OP #315: Human-Agent Teams, 235 hours, and looking ahead to 2026

The last OP of 2025... crazy how fast time flies!

Author

Darren Herman
December 23, 2025

Happy (early) holidays to all. My wife and I started a tradition of taking the family away for Thanksgiving which is a mental early start to the holiday season. I like it a lot and extends the festive period. It’s flown by and now we’re in the thick of it. For all of you who celebrate anything and everything, have a wonderful holiday and New Year.

Looking back on 2025, it was a transformational year for me. 235 hours of tracked time in the gym - and a huge thank you to my “team” for supporting, especially Liraz, Eric and Ryan. I dropped a lot of weight, reversed negative health results, got my mile run average down materially, and learned to like rowing (well, tolerate). Outside the gym, I made myself uncomfortable with posting reels, got back into photography, built a global network of horological friends, and have tried to spend extra time with my son who will be heading to college after he graduates HS in the Spring. I became a bit more selfish with my time and I believe I’m better for it. I feel better all around.

I feel that 2026 will be transformational OR a huge bust but being the ever optimist myself, I’ll bet on transformational. Why? Well, I’ll leave that to my writing below on the HAT - human-agent teams. I am also excited about the Silicon Alley Sports season that’s ahead of us… It’s going to be awesome and will start talking about it in mid-January. And of course, the New York Knicks… who cannot be excited about this season?

OK, enough reflections and such. Lets get this OP started.

Here’s a few fun links:

The next OP will come after the New Year - so thank YOU for subscribing and being part of the community. I answer every single email that comes in from OP subscribers so please never be a stranger.

Be well, do good.
Darren

From Point Solutions to Human–Agent Teams

We are finally moving past the "magic trick" phase of AI in marketing.

For the last two years, we’ve been drowning in point solutions. Everyone has a tool to write copy faster, generate images cheaper, or optimize a campaign by 1%. These are useful toys, but they are incomplete. They don't solve the core constraint of a growth organization: infrastructure.

The next evolution isn't about better prompts; it’s about org design.

Right now, marketing orgs are built to orchestrate humans. We take strategy, break it into tasks, and hire managers to coordinate the execution. Progress is governed by meetings, approvals, and sheer brute force. That management layer is the "orchestration layer"—we just don’t call it that.

But when the machine starts doing the actual execution, the org chart breaks.

The question for 2026 isn't "which tool do I buy?" It’s "how do I manage a workforce that is half silicon?"

This requires a shift toward Human–Agent Teams (HATs). In this model, humans retreat to the high ground: strategy, taste, narrative, and judgment. Agents handle the execution, iteration, and scale. The "manager" role shifts from managing people to orchestrating a system of humans and bots working as a single unit.

If you get this right, the leverage is massive. But the HAT model is uncharted territory, and as operators, we need to be honest about the "watch outs" before we start rewriting job descriptions.

Here are the five questions I’m asking growth leaders right now:

1. Where does the buck stop? When an agent executes the work and a human supervises the outcome, ownership gets blurry. If an agent hallucinates a discount or offends a customer segment, who is responsible? We need new accountability models for revenue and brand risk.

2. What is the new span of control? If one manager can supervise ten humans today, can they supervise fifty agents tomorrow? A hundred? We don’t know where cognitive overload hits, or where quality control falls off a cliff.

3. Can you automate "taste"? Agents are incredible at iteration. They are terrible at taste. Knowing when to intervene—and having the confidence to say "this is technically correct but emotionally dead"—will be the defining skill for future creative directors.

4. Who are we hiring? The highest-leverage marketers in a HAT-based org won't look like traditional executors. They will look like editors, systems architects, and orchestrators. Career paths and incentive structures need to be totally rewired.

5. What is actually proprietary? In a world where every competitor accesses the same LLMs, your durable advantage won't come from the tool. It will come from your proprietary data and how effectively your HATs are designed and integrated.

We are entering a period of rapid experimentation. The "point solution" phase was inevitable, but it was just the warm-up. The real alpha will be generated by the companies that figure out how to weave humans and agents into a single, cohesive operating system.

The map isn't fully drawn yet, but the race has already started.

Below are a few articles I came across this past week that I found interesting. While I may not agree with everything in each one, I think they're worth a read. If you stumble upon an article you think I or the Operating Partner community would enjoy, feel free to share it with me. Of course, I reserve the right to decide what gets featured in the OP.

Making Advisors (Fred Wilson)

Thank you again for subscribing to the OP! See you in 2026…

What did you think of today's OP letter?

Login or Subscribe to participate in polls.