Stop Talking About AI adoption or transformation. Start Running These Five Motions.
I sit on a handful of boards, and the spread between companies treating AI as operating-critical and those treating it as a side project is widening every month.
The good news: the difference isn't strategy, vision, or budget. It's a small set of repeatable motions that leading companies run every week.
Here are the five I'm pushing hardest on right now.
1. Demo Days
The best operators in my portfolio are running internal demo days — short, focused sessions where teams show the AI tools, prototypes, or workflows they've actually built. No over-polishing. No theater. Just real examples of AI driving leverage.
These sessions cut through the noise. They spark new ideas, surface hidden experts, and make it easier for teams to copy what already works. Seeing a peer demo an automated workflow is infinitely more compelling than reading a memo about "transformational AI opportunities."
If your company isn't doing this yet, start. Thirty minutes, three demos with Q&A for each. Add-in the list of tools. That's the entire format.
2. Board-Ready AI Updates
This is one I'm pushing from my own seat, leading on Lev, but on my boards too. The boards I'm on now get a concise quarterly AI update from the CEO:
- What we automated last quarter (with AI)
- What's planned for the current quarter
- Where we need resourcing, data access, or technical support, connections to other founders ahead of us
The discipline forces clarity. It also signals — internally and externally — that AI is not a science project. It's operating-critical. When the board tracks AI progress the same way it tracks revenue or product velocity, company behavior changes.
If you're a CEO and your board isn't asking for this, volunteer it. If you're a board member and you're not asking for it, start at your next meeting. Zero Health and their CTO, Greg Inman are doing an awesome job of this.
3. Evaluating AI-First Startups
The companies pulling ahead are getting sharp at evaluating early-stage, AI-native teams. Two reasons this matters:
- These teams often have capabilities you can partner with — or buy — immediately.
- The evaluation process itself keeps you calibrated on models, infrastructure trends, and emerging use cases.
The teams building from zero with AI are fast. They're scrappy, opinionated, and usually see around corners. Understanding how they think helps you build better systems internally.
The companies doing this best are running a rolling 30-day cadence: try one, swap one, kill one. Last month we tried X, switched out Y, canceled Z. Posthog for product and usage analytics has been one of these kinds of products for me and many of our portfolio. I highly recommend it! It's not a prt's not a procurement process. It's a learning loop.
4. Peer Group Calls
We've been running these with the CEOs of our PSL portfolio companies, and they're some of the highest-signal hours on the calendar.
The format is simple. A recurring call with a small group of CEOs from like-minded companies. Everyone shares what they're testing, what tools they've found, what worked, and what got killed. And they bring real questions to the group — the ones they don't want to ask in a board meeting yet.
The leverage shows up fast. You compress months of solo trial-and-error into an hour because someone else already ran the experiment. Tool finds, vendor warnings, prompt techniques, internal rollout playbooks — it all spreads through the group faster than any newsletter or analyst report could.
You don't need to be in a portfolio to run this. Find five or six CEOs or CTOs you trust who are equally serious about AI, get on a recurring monthly call, and commit to actually sharing what's working — including the parts that didn't. The group that shares most candidly wins the most.
5. Speaker Lunches
We're lucky at PSL. We've had speakers from OpenAI, Stripe, Microsoft, Google, AWS, to name a few. Most companies can't get those people in the room, and I want to be honest about that.
But the motion still works without that access. The point isn't the famous speaker. The point is pulling someone who lives one or two layers deeper than your team into the room for an hour. That bench is wider than you think:
- Vendor founders and senior engineers — most will jump at a 45-minute lunch with a real customer
- Practitioners at non-competing companies running the same playbook in their org
- Local academics, AI/ML meetup organizers, and applied research folks
- Customers using AI in interesting ways inside their own businesses
A one-hour conversational lunch with someone two steps ahead can replace weeks of speculation. These sessions also let people feel how fast this is actually moving — when someone says, "Here's what'll be possible in six months," they start planning differently. Cheap, fast, high-signal. There's no reason not to be running one a month.
None of this is a transformation program. It's a set of habits.
The companies winning at AI right now aren't winning because they have a better strategy deck. They're winning because they're running these motions every week while everyone else is still in the planning phase.
If you're on a board, push for them. If you're running the company, install them. The compounding starts immediately.