Andrew Crossley
AI PRODUCT MANAGER · UK & REMOTE

AI Product Manager UK

An AI Product Manager owns the customer outcome and the behaviour of the AI system behind it. I work with founders and product teams to define the opportunity, scope the product, create evaluation criteria, manage model trade-offs and failure states, ship the roadmap and connect AI quality to measurable customer and commercial results.

Contract · Fractional · Advisory · UK based · Remote worldwide

WHEN COMPANIES BRING ME IN

The AI works. The product still needs an owner.

The problem is rarely a lack of AI ideas. It is usually unclear ownership between customer need, product judgement and technical capability.

Your engineers can build the AI feature, but nobody owns whether customers actually want it.

You have a prototype, but no evaluation framework for deciding whether the AI is good enough to launch.

Product and engineering disagree about what a successful AI output looks like.

Your roadmap is a list of AI features rather than customer outcomes and commercial priorities.

You need senior AI product ownership, but you are not ready for a permanent hire.

The product works in a demo, but cost, latency, failure handling and reliability are still unclear.

WHAT I OWN

From customer problem to launched AI product

The role is not prompting. It is making the product decisions that determine whether an AI capability becomes something customers trust, use and pay for.

Customer problem & strategy

Find the customer problem where AI creates enough value to justify the complexity, then define the product outcome and scope.

MVP scope & prototyping

Turn the strategy into one complete user journey and use working prototypes to improve the quality of discovery before expensive build work.

Evaluation & success metrics

Define what good looks like, create evaluation criteria and connect AI quality to customer behaviour, retention and commercial outcomes.

Failure states & guardrails

Design confidence thresholds, human fallback, permissions and recovery paths so the product can survive real-world use.

Model economics

Balance output quality against latency and cost per successful task instead of choosing models on benchmark scores alone.

Delivery & launch ownership

Prioritise the roadmap, align product and engineering, ship the product and use customer evidence to decide what happens next.

I use the Crossley Method to move from Discover and Validate through Prototype, MVP, Launch and First Revenue without confusing speed with evidence.

AI PM vs TRADITIONAL PM

The core job stays the same. The evidence changes.

TRADITIONAL PRODUCT MANAGEMENT

  • Discovery relies mainly on interviews and static mocks
  • Analytics focuses on funnels and feature usage
  • Engineering owns most implementation trade-offs
  • QA is largely deterministic

AI PRODUCT MANAGEMENT

  • Discovery combines interviews with working AI prototypes
  • Analytics also measures output quality, failure rate and intervention
  • The PM needs fluency in model, latency, quality and cost trade-offs
  • Evaluation must account for non-deterministic outputs and regression

For the deeper role breakdown, read what an AI Product Manager actually does.

WAYS TO WORK WITH ME

Choose the level of ownership you need

The engagement model should follow the problem. I do not force every company into a retainer if a shorter intervention is enough.

Embedded · 3-12 months

Contract

Full-time product ownership for a specific AI product, launch or transformation when you need an experienced PM inside the team.

1-2 days per week

Fractional

Ongoing AI product leadership for founders and teams that need senior ownership without a full-time hire.

Focused senior input

Advisory

Roadmap, evaluation, model choice, product strategy and launch decisions for teams that already have day-to-day product ownership.

AI Product Manager or AI Product Consultant?

Choose an AI Product Consultant when you need senior diagnosis, strategy or a defined product decision. Choose an AI Product Manager when you need ongoing ownership of the roadmap, evaluation, delivery and outcome. If the need is company-wide product leadership, a Fractional CPO is the better fit.

PROOF

Products I have built and led

My product judgement comes from operating products, not only advising on them.

Wocal

Founded and led a SaaS platform from 2020 to 2025, scaling to 300+ venues with a reported £2.7M pre-money valuation.

Co-Ride

Founder and CPO of a community carpooling platform, giving me direct experience of early-stage product, marketplace and operational decisions.

AI-first product work

Hands-on work across AI prototypes, evaluation, production readiness, workflow design and commercial launch through founder products and client engagements.

FREQUENTLY ASKED

AI Product Manager questions

What does an AI Product Manager do?
An AI Product Manager owns the customer outcome and the behaviour of the AI system behind it. The role covers discovery, product strategy, prioritisation, evaluation, model trade-offs, failure handling, analytics, launch and commercial outcomes. The difference from a traditional PM is that AI quality, uncertainty, latency and cost are now product decisions, not purely engineering decisions.
When should I hire an AI Product Manager?
Hire one when AI is central to the product and somebody needs to own more than delivery. Typical signals are a prototype with no evaluation plan, engineers making product decisions by default, unclear AI success metrics, repeated model changes without customer evidence, or a founder who has become the bottleneck between customer, product and engineering.
What is the difference between an AI Product Manager and an AI Product Consultant?
An AI Product Consultant is usually brought in for a defined problem, diagnosis or strategy decision. An AI Product Manager is closer to the ongoing product owner: prioritising the roadmap, aligning the team, defining evaluation, shipping releases and staying accountable for the outcome over time. I work in both modes depending on what the team actually needs.
What is the difference between an AI Product Manager and a Fractional CPO?
An AI Product Manager owns a product or product area. A Fractional CPO owns the wider product function: company-level product strategy, portfolio decisions, team structure, hiring, operating cadence and often board-level product communication. Early-stage companies sometimes need one person to cover both, but the accountability level is different.
Do AI Product Managers need to code?
They do not need to be software engineers, but they should be technically fluent enough to prototype, understand model limitations, reason about APIs and data flows, and discuss quality, latency and cost with engineering. AI-assisted development has made hands-on prototyping a useful product skill because it shortens the distance between an idea and real customer evidence.

Need someone to own the AI product, not just advise on it?

Tell me what is built, where the team is stuck and what outcome matters. I will tell you whether the right shape is contract, fractional, advisory or a shorter consulting engagement.