Three engagements.
One outcome: agents in production.
We don't sell time. We sell deliverables. Each engagement has a fixed scope, fixed price, and a written readout you can put on a slide.
AI Workflow Audit
The diagnostic that starts every engagement. We map your stack, score your risk, and hand you a prioritized 90-day roadmap.
What we do
- Catalog every agent, workflow, model call, and tool integration across your stack
- Map data flows and identify where PII, secrets, and customer data move
- Score your setup against the 42-point Readiness Checklist
- Benchmark cost & latency against routing best-practice
- Interview 3–5 stakeholders to surface unwritten constraints
- Deliver a prioritized 90-day roadmap with effort & impact estimates
Deliverables
Timeline
Kickoff & access
Read-only access to your AI stack, tools, and observability. 60-minute scoping call.
Discovery
Stakeholder interviews, code & config review, traffic sampling. Async by default.
Analysis
Scoring, cost modeling, risk mapping. Mid-point check-in (30min).
Readout
Written report shipped. 60-minute live readout, recorded. Q&A & next-step planning.
Credit toward Build
If you proceed to a Build Sprint within 60 days, the full $2,500 is credited against the sprint fee.
Build Sprint
Pick one workflow from the audit roadmap. We ship it end-to-end with evals, observability, and runbooks.
What we ship
- One production-grade workflow, end-to-end, with your name on it
- Agent + tool wiring via MCP / A2A, hosted on your infra or ours
- Routing config: primary model, fallbacks, budgets, rate limits
- Policy & guardrails: PII redaction, scope checks, approval gates
- Regression eval suite that runs on every change
- Observability dashboards (Langfuse / Helicone / your stack)
- Runbooks for the on-call rotation
- 2 training sessions with your team
Scope tiers
How we work
One named engineer leads the sprint, with design & review from the partner team. Daily async updates in a shared Slack, weekly demo, mid-sprint & final reviews. You own the code from day one — we push to your repo.
30-day post-launch
Every sprint includes 30 days of bug-fix & tuning support after launch, at no extra cost. After that you can transition to a Governance retainer or run it yourself.
Gateway & Governance
We run the routing layer for you. On-call, policy reviews, model migrations, quarterly evals.
What we manage
- Routing gateway (Portkey / OpenRouter / LiteLLM / self-hosted)
- Policy & guardrails — quarterly review, monthly tuning
- Cost SLOs with per-route budget alerts
- Latency SLOs with p50 / p95 targets
- Model migration support when new releases drop
- Eval re-runs on every model upgrade
- Incident on-call (4-hour response, business hours)
Tier breakdown
What we won't do
We won't build new features under a governance retainer — that's what Build Sprints are for. We won't take ownership of business logic. We won't put ourselves on the critical path of your product roadmap. The point of governance is to make the AI layer boring, not bottlenecked.
Things people ask before booking.
Do you require us to use a specific gateway?
No. We're vendor-neutral. If you're already on Portkey, OpenRouter, LiteLLM, Helicone, or rolling your own — we'll work with what you have. If you're starting fresh, we'll recommend based on your team and traffic profile.
Can you work on top of n8n / Zapier / Make?
Yes — most of our clients have at least one no-code automation in the mix. We treat them as sources or destinations and route their traffic through the gateway like anything else.
Do you sign NDAs?
Yes, mutual NDAs before the audit. Build engagements include a more detailed data processing agreement.
What if we're not ready for an audit?
Start with the free Readiness Checklist on the home page. If you score below 60%, you're not ready — and you'll know exactly what to fix first.
Do you do hourly or T&M work?
No. Every engagement is fixed-scope. If we discover the scope was wrong, we eat the difference — that's our risk to manage.
Where are you based?
Distributed team across North America & Europe. We work async by default with 2–3 live touchpoints per week.
Start where we always start.
Every engagement begins with the Audit. Two weeks. $2,500. Credited if you continue.