The situation
Every team got faster. Nobody knows where Q4 stands.
It's the middle of Q3. Over the summer, each of your teams took up an AI tool. Platform works in Claude Code and ships small changes all day. Payments runs weekly cycles in Codex, with two approvers on anything that touches money. Mobile plans two-week sprints and codes with Copilot. Each team will tell you it's faster, and each team is right.
Then the CEO asks a simple question. Are we on track for the Q4 roadmap? You get three answers: a dashboard from one team, a spreadsheet from another, a thread of messages from the third. Each is accurate about its own team. None of them adds up to an answer about the roadmap.
The tools aren't the problem. The problem is that nothing sits above them.
Why it's hard
Two tempting fixes, and why both backfire.
Make everyone use one tool
- What it promises
- One way of working, and one report at the end of the quarter.
- What it costs
- Teams lose the tool that works for their work. You spend a quarter on a migration nobody asked for, and some engineers quietly keep their own tool anyway.
Let every team do its own thing
- What it promises
- No disruption. Every team stays fast.
- What it costs
- No shared view of the roadmap, rules that change with the tool, and AI use that nobody outside the team can see.
Let the ways of working differ. Share the layer above them.
Teams keep their tool and their rhythm. The roadmap, the rules and the record belong to the organisation, and they stay the same whatever tool a team uses.
The split
What should differ, and what should be shared.
Drawing this line is most of the work. Everything on the left can change from team to team without hurting anyone. Everything on the right has to be one thing, or leadership is looking at several separate systems.
How each team works
- The AI toolClaude Code, Codex, Copilot: whichever harness the team works best in.
- The rhythmContinuous, weekly or two-week. The work sets the pace, not the org chart.
- The agents a team builds for itselfA compliance checker for Payments, a UX reviewer for Mobile.
- Team rulesTwo approvers on anything touching money. Only the team that needs a rule carries it.
What leadership sees as one
- One roadmapEvery team's plans roll up into it, so "are we on track?" has one answer.
- The agent catalogueAn agent one team builds, another can use, pinned to the same version.
- Organisation-wide guardrailsThe rules that hold for everyone, whatever the tool.
- Identity and accessOne identity provider decides who can do what.
- One audit trailWho asked, which agent did it, and who approved it, in one place.
- One view for leadershipDelivery across every team, without forcing one process on them.
The questions
Five questions for your next leadership review.
Ask your team leads. If the answers depend on which team you ask, you've found the part of the layer that's missing.
Can you see the roadmap across teams in one place?
RoadmapIf each team plans in its own tool, the roadmap lives in a slide deck that someone rebuilds by hand before every review.
When one team builds a good agent, can another use the same version?
AgentsOtherwise the best prompt in the company is a file on one engineer's laptop, copied and changed by everyone who finds it.
Are your organisation's rules the same, whatever tool a team uses?
GuardrailsA rule written into one tool's settings doesn't exist in the others. The team that switched tools last month may have left it behind.
If one team switched AI tool tomorrow, what would break?
The layerIf the answer is "their plans, their rules and their history", those were living inside the tool instead of above it.
Can you compare delivery across teams without forcing one process?
Leadership viewWeekly cycles and continuous flow look different on a board. The comparison has to come from the record, not from making every team work the same way.
Try this on your organisation
Map your teams, and the layer above them.
List your teams and how they work today, then tick what they genuinely share. It takes two minutes, and nothing you type leaves this page.
This shows what your teams share, not how well they work. Different tools and rhythms are fine. That's the point.
With Wakalix
Each team works its own way. Leadership sees one system.
Wakalix is built on that split. It doesn't replace the tools your teams chose. It gives the organisation the layer above them.
- 1Their tool, their agents, their rhythmTeams choose their harness (Claude Code, Codex or Copilot), the agents they need and the rhythm they work at.
- 2One roadmapEvery product keeps a ranked roadmap. A cycle opens from any item, and the roadmap keeps track of every cycle it spawned and what each one produced.
- 3Agents the organisation ownsAn agent is defined once, owned by your organisation, released like software and pinned for every team: "Developer agent v3, owned by the Platform team, pinned for every team."
- 4One set of guardrails, one recordLeadership sees one roadmap, one set of guardrails and one record across every team. Sign-in comes from your own identity provider, and every step is on the audit trail.
Your teams keep the AI tools they chose, on the subscriptions you already have. Wakalix never runs the model.