top of page

All Posts


ree

This one’s for an audience that’s probably about three people wide, but here we are. I feel very moved to talk about how seeing an AI Consultant on Big Brother hit me in a specific way.


I’ve watched the show for more than a decade. I once threw myself a full Big Brother birthday weekend: competitions, secret meetings, a makeshift memory wall (fine, that was this year, when the premiere landed on my birthday, but whatever). I have a BB tattoo. I met one of my closest friends after bringing up a BB podcast at a music festival. I once sat next to Cory Wurtenberger’s brother in an Uber and spent the entire ride pretending to be so chill.


Ten years ago I watched Da’Vonne Rogers and Jason Roy on the live feeds. Their friendship was so easy it made me feel like I was in the room. I haven’t had that kind of parasocial hit since.


Every summer my schedule tilts around the show - episodes, podcasts, feeds. I’ll talk for hours about the reality of the season versus the edit. I’ll spend the warmer months debating best and worst moments like it’s an actual sport. I have strong opinions about production choices, season themes, and casting “flops” who deserve a second chance.


All of that to say, I really love Big Brother. But let’s get to the point.


Last year I added AI consulting to my work. The tools help, sure, but most of it’s just paying attention. You hear what people say they want, you watch what they actually do, and you notice the gap between the two. Then you figure out how to close it without breaking the rest of the system.


So when Jimmy introduced himself in the BB27 preseason as an AI Consultant, I flinched. Not because I doubted him, but it’s weird seeing your relatively new job title on TV. Hearing my favorite podcasters speculate what it actually means, knowing in the back of my mind that day-to-day his job is basically training for this game. Read the room, track patterns, act at the right time. I wanted to cheer but mostly I just sank lower into the couch cushions.


Then I watched him play.


Week two the house went dark - literally. Jimmy won the Black Box HOH, a pitch-black maze where most players stayed locked on their own puzzle task. He not only navigated it cleanly, he clocked Kelley pocketing a power in the dark. That’s high-level awareness: catching the quiet move while you’re still in motion.


With power in hand he targeted the solid early threat in Keanu. Spotting comp potential and momentum was a good read, even as the week was swallowed by extra powers. When Keanu and Kelley both saved themselves, the week turned into a minefield. Twist powers forced two renoms, meaning Jimmy ultimately put five different people up for eviction during his HOH reign. He pivoted to a safer choice of nominating Amy and Will - a decision that frustrated some of his alliance but kept him in line with broader house consensus. It wasn’t the biggest move, but it was one that kept his threat level low in a week where overreach could have blown up his game.


Good consultants read the flow of information before acting on it. From the start, Jimmy positioned himself where those currents moved. In week one, he linked up with Mickey and Morgan to form Triple Threat - the first named alliance of the season - and built lines into a larger voting group. He also built a genuine connection with Rachel, giving him a personal link outside the core that could feed intel and offer cover if the main alliance fractured. That’s network design, not noise, and it’s the kind of foundation that can carry you deep.


By week three, the house was still wobbling from the early-game chaos. Triple Threat was showing cracks: Mickey wasn’t fully looping Jimmy in, Morgan was exploring side channels, and the larger voting bloc he’d tapped in week one hadn’t solidified. Bonds like his with Rachel were genuine but untested under real vote pressure. And players were already floating aggressive moves before jury, chasing short-term momentum in ways that risked alienating multiple factions. In that climate it was easy for the wrong read to take hold.


Mickey gained control via the last twist on the board, mistook Jimmy’s visibility for disloyalty, and cut him - the kind of misread that tanks a strategy before it has a chance to pay off. He was evicted this past Thursday. The irony: he was one of her most loyal allies, and removing him was a short-term win for her but likely a long-term loss.


Jimmy left with the right reads, the right allies, and the ability to spot a hidden power in the dark. You can make the right read and still lose to timing, twists, and feelings. That’s not an excuse; it’s the job. Strategy lives in the gap between what should happen and what people actually do.


Seeing “AI Consultant” in a house built on chaos didn’t make the work feel loud. It made it feel accurate. The edge isn’t the tools; it’s knowing when to speak, when to wait, and spotting the one thing no one else in the dark is paying attention to.


I started the preseason wary, but from his intro package on night one I was in Jimmy’s corner. He had the reads, the measured play, and the network instincts that could have carried him deep. It’s such a loss to see him cut so early. Not because of bad reads but because of timing, paranoia, and a twist-heavy board. The season is a little less sharp (and a lot less fun) without him in it. I’ll be watching for the next time an AI Consultant ends up on a social strategy reality show. I don’t think I’ll have to wait long.


