Web4Guru AI Operations
Docs · Playbooks · Writing your own
Roadmap — Enterprise tier Target: post-v1.0

Writing your own playbook

Enterprise-tier customers will be able to author their own playbooks. Here is how the feature will work and where it sits on the roadmap. Nothing on this page ships in v0.1.

This is a roadmap note

The feature described below is not shipped. Today, new playbooks arrive through the Skill Pack release cadence (v0.2, v0.3, etc.). User-authored playbooks land after v1.0 as part of the Enterprise tier.

Why it's on the roadmap

The 11 playbooks we're planning will cover the vast majority of solo-service-business workflows. But every business has a few moves that are uniquely theirs — a specific proposal format, a peculiar onboarding flow, a nightly data-pull from a bespoke CRM. Those moves deserve first-class playbooks too. For Enterprise-tier customers, that means the ability to author them.

What authoring will look like

The shape is pinned by the 11 we've already written. A custom playbook will follow the same contract:

  • Name + triggers — the phrases and situations that should route to this playbook.
  • Preconditions — required connections and config.
  • Steps — ordered calls to specialists with explicit briefs.
  • Fallbacks — how to handle each step failing (usually "retry once, then surface an error_alert").
  • Success criteria — the evaluator rubric the output must pass before the owner sees it.

Authoring will be in plain English plus structured fields — not code. You describe the intent; the CEO helps you formalise it.

How we think about timing

User-authored playbooks are post-v1.0 for two reasons. First, we want the hand-tuned library to cover enough ground that most customers never need to write one. Second, authoring is only safe once the evaluator rubric language is stable — otherwise custom playbooks would ship brittle work to real customers. Both prerequisites land inside the v0.5 / v1.0 window.

In the meantime

The CEO can already run arbitrary orchestration when no named playbook fits. It will decompose the work, delegate to specialists, run the evaluator, and ship a result. The output is not as polished as a named playbook, but it's real. See triggering playbooks for how improvisation works in v0.1.

If a specific workflow is slowing you down today, tell us. We ship new playbooks every release — your use case may become the v0.3 entry.

Key takeaways

  • User-authored playbooks are planned, not shipped.
  • They will land on the Enterprise tier after v1.0.
  • Authoring will be structured plain-English, not code.
  • Until then, the CEO improvises when no named playbook fits.

What to read next

Have a workflow in mind?

Tell us what you run weekly. We ship new playbooks every release — the best ones come from real owner requests.