The All human
AI Journey
From curious to orchestrator, we are all somewhere on the AI journey.
AI is already changing what's possible in almost every role at All human. It is the most significant shift in how we work in a generation, and increasingly the difference between good work and exceptional work comes down to how well we use it. That makes building the skill worth real attention, for your own craft and for where the company is heading.
This is already underway across All human.
Designers are building HTML prototypes and testing real interactions, not just wiring up screens in Figma. Developers are building things that were out of reach only a few months ago, and in a fraction of the time. QA are running an agentic test council, and PMs are planning and reporting with it.
5 stages.
From first curiosity to directing AI at scale. Each stage builds on the one before.
Where are you on the journey?
5 stages run from curious to orchestrator. Find where you are today, and what your next step looks like.
Curious
You've watched it work for others, but haven't put it to work yourself yet.
Set up Claude as a work assistant living in your project, and give it permanent ownership of a task you would usually do yourself.
Adopter
You're trying it on real tasks and starting to get a feel for what it can do.
Give AI ownership of multiple tasks at once, and let it automate an entire routine.
Operator
It's become routine, a normal part of how you work day to day.
Build out skills, not just to automate single tasks, but so you can reuse the same workflows across multiple projects.
Builder
You're creating prompts, tools and workflows that others can pick up and reuse.
Run several agents at once across multiple workstreams and projects, and share the skills you build so the rest of the team can use them too.
Orchestrator
You're directing several agents at once, and setting patterns others follow.
You're at the leading edge now. The question is how far you can take it: push what's possible, and boldly go where no one has gone before.
What it looks like for you.
AI lets you spend your time on decisions and people, not document wrangling.
- Draft backlogs, tickets, release notes and status from real project context.
- Plan against capacity with AI doing the heavy lifting.
- Turn meeting notes into actions, summaries and follow-ups automatically.
- Spot risks, blockers and slippage early by having AI watch the project data.
AI lets you turn raw requirements into clear, testable specifications far faster.
- Draft user stories and acceptance criteria straight from meeting notes and context.
- Interrogate data and documents to surface the real requirements.
- Map and document processes, then keep them current as things change.
- Spot gaps and edge cases early, with AI pressure-testing the brief.
AI lets you test real journeys in working prototypes, not just static flows.
- Stand up interactive prototypes in the browser and test real interactions.
- Generate flows and wireframes directly from a brief or a problem statement.
- Synthesise research and feedback into clear insight far faster.
- Produce production-ready components, not just visuals.
- Explore several directions in parallel before committing to one.
AI lets you design in real, working code, not just static screens, and at a pace that changes what you can explore.
- Turn Figma designs into working front-end through Code Connect and MCP.
- Build interactive HTML prototypes instead of static mock-ups.
- Generate and iterate UI directly from a brief or a reference.
- Keep a design system consistent, with AI applying it across every screen.
AI lets you build at a pace that was not possible before, with you owning the architecture and the agents doing the typing.
- Pair-program in Claude Code and Cursor as your default way of building.
- Generate components straight from designs via Figma MCP.
- Stand up whole products fast (AhCMS went from nothing to deployed in four days).
- Keep ownership of the architecture and the decisions that matter.
AI lets you run quality at the scale of the whole product, with you owning the sign-off.
- Run an agentic test council (AhTesthub) that generates cases and runs regression in parallel.
- Have AI triage failures and surface what matters, with a human signing off.
- Move toward agents finding, fixing and raising PRs for bugs on their own, as on FITRADE.
AI lets you build the capability to win visibility as search shifts from search engines to answer engines.
- Build out GEO and AEO as a service the company can offer, alongside SEO.
- Optimise content and structure to be cited by AI answer engines, not just ranked.
- Research keywords, briefs and competitor gaps in a fraction of the time.
- Stay ahead of how AI search ranks and surfaces content as it evolves.
AI lets you go from data to decision faster, and act on what you find.
- Mine analytics for patterns far faster than by hand.
- Draft and rank the hypotheses worth testing.
- Run experiments end to end, from idea to readout.
- Automate recurring reporting and dashboards end to end.
What will you try next?
Our industry is changing faster than any of us have seen before, and AI adoption is how we stay ahead. The 5 stages of the journey are a way to illustrate where you are and how to advance, so we can continue our success as digital market leaders. We can all read about how to use AI, but it really only clicks when it lives in your projects, when you use it on your own tasks, in your own way. What it does for you will not look the same as it does for anyone else. Find your stage above, and see how to reach the next step.
