Workflow design
Shape the inputs, hand-offs and human review into something people can use.
Services
From the first useful workflow to a system that keeps delivering.
Most teams don't need more AI subscriptions. They need a clear place to start, the technical judgement to connect the pieces, and enough momentum to make it stick.
01
Diagnose
AI Workflow Audit
A fixed-scope two-hour diagnosis of the tools, workflow, risks and ownership around the work worth fixing first.
Tool and spend map
Workflow diagnosis and exclusions
Risks, ownership and data boundaries
Written recommendation for the next move
02 / Design
AI Enablement Day
Guardrails and ownership
Decide what stays with the team, what needs review, and where the boundaries sit.
Measure and hand over
Agree a useful success signal and leave a practical plan for adoption.
03
Implement
Implementation Sprint
A scoped delivery engagement for the workflow that is ready to be connected, tested and put into everyday use.
Integration spike
Connect the systems and data needed to test the workflow without replacing the tools people already rely on.
Build and testing
Make the workflow observable, reviewable and robust enough for everyday use.
Rollout and documentation
Introduce it with the team, document ownership and leave clear fallbacks.
04
Operate
Managed agent operations
Hermes AgentA managed agent, shaped around the work it needs to do.
For workflows that need to keep running after the workshop, I can host and manage a Hermes agent — including the workflow adaptation, integrations and ongoing technical care. It can be supplied with an AI usage allowance, or run against your own model keys when you want direct control of spend.
Managed allowanceOne predictable service covering hosting and agreed AI usage.
Bring your own keysYour model accounts and usage; my infrastructure and support.
A sensible first step
Start with the 5-minute workflow check.
Nine practical questions, an immediate brief, and no sales call required. Add your email only if you want a PDF copy.