Skip to main content
Forge

Strategy · 5 min read

An automation strategy that scales past the pilot

Most automation programmes stall after the first success. The pattern is predictable, and so is the way out of it.

Marcus VaneHead of Platform Strategy

Why the second workflow is harder than the first

The first workflow gets executive attention, a dedicated team and permission to cut corners. The second gets none of those. If the first one left behind bespoke infrastructure and undocumented decisions, the second effectively starts from zero.

The organisations that scale are the ones that treat workflow one as a platform investment rather than a project. The measure of success is not that it shipped — it is how much of it the next workflow reuses.

Pick candidates by shape, not by noise

The loudest process is rarely the best first candidate. Look instead for shape: high volume, low variance, a clear definition of correct, and an owner who can say yes.

  • High volume — the savings are real and measurable
  • Low variance — the rules can actually be written down
  • Verifiable outcome — you can tell whether a run was right
  • A single accountable owner — nobody needs to convene a committee

Budget for the controls up front

Access, audit and residency reviews will happen whether you planned for them or not. Doing that review once, at the platform level, is the difference between a programme that compounds and one that re-negotiates the same questions every quarter.

Keep reading

Related guides

  • Workflow design

    Designing AI workflows that survive production

    Read guide
  • Engineering

    API architecture for platform teams

    Read guide
  • Operations

    Scaling operations without scaling headcount

    Read guide

Apply it

See these ideas running in the platform

The product and platform pages show how these principles are implemented in Forge itself.