate — removing variability as a trust liability.
Dave Vu at Ribbon framed trust as a two-sided problem in high-volume hiring: candidates need to trust that AI screening is fair and that their time is respected, while organizations need to trust that AI outputs are consistent enough to act on. His point that transparency about AI usage actually increases candidate adoption — rather than creating resistance — was something I found genuinely counterintuitive and worth thinking more deeply about.
Stephen Messer at Collective[i] and Intelligence.com took the longest view on this. His argument is that trust isn’t just a nice-to-have in AI-powered commerce — it’s a transaction accelerant. Organizations with trusted relationship graphs move faster, close more, and waste less. The friction that most companies accept as normal is, in his framing, a trust deficit masquerading as a process problem.
Jerry Jao at Employ, processing nearly 100 million applications annually, is watching trust erode in real time from a different angle — AI-optimized resumes are degrading signal quality, deepfakes are appearing earlier in hiring funnels, and the volume problem has become so severe that meaningful evaluation is genuinely at risk.
And Laura Maffucci at Globalization Partners raised the sharpest version of a concern I keep hearing: when people turn to unreliable general-purpose AI instead of trusted, expert-grounded data, the consequences in compliance and legal contexts can be severe. Her line about hoping Reddit isn’t anyone’s first stop for a legal question landed hard.
The through-line across all six of these conversations is that trust isn’t a feature you add to an AI system. It’s the condition under which the system gets used at all.
AI adoption stories that sound anything like “deploy tools, track productivity improvements, declare successful adoption” abound. I have spent much of 2026 talking to people who acknowledge that this is an immature approach, not to mention that it ignores the lessons we should have learned from prior tech-influenced transformations. Most transformations are failing, and that result was entirely predictable where leaders treat AI transformation like a technology deployment.
Russ Fradin at Larridin has built his entire advisory practice around the measurement gap. His observation is blunt: most organizations have no idea which AI tools their employees are actually using, let alone whether those tools are driving outcomes that matter at the enterprise level. Individual efficiency gains — time saved per task, outputs per hour — are real, but they’re also necessary but not sufficient. The companies that will win aren’t the ones that deployed the most tools. They’re the ones that redesigned work around the tools and reinvested the gains into growth rather than just margin.
Paul Rubenstein at Visier brought a framework that I found genuinely clarifying. He tracks three curves for CHROs navigating AI: the human-machine efficiency frontier, the ROI curve on AI investments, and what he calls the humanity index — a composite of talent density, engagement, and culture. Most organizations, he argues, are stuck in what he calls the “gym membership phase” — they distributed the tools, paid the subscription fees, and assumed transformation would follow. It doesn’t. Real returns require intentionally deconstructing jobs and reassembling them around AI capability, which is harder and slower than buying software.
Charlene Li, returning to the show fresh off the release of Winning with AI (co-authored with Dr. Katia Walsh), made the most provocative argument in this cluster: organizations should stop running pilots. Pilots, she argued, are just a socially acceptable way to procrastinate the real decision — yes or no — and they give everyone cover not to commit. The organizations she studied that are actually winning led with business strategy first and asked how AI could serve goals they already had, rather than searching for problems worthy of their new tools.
Melissa Reeve, whose Hyperadaptive framework maps a five-stage journey to becoming AI-native, diagnosed the same problem from the organizational design side. She calls individual productivity gains a vanity metric. The real question is whether AI is unlocking new organizational capabilities and a more ambitious mission — whether it’s expanding what the organization can be, not just how efficiently it can do what it already does.
What strikes me across these four conversations is that the measurement problem is also a courage problem. Knowing you’re in the gym membership phase requires admitting that significant investment hasn’t yet produced significant transformation — and that the path forward involves more disruption, not less.
This is the question I keep returning to, episode after episode, because I don’t think most organizations have seriously answered it. They’ve answered the adjacent question — what tasks can AI do? — but not the harder one: given that, what are humans distinctively for?
The most rigorous answer I heard this year came from Jonathan Aberman at Hupside. His Original Intelligence Quotient framework is built on a precise definition: originality equals novelty plus salience. AI can generate novelty — it’s quite good at it. What it cannot do is determine what’s meaningful to the humans consuming its output. That judgment remains a human attribute. Organizations that misunderstand this, that treat AI output as inherently valuable rather than as raw material requiring human curation, are building on a faulty foundation.
