How it works

Three stages. The same method on every engagement. Clear controls at every step.

The problem with most automation

Most automation gets deployed without governance. Tools go live fast. Approval steps get added to the backlog and stay there. Emails go out from workflows nobody can audit. A step runs twice, or not at all, and nobody knows until a client complains.

The problem is not the automation. It is that the controls were never built. Approval gates, named ownership, audit trails, and exception paths are not hard to build -- they just get skipped because they slow down the initial deploy.

We build them in from the start. Not because it is slower, but because it is the only way to build something you can actually trust and hand off.

The three stages

Operations Review

Understand before you build.

Before any automation is written, we map what you actually do. Not what you think you do -- what a typical week looks like in practice. Where time goes. Where things get dropped. Where the approval or escalation path is unclear.

The review produces a ranked workflow backlog: the highest-ROI, lowest-risk workflows to automate first, with control points and ownership defined before a single line is written.

  • Current state map (where time goes, where things drop)
  • Ranked workflow backlog (3 recommendations with rationale)
  • Controls pack: approval matrix, audit log format, rollback notes
  • Risk flags: what must not run automatically

from £497

Build Sprint

Draft-first. Approval-gated. Documented.

We build 3 governed workflows inside your existing stack. Every workflow follows the same pattern: outputs are drafted before any action is taken externally. A named owner reviews and approves. Only then does anything execute.

Draft

Output is created as a draft. Nothing external happens. The owner reviews.

Approve

Owner reviews the draft and approves, edits, or blocks. Reason is recorded.

Execute

Only after approval does the workflow act externally. Evidence is logged.

  • 3 workflows built inside your existing tools (Notion, Google Workspace, Slack, CRM)
  • Approval gates on all external actions
  • Audit trail on every run: what ran, who approved, evidence link
  • Exception path: defined pause-and-escalate process for failures
  • Runbooks and handover documentation

from £2,500

Oversight and Optimisation

Optional. Your team can run without us.

Every workflow ships with a runbook. Your team can operate, edit, and update without dependency on us. If you want ongoing reliability checks, change management, and a monthly optimisation queue, that is what Oversight covers.

  • Reliability checks and monitoring
  • Change log and versioned updates
  • Monthly optimisation queue
  • Exception handling and incident log

from £750/month

What an audit trail looks like

Every run writes a row. You can see what ran, who owned it, what the output was, and whether it passed or failed -- without digging through Slack or asking anyone.

Time Workflow Step Owner Status Evidence
07:27 Morning briefing Collectors ran Operaitions OK Draft Pack link
07:31 Morning briefing Notion upsert Operaitions OK Page link
09:05 Follow-up queue Draft hygiene guard Operaitions OK Guard summary
09:08 Follow-up queue DM draft created Owner Pending approval Draft link
09:41 Follow-up queue Owner approved Owner Approved Approval note

Start with an Operations Review

The review maps what you actually do, finds the highest-ROI workflows to automate first, and defines the control points before anything is built. Most clients have more automatable patterns than they realise.