The Hop Labs 8

An AI Assessment Framework

The Hop Labs 8 scores your organization on the eight things that decide whether AI work ships, compounds, and thrives alongside the people who built it. It's based on a decade of Hop Labs' engagements with life sciences, deep tech, and platform teams, and calibrated against what we've seen work and fail inside real organizations.

Who this is for

You run or sponsor AI inside an organization that takes technology seriously: you have access to some researchers, models, maybe even a platform team. You might be a Director or VP of AI, a Head of Data Science, a Head of Platform, or a CTO at a smaller biotech.

Crucially, you've begun asking yourself the question every one of our clients eventually poses:

"Where are we compared to everybody else? We don't want to fall behind."

But comparing yourself to everybody else is the wrong anchor; in our experience, many of your peers are struggling with the same obstacles.

Instead, the Hop Labs AI Assessment Framework scores you against your potential: what your organization could be doing with your people, your data, and your challenges.

A note: Companies with thousand-person platform teams hold no monopoly on the top of that scale; a small team with a clear point of view about AI can reach it just as easily.

The eight dimensions

The framework measures two things about your organization, through eight dimensions.

Your point of view

Do you know what AI is for here, and is the org built to act on that knowledge?

  1. Theory of value. Where does AI create the most value for your business — and would a researcher, an engineer, and a leader give the same answer?
  2. Product orientation. Is someone's actual job the problem definition — what success looks like, who the user is, what to cut? Or is it research with no customer?
  3. Expert access. Who labels the things only your domain experts can label, and how much of their week does your AI team realistically get?

Your operational maturity

Is the machinery in place for AI work to ship, repeat, and survive?

  1. Top-of-license. Are your researchers doing research, or are they doing IT?
  2. Velocity. Are you faster this year than last — and can you point to why?
  3. Feedback loop. If a training label is wrong today, how long until the fix is live? (No customer-facing product? Read "live" as wherever your model's output gets used for real — the analysis someone trusts, the prediction someone acts on.)
  4. Ownership. Can you name the person who owns data quality? Model quality? The system that tells you a model is misbehaving?
  5. Durability. How many departures would it take to meaningfully derail your AI goals?

Scoring

Each dimension scores 1–3 from your answers.

The first three sum to your point-of-view score; the other five sum to your operational maturity score. Each axis lands Low, Medium, or High, and the two together place you on the map below.