Bob Danna, physicist, former Deloitte partner, and author of My Curious Life, lived a version of this question when he and collaborator Joe DiDonato built “Bot-Bob” — a digital twin trained on his memoir, writings, and decades of accumulated experience. What emerged from that experiment wasn’t a replacement for Bob Danna. It was an amplifier — a way of making his judgment, values, and perspective available at scale. His framing stuck with me: the knowledge worker of the future isn’t someone who competes with their digital twin. It’s someone who collaborates with it.
Jacob Bank at Relay.app has been thinking about AI agents longer than almost anyone I’ve talked to, having run the multi-agent systems lab at Stanford before founding an AI calendar company that Google acquired. His current argument is that the humans best positioned for the AI era aren’t necessarily the best coders or the most technically sophisticated. They’re the best managers — people who know how to give clear direction, set expectations, provide feedback, and hold agents accountable. “We’re all managers now” is how he put it, and I think it’s one of the most practically useful framings I’ve encountered this year.
Oded Dubovsky, my former IBM colleague who now runs STRAIX, brought the longest institutional memory to this question. Having spent over two decades at IBM Research building cognitive computing systems before they were called AI, he’s watched multiple waves of “this changes everything” wash through organizations. His contribution was a kind of epistemic humility: slow down to think carefully before prompting or building, because the questions you ask determine the quality of what you get. Einstein’s 55/5 rule — spend 55 minutes on the problem, five on the solution — is more relevant in an age of instant AI output than it has ever been.
The concept that “leaders need to get on board” has been a recurring theme in AI transformation conversations, but it doesn’t tell us enough. What I heard in 2026 is more specific than that; and perhaps more uncomfortable. The leaders who are getting this wrong aren’t necessarily skeptics or technophobes. Many of them are enthusiastic AI adopters in their personal workflows. What they haven’t done is the harder internal work: genuinely reckoning with what AI means for how they lead, what they model, and whether they’re willing to be visibly uncertain in front of their teams. You can’t ask your people to embrace disruption and self-reinvention while projecting certainty from the top. That gap — between what leaders say about AI transformation and what they’re actually willing to risk personally — is what I kept running into this year. (I should also note that Charlene Li called out a variation of this in her first guest appearance (Episode 19) when she talked about the Knowing-Doing-Leading gap.)
Juan Garcia at Tuio, a digital-native insurer in Spain that has genuinely reimagined how AI works inside an insurance company, named it directly: the real barrier to AI transformation is organizational courage. Not resources, not technology, not talent — courage. The willingness to rework processes and structures around AI in ways that require leaders to take personal risk for uncertain outcomes. His team made a decision I found remarkable in its simplicity: they would never automate negative customer decisions. Not because regulation required it, but because their values did. That kind of clarity requires leadership that has actually thought about what it believes, not just what’s efficient.
Meg Bear, former President of SAP SuccessFactors and one of the sharpest observers of organizational behavior I’ve had on the show, came at the courage question through her critique of what she calls “founder mode” — the belief that a single visionary individual drives organizational success. Her counter-argument is that sustainable transformation requires intellectual humility, collective intelligence, and the willingness to bring your learning self rather than your knowing self. Her line that the only way to preserve your relevance as a worker is to make your own job obsolete is the most honest statement of the individual challenge I’ve heard anywhere this year.
Tim Borys, executive coach and host of the Working Well podcast on the WRKdefined network, located the courage gap in the individual as much as the organization. His framing — that humans have a “human operating system” that most people never optimize — resonated with me because it puts agency back in the hands of individuals rather than waiting for organizations to get it right. Adaptability, in his view, isn’t a soft skill. It’s the master skill that determines whether everything else compounds or decays.
Lisa Cole at 2X brought the marketing leader’s perspective on a specific kind of courage: the willingness to decide what should remain distinctly human even when AI could technically do it. Her argument — that deciding what to say and how to say it should stay human because that’s where competitive differentiation lives — is one I think applies far beyond marketing. Her observation that guardrails are what actually liberate AI adoption, rather than constrain it, reframed a conversation I’ve been having with guests for two years.
The final theme from this year’s episodes is the one that will age fastest, because the landscape is moving so quickly that anything I write about specific tools or architectures may already be outdated by the time you read it. What I’ll focus on instead is the pattern.
In early 2026, the guests who were furthest along in AI deployment weren’t talking about AI as a technology layer sitting on top of existing processes. They were talking about agents as the primary unit of work — systems that operate proactively, in the background, across workflows that used to require sustained human attention. The shift isn’t from human work to AI assistance. It’s from human execution to human orchestration.
