What corporate innovation owes the person doing the work

The wall is covered in clusters of notes. A sponsor can stay for the afternoon. The person who runs Tuesday's queue is still on the floor, because the queue doesn't pause for a workshop.

That absence decides what the room can learn. The notes describe a process the sponsor recognises. They miss the workaround that clears the queue when the official steps jam. Corporate innovation teams repeat this scene often enough that the absence starts to look like part of the method. Nothing in the method says the operator should be elsewhere. The calendar is what put them there.

What human-centred design asks you to observe

The Nielsen Norman Group sets out design thinking as a hands-on, user-centred way of solving problems, and puts the work in three movements: understand, explore and materialise. Inside them sit six phases: empathise, define, ideate, prototype, test and implement.

Human-centred design in corporate innovation is that sequence run on a live workflow, with the person who does the work in the room. Empathise is the phase where you develop knowledge of what people do, say, think and feel. The instruction is ordinary and easy to skip. Talk to a range of actual users. Watch what they do. Ask what motivates them, what discourages them and where they get stuck. The goal is enough observation that the team can see the work from that person's side.

Define comes next. You pool the observations and look for the shared pain and the unmet need. Ideate is a wide pass at responses to that need, with quantity ahead of polish, and with people building on each other's sketches. None of this is a mural for its own sake. Each phase is supposed to leave an outcome strong enough to steer the next one.

Prototype, test and the phase teams drop

Prototype means making a subset of the ideas tactile enough to learn which parts hold. The point of that phase is weighing impact against feasibility, using feedback on something a person can react to. A wireframe, a revised script, a form the operator can try on a real case: the artefact can be rough. It has to be real enough that a flaw shows up in the room, before it turns up in a later steering pack.

Test sends that artefact back to the people who do the work. The questions are direct. Does this meet their needs? Has it improved how they feel, how they think or how they do the task? A team that stops at a demo hasn't asked those questions. It has shown a sponsor a picture.

Implement is the phase organisations forget, and I'd argue it's the one that matters most. The idea has to change the day of the people who use it. A concept aimed at the official process, and never put into the queue, solves the wrong problem and never reaches the work.

The phases are meant to loop. Teams return to empathise and define after a prototype, because only then can they see whether the solution works and what they still need to learn. A corporate innovation calendar that runs each phase once, in a straight line, and then files the deck, has used the names and dropped the loop.

When people won't describe the real work

Empathise fails in a particular way inside a large organisation. If describing a miss gets someone pulled up in front of their manager, the research records the official process. The workaround stays in the person's head, because saying it out loud carries a cost. Psychological safety, in this setting, is the condition that lets the operator describe the step they're uneasy about. The term is easy to print on a poster. The test is smaller. Can the person who runs the queue say, in front of the sponsor, where the documented steps are already a fiction?

Without that condition, human-centred design becomes a tour. The team is polite. The notes are full. The hours still sit in the step nobody wanted to admit. The same distortion hits AI training when the curriculum follows the documented process and the floor has already abandoned it. People practise a tool against a workflow that doesn't exist on Tuesday. They get faster at the fiction.

A leader who wants the real observation has to make the first honest description cheap. Thank the workaround when it appears. Ask where quality usually fails before you ask for ideas. Keep the operator in the room when the sponsor reacts, so the operator can see that the truth didn't cost them. That's a design choice about the session. It's also the only way the later phases have true material to work on.

What a portfolio owes the implement phase

Corporate innovation often keeps a portfolio of bets that stall at ideate. Each bet has a title, a sponsor and a slide. Few of them have an owner for the change in the live work, a date by which the operator will have tried the new step and a line for what will be stopped if the trial is wrong. For years design was a coat of finish at the end of product development, applied after the real decisions were made. The companies that got value moved that work to the start and let what people needed shape the offer. A shared sequence of phases helps a large organisation, but it won't do the implement phase for you.

A portfolio review that can't say which bets reached implement is a review of intentions. Ask, for each live bet, who was observed, what they were seen doing, what was made tactile, what the operator said when they tried it and what changed in the queue afterwards. A bet that can't answer the last of those hasn't left the workshop. Park it or kill it.

I'd rather see one bet reach implement than ten sit in discovery. Pick the workflow where the operator is most frustrated, because that person will tell you where the documented steps break. Their frustration is the best research brief you'll get, and it costs nothing to ask for it.

What to run on one workflow

1. Put the operator of one workflow in the room with someone who can change a handoff, and spend the opening of the session on what that operator actually does when the official steps jam.
2. Define a single unmet need from that observation, then make one rough artefact the operator can run on a real case beside the current way.
3. Decide in the same week to adopt, adapt or stop, and name the person who owns the changed step after the room empties.
4. If people won't describe the workaround, fix that reply from the sponsor before you book another session. More notes won't repair a room that punishes the truth.

A design thinking workshop can hold that week for an intact team. You'll still need an owner for implement once the sponsor leaves, and you'll need the person who does the work to say what it really involves. Innovation, digital and L&D leaders can ask for both on one workflow this month. If the answer is a wall of notes and an unchanged queue, the programme hasn't reached implement yet.