Living Playbooks: Knowledge That the Next Executor Can Run

Most playbooks are written once and forgotten. They launch with a new process, sit in a shared drive, and drift from how the work actually runs within weeks. The team knows they are outdated, but nobody is accountable for fixing them. So the knowledge stays tribal, the playbook stays stale, and every new team member learns by shadowing and asking.

This pattern does not just slow onboarding. It creates a structural ceiling on how much improvement can compound. When the way work gets done lives in people's heads rather than in assets the next executor can pick up and run, every improvement has to be re-taught, re-explained, and re-discovered.

Why playbooks decay

Playbooks do not decay because people are careless. They decay because no one owns them after go-live.

The implementation team built the playbook as part of a project. The project ended. The playbook became a snapshot of how the process was supposed to run at launch — not how it actually runs now. Updates happen informally: a workaround discovered, a policy change absorbed, a step abandoned because it never worked. None of it makes it back into the document.

The result is a familiar pattern:

• New hires are handed the playbook, told it is "mostly accurate", and pointed toward someone who can explain the real version.
• Experienced staff carry the current process in their heads. They are effective, but the team depends on their availability.
• Improvement initiatives start from scratch because the documented baseline cannot be trusted.

This is not a documentation problem. It is a governance problem. The playbook was treated as a deliverable, not as an asset that requires ongoing ownership, testing, and update cycles — the same discipline applied to software or other operational infrastructure.

What living playbooks look like

The teams that compound improvement treat operational knowledge differently. They treat playbooks as living assets that need the same rigour applied to any critical system: versioning, ownership, periodic review, and structured formats that support multiple consumers .

Living playbooks share several characteristics:

Owned, not orphaned. Someone is accountable for accuracy. That does not mean one person maintains every playbook — it means every playbook has a named owner who reviews it on a defined cadence and is responsible for flagging when it has drifted.

Versioned and auditable. Changes are tracked. If the process changes, the playbook reflects it — and the team can see what changed, when, and why. This matters for compliance, for troubleshooting, and for understanding how the process evolved.

Structured for multiple readers. A playbook that only a human can interpret is a playbook that cannot scale. The shift toward agent-assisted and agent-led execution means playbooks increasingly need to be structured so machines can parse them: explicit inputs, explicit outputs, explicit decision logic .

Tested in use. The real test of a playbook is whether someone unfamiliar with the process can follow it to completion. Teams that take documentation seriously use new hires or cross-functional reviewers as a test: if they struggle, the playbook is not ready.

The human-or-agent readiness test

As operating models evolve toward a mix of human and AI execution, playbook readiness becomes a prerequisite for capturing the value . A useful frame is to ask, for each playbook: could the next executor run this without asking questions?

That executor might be:

• A new hire starting next month.
• A team in another region picking up the workload.
• An AI agent that needs explicit instructions, not contextual assumptions.

If the answer is no — if the playbook requires interpretation, institutional memory, or verbal clarification — then the playbook is incomplete. The knowledge lives outside the document, which means it cannot be reliably transferred, scaled, or automated.

This does not mean every playbook must be agent-ready on day one. It means the gaps should be visible. When you know a step depends on tacit knowledge, you can decide whether to invest in documenting it, simplify it, or accept that it stays human-led. What you cannot afford is to assume the playbook is complete when it is not.

Making playbooks live

Shifting from project deliverable to living asset requires three changes:

1. Assign ownership. Every playbook needs a named owner with explicit accountability for keeping it current. This is not a suggestion to review when convenient — it is a scheduled review cycle, typically quarterly, with documented sign-off. Ownership can rotate, but the accountability cannot lapse.

2. Embed in the operating rhythm. Playbooks that sit in a separate knowledge base tend to drift faster than those embedded in the flow of work. If the playbook is what the team opens to run the process, discrepancies surface immediately. If it is a reference document nobody touches day-to-day, drift accumulates invisibly.

3. Treat updates as change management. When the process changes, the playbook changes — and the team is notified. Major updates warrant the same attention as a code release: review, testing, and communication. This is what McKinsey's research describes as treating skills "the way they treat software: versioned, tested before deployment, updated when requirements change, and reviewed on a defined cadence" .

None of this requires new technology. It requires treating operational knowledge as a system that needs maintenance, not as a static document that gets written once.

The compounding advantage

The payoff is not just better documentation. It is a structural advantage that compounds.

Teams with living playbooks onboard faster. Improvements propagate. Cross-training becomes possible because the knowledge is portable. Handoffs are cleaner because expectations are explicit. And when the organisation is ready to introduce AI agents, the groundwork is already done — well-documented workflows become blueprints for redesign, not obstacles .

The alternative is the pattern most organisations live with today: playbooks that exist but are not trusted, knowledge that lives in people rather than systems, and a ceiling on how much improvement can scale. Every initiative to automate or optimise starts by re-documenting what should already be documented. Time that could go toward value creation goes toward reconstructing what was never maintained.

Jobs are changing rapidly, and people and organisations that thrive are those equipped not just to understand how things used to be, but to succeed in what comes next — and that starts with investing in how knowledge is captured and kept current .

---

How does your organisation treat playbooks today — as living assets, or as snapshots that drift? If the next executor were an AI agent, which processes would be ready?

Sources

1. 1
   McKinsey & Company. Powering productivity: Operations insights for 2025. 2025.Jobs are changing rapidly; people and organisations that have talent around them understand the way things used to be but are also equipped to be successful in the brave new world — invest in people and their capabilities.View sourceBack to citation
2. 2
   Deloitte Insights. The path to agentic transformation. 2026.Survey of enterprise leaders; 74% expect nearly half of their business processes will be redesigned or rebuilt around AI agents within four years. Well-documented workflows, including prior RPA documentation and data maturity assessments, provided valuable blueprints for redesign efforts.View sourceBack to citation
3. 3
   McKinsey & Company (QuantumBlack). Agent skills: Turn human expertise into AI advantage. 2026.A skill is a set of instructions that define how a task should be done, packaged in a reusable, versioned format. Skills that are not maintained drift from current requirements; teams that sustain improvement treat skills the way they treat software — versioned, tested, updated when requirements change, and reviewed on a defined cadence.View sourceBack to citation
4. 4
   Stanford HAI. The 2026 AI Index Report — Economy. 2026.Productivity gains are largest in structured, measurable work.View sourceBack to citation
5. 5
   Deloitte Insights. Rewiring the operating model for AI. 2026.42% of surveyed leaders believe more than 40% of organisational processes will be automated or AI-enabled by 2028, up from just 6% today — a sevenfold increase. At the core of the AI-native operating model is a shift from managing people to orchestrating work across humans and AI agents.View sourceBack to citation