Jimmy didn’t get the ending he wanted. But the way he played - steady, directional, and early to structure - looked a lot like real life to me.


Maybe we’re all just trying to read the room a little better.

 
Elegant room with green walls and gold accents. A chair and table with a laptop sit below a glowing wall square. Sunlight casts shadows.

Why the best systems get built in a mess.

It’s easy to imagine what building strong systems looks like: a clear runway, a calm team, the luxury of time to map and iterate. That’s the fantasy. And for most founders? That’s never how it starts. Instead systems get built in motion. Processes mapped out mid-launch. Team check-ins scheduled post-burnout. Everything on the heels of a cash flow panic or a team shakeup. Not because it’s ideal, but because it’s necessary.

And here’s the part that doesn’t get said often enough: that kind of systems work - the duct-taped, decision-fatigued, just-in-time kind - is still deeply valid. Strategic, even. Not just a placeholder until something better comes along.


The Fantasy vs. The Reality

The idealized version of operations work is clean and color-coded. SOPs built before the first client. Airtable bases that anticipate everything. But most small businesses don’t build that way.

Only 49% of U.S. companies report having a formal crisis plan. And among those fewer than a quarter actively practice it. That gap between theory and readiness? It shows up fast when pressure hits.

And under pressure perfectionism doesn’t help you or your team. Research shows self-oriented perfectionism has risen significantly over the last few decades - and so has burnout. Perfectionist expectations might look like high standards, but in a crisis, they stall momentum.


How to Build Systems in Business (When You’re in a Crisis)

Real systems built mid-crisis aren’t polished. They’re functional. Often held together by Google Sheets, Slack threads, and institutional memory. They prioritize what matters now, not someday.

Frameworks like triage models (red/yellow/green) or the Incident Command System come from emergency response, and they work. In a small business that might mean shifting launches based on cash flow realities or tagging team updates by urgency. It’s not glamorous, but it keeps things moving forward.

Minimum viable systems follow the same logic. Start with the core functionality. Deliver stability fast. Iterate when things calm down.


The Weight of Building While Holding Everything

Crisis mode doesn’t just drain capacity - it messes with clarity. Decision fatigue creeps in. Strategic thinking narrows. You start reacting, not building.

Research backs this up: under high cognitive load, decision quality drops. People default to what’s easiest, not what’s best. That’s not a personal failing. It’s a brain thing.

This is where light structure helps. Tools like OKRs or even simple “Go/No-Go/Wait” choices reduce load and preserve capacity for the decisions that actually matter.


Trade-offs Are the Strategy

In a perfect world you’d fix everything, but mid-crisis, trade-offs are the strategy. You choose what to stabilize first. You let nonessentials slide.

This isn’t about settling. It’s about sequencing. MVP thinking applies to backend systems too: what’s the smallest version of this process that can work right now?

And more importantly: what’s draining you that doesn’t need to?


Building Adaptive (Not Perfect) Systems

The goal isn’t to finish. It’s to keep evolving. Adaptive systems respond to pressure. They shift with you. That’s resilience.

Chaos engineering (used by companies like Netflix) tests systems by breaking them on purpose to see where they hold and where they don’t. Not to punish, but to learn. In a small team this might look like stress-testing your client onboarding by having someone new follow the steps without help. You’ll learn a lot, fast.

You don’t need to simulate failure. You’re living it. But you can treat this phase as data. Every workaround, every breakdown is information you can build from.


A Simple Framework for Crisis System Building

Red (Critical) - What breaks the business if left unresolved? Fix it now.

Yellow (Urgent) - What’s causing friction but still functioning? Triage this next.

Green (Supportive) - What can wait? Schedule or shelve it until capacity opens up.


Pair that with a few mental load reducers:

  • Daily decisions? Standardize what you can.

  • Priorities? Name the top three and let the rest be background noise.

  • Tools? Use what’s already working. Skip the shiny new system for now.


If you’re building systems mid-crisis, you’re not behind. You’re just building in real life, with real constraints. And what you build now might not be pretty. But it will be lived in. Yours. Strong in ways that polished systems rarely are.


Let it be messy. Let it hold. You can refine later. For now you’re doing the real work.



Sources:

 
Man at desk in room with "AI READY" sign. The sign is only half lit, with the word "READY" not working. Cursor points to an empty "Enter prompt:" text box. Papers and pen on table. Retro, focused mood.

