How We Work
A four-stage engagement, fixed at every step
Most AI engagements are open-ended consulting bills with a "discovery phase" that never ends. Ours is the opposite. Each stage has a fixed scope, fixed timeline, and a clear deliverable. You can stop after any stage with no further commitment.
The shape below is what 90% of our engagements look like. If you have something genuinely unusual, we'll tell you and propose a custom shape.
Stage 1
Discovery call
30 minutes · free
A direct conversation with Daniel. You describe the workflow that's costing you the most. We tell you honestly whether AI is the right tool, what the realistic shape of a solution looks like, and whether we're a fit. No deck, no follow-up sequence. If we're not the right team, we'll say so and point you at someone who is.
Stage 2
Operations audit
Short engagement · fixed price
If discovery surfaces a real opportunity, we run a paid audit. Our team spends time embedded with you: shadowing the workflow, mapping data flow, identifying the constraints, and writing up a focused diagnostic with three or four concrete automation candidates ranked by ROI and feasibility. You own the document. If you take it to another vendor, that's fine.
Stage 3
Pilot build
Several weeks · fixed scope
We build one workflow end-to-end and ship it into production. Fixed scope, fixed timeline, fixed price. The pilot is real software your team uses, not a demo. We deliver the code, the deployment, and the integration into your existing systems. By the end you have either a working automation or honest evidence that the approach doesn't fit your operation. Both outcomes are better than months of slideware.
Stage 4
Scale phase
Monthly retainer or per-project
Once the pilot proves value, we extend. Either as a monthly retainer that covers ongoing improvements, model updates, and new workflows; or as discrete projects you commission against the original architecture. Most clients move to retainer once a few pilots have shipped and they want a single team owning the AI surface across their operation.
Pricing
What it actually costs
Every other consultancy hides numbers. We won't. Here's the real range so you know whether the conversation is worth having before you book the call.
Discovery
Free
30-minute conversation, no obligation.
Operations audit
Fixed price, scoped per engagement
Our team embeds with you to produce a written diagnostic, an in-person workshop, and a recorded debrief. Indicative pricing on request, varying with team size and complexity.
Pilot build
Fixed scope, fixed timeline
Smaller automations sit at the lower end of a typical mid five-figure investment. Voice agents and full mobile apps run higher. The audit produces a precise quote before any build commitment.
Scale retainer
Monthly retainer, sized to engagement
Senior team capacity, model costs passed through at-cost, ongoing maintenance and new workflow work. Cancel any time.
Pricing is indicative only. Actual quotes depend on scope, integrations, data volumes, and team requirements. The audit produces precise numbers before you commit to any build work. Model and infrastructure costs (Anthropic, OpenAI, ElevenLabs, cloud hosting) are billed at-cost with no markup.
What we need from you
The minimum, not the maximum
Most consultancies ask you to fill out an RFP. We ask for four things, and we ask for them in person.
- An operational sponsor who can describe the current workflow and approve trade-offs.
- Read-only access to the data and systems involved (we'll spec exactly what's needed).
- One or two practitioners who actually do the work, willing to spend a few hours in interviews.
- Realistic acceptance that AI isn't magic. We'll be straight about what it can and can't do.
Team structure
Senior, small, owns the outcome
Every pilot runs with our team paired with yours for the duration of the build. Senior people only — we don't add juniors, account managers, or "delivery leads". Your sponsor talks directly to the people writing the code.
We run a small number of concurrent pilots so each one gets the attention it needs. When we're at capacity we'll tell you and slot you for the next start window rather than over-promise.