David Arnoux, co-founder of AI-native venture studio Humanoidz and community leader of the Gen AI Circle — a global network of nearly 500 heavy AI adopters — gave me the most honest map of where this is actually going. His four-stage framework — co-intelligence, augmentation, automation, redundancy — doesn’t flinch at the uncomfortable end of the spectrum. Some job categories, he said, simply won’t survive. Pretending otherwise isn’t optimism; it’s how you end up unprepared. His practical contribution was the concept of the “skills file” — a simple markdown document that turns any repeated workflow into a reusable, transferable, shareable asset — which struck me as one of the most accessible on-ramps to agent-powered work I’ve seen.
Lance Thompson at VIVI is deploying this in hospitality right now — voice AI agents handling reservations, room service, HR inquiries, and golf tee times across multiple languages in real time. His framing was one of the most human-centric I encountered: “We don’t want to replace Janet in Reservations — we want to scale her.” That distinction — amplification versus replacement — is doing a lot of work in every organization that’s getting this right.
Laura Maffucci at Globalization Partners is living it from the inside, piloting GIA — G-P’s agentic AI platform — as an internal HR agent handling employee inquiries, while simultaneously watching what it can do for clients at scale. Her candor about the perception gap between executives (who believe AI is driving efficiency) and employees (who often feel it’s adding to their workload) was one of the most grounded reality checks in any conversation this year.
What I take from these conversations collectively is that the window for “we’re still figuring it out together” is narrowing, as Meg Bear observed. The organizations that are ahead aren’t necessarily smarter or better resourced. They’re the ones that made a decision to act — and then kept adjusting based on what they learned.
I started Elevate Your AIQ because I was convinced that the human dimensions of AI transformation — readiness, responsibility, talent, ethics, culture, leadership — were being systematically underweighted in the public conversation about AI. The technology coverage was relentless. The human coverage was an afterthought.
Two years later, I’m more convinced of that than ever. And I’ve developed some sharper views from sitting across from over 120 people who are actually doing this work.
I believe trust is the real substrate of AI adoption — the condition under which everything else either works or doesn’t — and that most organizations are spending heavily on capability while underinvesting in the conditions that make capability usable and sustainable.
I believe the measurement gap and the courage gap are related, though not identical. You can’t honestly measure transformation without being willing to see what the numbers actually say — and then act on them, even when what they reveal is inconvenient. That second part is where courage comes in, and where most organizations stall.
I believe humans aren’t “in the loop” as a compliance formality. They’re in the loop because judgment, originality, passion, curiosity, and creativity are innately human contributions. The organizations that internalize this, rather than treating human oversight as friction until AI gets good enough, will build fundamentally different and more resilient solutions than the ones that don’t.
And I’ll admit something I think about more than I say publicly: I worry about the pace. Not the pace of AI development per se, but the pace at which the human infrastructure around it — the governance, the literacy, the cultural norms, the honest conversations — is keeping up. These recent conversations gave me real reason for optimism on that front. But optimism and complacency are not the same thing.
As I’ve said countless times, we are all responsible for Responsible AI, so we need to stay vigilant about doing the right thing and holding each other accountable. Which brings us full circle back to trust.
To every guest who joined me this year — Vijay, Adam, Lance, Tim, Dave, Stephen, Russ, Juan, Jonathan, Melissa, Jacob, Oded, Charlene, Paul, Jerry, Laura, Bob, Lisa, David, and Meg — thank you. These conversations are genuinely the best part of what I do. You brought your thinking, your candor, and in many cases your willingness to say things that don’t always get said in polished keynotes or press releases. That authenticity is key to Elevate Your AIQ.
I’ll also say this, because it’s true and I don’t say it enough: the timing of starting this show, while it might look obvious in retrospect, never felt inevitable to me. Everything I did before — the enterprise transformation and innovation work, the 25 years at IBM and NBCUniversal, the research, the perpetual curiosity about where technology and people intersect — it all pointed here eventually. I’m glad I followed where it led. And I’m glad you’re here with me for what comes next.
If any of this connects with challenges you’re working through, I’d love to hear from you. In the meantime, stay human-centric, keep learning, and innovate responsibly.
— Bob
Elevate Your AIQ is available on all major podcast platforms. Subscribe on Substack for written content, video clips, and upcoming live conversations. New episodes weekly.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.