Why enthusiasm isn’t enough in the age of intelligent tools


Last week I sat in on a team-wide AI conversation. It had good bones - curious energy, practical takeaways, people sharing tools and prompts like trade secrets.

But the tension was familiar: the founders were excited… and quiet.

They weren’t leading the learning. Not intentionally. Just by omission.

They applauded the team’s experiments (automated research agents, custom GPTs, smart prompts tuned for deal flow) but didn’t push deeper. No “could this replace our current process?” or “what would it take to scale that?” They showed up as supporters, not stewards.

And that’s the part that stuck.

Because if you’re at the top of the org chart and you’re not actively shaping how AI is understood, used, and evaluated you’re not leading the change. You’re just riding it.

A recent MIT study backs this up: when people used ChatGPT to draft essays from the start, their originality and neural engagement declined. They got results but lost something essential in the process. Without that upfront engagement, they missed the chance to define the problem, challenge assumptions, and direct the output with clarity.


Excitement ≠ Readiness


A recent MIT study found that participants who engaged with AI after doing the work themselves showed increased neural activity in areas tied to memory and problem-solving. In contrast, those who relied on AI from the outset saw a drop in originality and long-term retention. The takeaway? Thoughtful use isn’t just strategic. It sharpens your thinking.

Greg Shove, CEO of Section, names this plainly: you’re either an AI freeloader or an AI manager.

Freeloaders offload to AI and copy-paste the result. Managers use AI to think deeper, explore blind spots in decisions, and raise the bar.

And for founders, that distinction matters more. Because if you’re not fluent in the tools your team is adopting you can’t ask the right questions. Can’t spot the strategic gaps. Can’t model responsible use.

You end up with innovation that lives in pockets - not products.


A Simple Case: “Will this scale?”


One founder I worked with recently started using GPT to run through five different onboarding workflows. Not to write SOPs, but to interrogate them:

  • Where are we losing clarity?

  • What steps could be automated?

  • What’s redundant if we’re using tools like this daily?

It wasn’t flashy. No custom agents or advanced scripting. Just better questions, faster iteration, and the courage to rebuild systems from that clarity.

That shift freed up ~6 hours/week across a lean ops team.

And it came from the founder’s willingness to go first, not just delegate improvement down the line.


Where to Start (Without Overhauling Your Life)


If you’re wondering, What does “good enough” even look like for me? here’s the short version:

  • Use the tools daily - in your own workflow, not just to “support” the team.

  • Pressure-test decisions - run strategy drafts or investment questions through through your preferred LLM (ChatGPT, Claude, Gemini) and see what gets surfaced.

  • Audit one process per month - ask: could this be automated, augmented, or made clearer with AI?

  • Build a tiny system - even if it’s just an SOP writer or email tone checker. Not because it’s efficient, but because it teaches you how the pieces work.

  • Write your own AI manifesto - articulate what responsible, strategic AI use looks like for your company, in your language.

These aren’t projects. They’re practices.


And what isn’t your job?


You don’t need to become your own AI engineer.

You don’t need to build every workflow from scratch.

And you don’t need to be the most advanced user on the team.

Your job is to stay close enough to the tools that you can:

  • Spot the difference between novelty and leverage

  • Ask better questions about systems, not just outputs

  • Know when it’s time to bring in support - and what kind

The deeper builds, custom integrations, and team-wide enablement? That can come later, with the right people in place.

But fluency has to come first. Without it, you’re not delegating. You’re just handing off decisions you don’t yet understand.


The Risk of Staying Shallow


Shallow use has real cost. MIT researchers found that over time, habitual reliance on AI led to more formulaic thinking. That loss of creative engagement showed up in how little ownership people felt over the work itself.

When leaders don’t get hands-on, a few things happen:

  • Bright spots never turn into strategy.

  • Investment gets guided by flash, not function.

  • Culture flattens - because no one is modeling depth.

You don’t need to master every tool. But you do need to be in relationship with them. Otherwise, you’re asking your team to steer the ship without a shared map.


Bottom Line


You don’t need a Head of AI.

You need to stop outsourcing your AI fluency.

The companies that will thrive in this next chapter aren’t the ones with the best stack. They’re the ones where leadership knows how to use it: slowly, imperfectly, but with intention.

So the better question isn’t “are we using AI?” it’s “are we learning how to lead with it?”



Sources:

-Greg Shove, Section.AI CEO. Leaking Our Own AI Manifesto 

 
bottom of page