Diagnose the work.
Trace the work, handoffs and decisions. Identify the constraints behind them, then test where AI changes the problem — and where it doesn’t.
Rebuild how your team
works with AI.
Don’t just make the old process faster. Work out what AI changes,
then build the tools your team needs to work differently.
Rethinking the weekly reportExample working brief
Diagnose
Does the team need a faster report — or a different flow?
An example: a team gathers project updates into a weekly report so leaders can spot issues and make decisions.
The work today
Collect updates. Chase missing details. Rewrite them into a report. Wait for the weekly meeting to raise an issue.
Questions before a build
- Which decisions actually need this report?
- Is gathering information still the constraint?
- Could AI surface issues from agreed sources?
Output: a map of the work, its constraints and a specific AI change worth testing.
A faster process isn’t always
a better way to work.
Many processes exist because gathering information, coordinating people or producing work used to be hard. AI may change those constraints. Before automating the steps, ask which ones still need to exist.
- Start with a real business problem
- Test what AI actually changes
- Advice and hands-on implementation
Redesign the workflow.
Work out what can change, what can stop and what people still need to own. Define the new workflow, decision rights and review points before building.
Build and test it.
Build the application or automation that supports the new shape. Test it on real work, check the output and help the team learn to run it.
Clear responsibilities.
Tools that support the work.
For the team: a workflow they can run.
Useful tools, clear owners and a practical way to check outputs and handle exceptions. The team knows how to run the workflow, not just how the demo works.
Example workflow handover
Who does what, and when
- AI organises evidence from agreed sources
- The reviewer checks gaps and source quality
- The decision owner makes the final call
For leaders: a clear case for what to change.
What is slowing the work, what should change and what is worth building. Based on what AI can do here, not a list of tools to buy.
Example rebuild brief
Question to resolveOutput
What is slowing the work?Trace the underlying constraintDiagnosis
What should change?Rethink handoffs and decisionsNew workflow
What needs building?Define scope and review criteriaBuild brief
Better tools.
The same old bottlenecks?
More output. Same decisions.
AI produces more reports, but the team still waits for the same approvals. The bottleneck may be the decision, not the draft.
New tools. Old handoffs.
A pilot helps one person, while everyone around them still works the old way. Look at the whole flow, not one isolated task.
A demo. No owner.
Something impressive gets built, but nobody owns the checks, exceptions or everyday use. Design that responsibility into the work.
An operator’s view.
A builder’s approach.
I’m Marc Vallverdu, a former COO and commercial operator who builds with AI.
I’ve spent my career making work happen across teams: the handoffs, the decisions and the places it gets stuck. If AI changes what’s possible, which steps should remain — and what should be rebuilt?
Thinking Shape combines that operator’s view with hands-on building. Sometimes the answer is a new application. Sometimes it’s fewer steps, or better use of what you already have.
Marc VallverduFounder, Thinking Shape
Questions
Bring the work.
Let’s find a useful starting point.
Tell us about your team, the work that gets stuck
and what you want to change.