Before we propose a thing, we listen. Assess is the stage where we map your workflow, talk to the people who run it, and identify where AI fits — and where it doesn't. The deliverable is an Opportunity Report: prioritized, costed, honest about what AI can't fix.
What we ship
- Stakeholder interviews (3–8 people)
- Workflow map of the area in scope
- Opportunity Report with ranked use cases
- Honest 'AI isn't the answer here' callouts
- Recommended next stage + scope
What you do
- Open the door to your team
- Share existing process docs, even rough ones
- Tell us what's already been tried
- Read the report. Push back hard.
Exit criteria
- Opportunity Report delivered + reviewed
- Top 3 priorities ranked + budgeted
- Decision: continue to Design, or part ways
What if it doesn't go to plan- We find AI isn't the right tool. That's a fine outcome.
- We find more opportunity than you expected. Also fine.
Design is where we get specific. Architecture. Tool selection. User flow. Data model. The output is a Solution Spec your team can read in 20 minutes — and a working prototype of the riskiest part. If we can't prototype it in a week, that's a signal, and we say so.
What we ship
- Solution Spec (20-page max — readable)
- Working prototype of the riskiest piece
- Tool / model selection memo
- Cost model with monthly run-rate
- Fixed-bid quote for the Build stage
What you do
- Review the spec — every page
- Stress-test the prototype with real workflow
- Approve the quote, or renegotiate scope
Exit criteria
- Spec signed off
- Prototype demonstrated against real input
- Fixed-bid Build quote accepted
What if it doesn't go to plan- Prototype reveals a blocker. We replan together, on us.
- Scope grows. We re-quote transparently before we touch keyboard.
03 / 05Build
Ship the thing.
The Build stage is where most consulting firms disappear. We don't. Build runs in weekly check-ins, visible inside the COS, with a working artifact at the end of every week. By the end of Build, your team has the tool deployed in your environment, the code in your repo, and a one-page handoff doc.
What we ship
- Weekly working build, no exceptions
- Deployment to your environment
- Source code in your repo (not ours)
- Documentation in your wiki
- 30-day post-launch support included
What you do
- Show up to the weekly check-in (30 min)
- Test what we shipped that week
- Tell us when something doesn't match the spec
Exit criteria
- Acceptance criteria from Design — all met
- Production deployment verified
- Handoff doc accepted by your team
What if it doesn't go to plan- We under-quoted the work. We eat the overage. Always.
- You change scope. We re-quote in writing before continuing.
Software you can't use is software you don't need. Train is where we transfer the operating knowledge to your team — not in slides, but in shoulder-to-shoulder time. The goal is competence, not certification: by the end of Train, the people who'll use this every day actually use it confidently.
What we ship
- Hands-on training sessions (live, recorded)
- Prompt library + playbook tailored to your team
- Office hours via Slack for 30 days post-launch
- Champion identification — who's your in-house lead?
What you do
- Send the people who'll actually use the tool
- Schedule the sessions inside work hours
- Designate the in-house Champion
Exit criteria
- Champion identified + trained
- Team demo'd live, using the system on real work
- Playbook accepted, in your wiki
What if it doesn't go to plan- Team isn't ready to adopt. We extend Train or pause to Optimize.
- Champion turnover. We re-train under AaaS retainer.
Optimize is the ongoing stage — and the one most engagements forget exists. We measure against the success metrics defined in Assess, tune the system monthly, and aim to make ourselves unnecessary. The COS tracks the work. The AaaS retainer covers the cost. Graduation — when you don't need us anymore — is the goal, not the threat.
What we ship
- Monthly ROI memo against agreed metrics
- Active tuning of prompts, agents, dashboards
- Quarterly playbook updates
- Graduation roadmap — when you stop needing us
What you do
- Approve monthly priorities (30-min sync)
- Surface friction the moment you see it
- Tell us when you're ready to graduate
Exit criteria
- Metrics moving in the agreed direction
- Champion running internal office hours
- Or: graduation — engagement closed clean
What if it doesn't go to plan- Metric doesn't move. We diagnose + tune at no charge.
- Team graduates faster than expected. Best outcome.