The handoff is no longer a Figma file

In conversations with other designers, I keep hearing a similar concern: the process feels different now. Sometimes it feels like there is no process at all.

I recognize that feeling. My work is more scattered than it used to be. I can start in Figma, try something in Claude or directly in code, discover a constraint while building, change the interaction, and end up with useful thinking spread across a design file, a branch, screenshots, conversations, and a pull request.

Then comes the awkward part: handoff.

For a long time, the answer was obvious. Clean up the Figma file, make sure the flows and states are there, add whatever annotations are needed, and share it. One place. Done.

That model does not fit all of my work anymore. Trying to force everything back into Figma just to make the handoff look complete started to feel artificial.

The handoff has moved

Figma is still important. I use it for flows, visual systems, components, and for working through ideas with other designers. I am not trying to replace it.

But some things become clearer somewhere else.

I might try an interaction in code because I want to feel it rather than look at frames. While doing that I might hit a technical constraint that changes the design. The branch now contains something the Figma file does not. A conversation might contain the reason I dropped an earlier direction. The PR might show the version that actually works against the product.

That is fine. The problem is expecting the next person to reconstruct the decisions from all of it.

Sending someone a collection of links is not the same as handing over a decision.

Which thing is current? Which one was just an experiment? Did the Figma version change after I tried it in code? Is the PR ready to review, or am I using it as a prototype? Those details are obvious to me because I was there. They are not obvious to the next person.

What I started doing instead

I started creating a lightweight page around the work.

Nothing sophisticated. It can be a Confluence or Atlassian page, a Google Doc, or whatever shared space the team already uses. I think of it more like a folder with an explanation than a new specification.

I am not copying every artifact into it. I am connecting them.

Usually I put the problem near the top, then the explorations that actually mattered and the direction we chose. If something changed while I was building it, I write that down. Otherwise somebody has to compare two artifacts and guess which one won.

If there is a working branch, preview, prototype, or PR, I link it. For interaction-heavy work I would rather have someone use the thing than interpret another set of static screens. I also call out what is still open so people know where feedback is useful and what has already been decided.

That is usually enough.

If the visual system is best understood in Figma, go there. If the interaction needs to be experienced, open the preview. If a technical discussion changed the direction, link to it and explain why it matters. The page is there so I do not have to pretend one of those artifacts contains the whole story.

The PR is one artifact, not the handoff

At one point I thought the pull request might be the place where all of this comes together. I understand why. A PR has the implementation, discussion, changes, and eventually a record of what shipped.

But that asks too much of the PR.

I also use branches and PRs earlier than that. Sometimes the point is not to ship anything yet. I want to test an idea against the real product, see what breaks, and have a conversation around something that actually runs.

In that situation the PR is part of the design work. It is not the final source of truth.

A designer can try the responsive behavior instead of imagining it. An engineer can find a constraint that changes the interaction. A product partner can click through the flow. Sometimes the design changes because of what we learn there.

So I keep the PR where it is useful and give it context somewhere else: why this branch exists, what question it helped answer, and whether it represents the current direction.

Keep the artifacts. Connect the decisions.

I am less interested now in finding the next source of truth. The work is already happening across several places.

What matters to me at handoff is that someone else can understand the decision without replaying the entire project in their head. That is why the lightweight page has been useful. It does not contain everything. It tells you what matters now and where to go next.

I am still refining this, but I no longer think of handoff as transferring a finished design file from design to engineering. By the time I am handing something over, design and implementation may already have been feeding into each other for a while.