top of page

ONE-DAY TEAM AI PROCESS WORKSHOP FOR AGILE TEAMS

Your team doesn’t need another framework. It needs to know what to keep, what to add, and what to let go.

Most enterprise agile teams run some version of Scrum. Now AI is in the loop. Code arrives faster, but your Sprint Reviews still look the same, backlogs fill with tasks and nobody is quite sure who answer for what an agent has produced.

​

You don't need a transformation program. You need one focused day where your team comes together in an AI process workshop for agile teams to decide what to keep, what to add, what to let go, and what to try next. 

Sound familiar?

  • Sprints are busy, but it's hard to say which outcome moved.

  • Developers use AI on their own. The team's way of working hasn't changed.

  • The Product Backlog is a list of tasks that nobody has time to order.

  • The Definition of Done was written before an agent wrote code.

  • Sprint Reviews demo output instead of inspecting evidence.

What your team leaves with

A plan the team writes and owns, to evolve how they work:

  1. A keep / let go / try map of every practice you run today, sorted honestly. 

  2. Three to five experiments written as bets: What you believe, how you'll measure it and by when.

  3. A Definition of Done for agent work: It works, it's explainable and its instrumented.

  4. Clear accountability for who answers for what agents produce.

  5. A 30-day checkpoint: The date and the evidence you will inspect together.

THE LENS WE USE

Bet. Build. Prove.

Bet

Test cheaply fast

Frame the problem, name the riskiest assumption and run the cheapest test that could prove you wrong

Build

Only what earns it

Work flows continuously. People and agents build against a bet, not a batch and tasks become disposable.

Prove

Measure the outcome

Inspect the evidence against the bet, then decide to scale, adjust or stop.

The rule we hold to: Agents do work; only people are accountable.

The day (sample agenda)

TIME

SESSION

TEAM OUTPUT

09:00

Where you really are. Scrum as written versus Scrum as practiced and where AI already shows up in your work.

Current-state map

10:30

Keep, let go, question. Test each practice: Empiricism, goals, Done, the Sprint as a container, your backlogs.

Keep / let go / try map

12:00

 Lunch

12:45

From tasks to bets. Rewrite a slice of your real Product Backlog as bet and place each on the evidence ladder.

Two or three live bets in Product Backlog

14:15

Done for agent work. Redraft your Definition of Done and agree on who answers for what agents produce.

Draft Definition of Done and Accountabilities

15:30

Build the plan. Choose experiments, owners and dates. Your sponsor joins for the readout.

The team's evolution plan

Who attends

One whole Scrum Team of 5-10 people: Product Owner, Scrum Master and Developers. The sponsor joins for the last 45 minutes.

​

Some form of Scrum (it doesn't have to be pretty) and basic AI familiarity. No fluency needed.

How it runs

Onsite: One day at your office.

Virtual: Two half-days.

​

Before: A 15-minute team survey, plus your Product Backlog and Definition of Done.

​

After: A 60-minute check-in at 30 days to inspect the first evidence.

Ready to evolve?

Tell us about your team and how you work today. We'll be in touch to plan your workshop day.

bottom of page