Services
Audit. Build. Operate.
Fixed pricing on every engagement. Each phase ships something concrete. The audit sets up the build; the build sets up the operation.
01 — Audit
Three tiers, scaled to org size.
SMB Workflow Audit
Founders and ops leads at 10–50 employee firms who suspect their back-office is costing more than it should — whether or not they can pinpoint where.
In scope
- Up to 3 core workflows
- 3–5 stakeholder interviews
- One 90-minute kickoff session
- One 60-minute delivery session
Deliverables
- — Current-state process map for each workflow
- — Friction inventory (time cost + error rate per bottleneck)
- — AI-fit assessment for each friction point
- — Build proposal with scoped recommendation and fixed-price estimate
Mid-Market Operations Audit
COOs and ops leads where workflows cross multiple teams and the handoffs between departments are where time and money leak — usually after a quarter or two of trying to fix it internally.
In scope
- 3–5 workflows across up to 2 departments
- 6–10 stakeholder interviews
- Weekly progress notes
- Mid-engagement review session
Deliverables
- — Everything in the SMB tier
- — Data audit (what systems hold what data, accessibility for automation)
- — Phased build roadmap with decision gates
Enterprise Audit
Operations or transformation leads at 500+ employee firms who need an independent assessment before pitching AI internally — typically when there are already competing recommendations or vendor proposals on the table.
In scope
- Custom — defined in a scoping call before proposal
- Multi-stakeholder discovery sprint
- 10+ interviews across leadership and operations
- Typically a single department or division
Deliverables
- — Full discovery sprint output
- — Stakeholder alignment summary
- — Technical architecture review
- — Build roadmap with ROI model
- — Executive-ready summary (separate from the technical report)
02 — Build
Three build tiers, scoped from the audit.
We don't build anything that wasn't identified in the audit. The build proposal — system, integrations, outcomes, fixed price — is the final section of every audit.
Simple build
Single integration, one process. Best for proving AI fit on one workflow before scaling further — the ideal first engagement when you want to test whether automation actually works on your real data.
Scope
- 1 integration
- 1 core process
- Single workflow automation
Examples
- — Document processing pipeline (extract, validate, classify)
- — Single-workflow automation with structured outputs
- — Standalone AI module slotting into existing systems
Standard build
Multi-step agentic pipelines across 2–3 integrations. The most common engagement for mid-market firms.
Scope
- 2–3 integrations
- Agentic decision pipeline
- Exception handling and routing
Examples
- — Multi-step automation across 2–3 systems
- — Agentic decision routing (intake → triage → action)
- — Document → action pipelines with human-review gates
Complex build
Multi-system, multi-department orchestration. Replaces a sequence of manual handoffs across organisational boundaries.
Scope
- Multi-system, multi-department
- Cross-functional orchestration
- Full process replacement
Examples
- — Cross-departmental workflow orchestration
- — Exception handling at scale (hundreds of cases/week)
- — Full back-office process replacement systems
03 — Operate
Production AI systems need ongoing operation.
Drift, prompt tuning as inputs change, model upgrades, performance reviews, small feature additions. Retainers offered at delivery — never promised in the proposal.
Monitoring & Maintenance
2–4 hrs/mo
Monitoring, drift checks, 2 prompt-tuning cycles, 1 check-in call
Operational Optimization
6–10 hrs/mo
Everything in Monitoring & Maintenance, plus feature additions and model upgrade management
Embedded AI Ops
10–20 hrs/mo
Everything in Operational Optimization, plus architecture decisions, team training and vendor evaluation
Boundaries
What we don't do.
Defining a studio by exclusion is sometimes more useful than defining it by service.
We do not build before auditing.
If you already know exactly what you want built, we will ask you to walk us through the process first. If you cannot describe it in a structured walkthrough, we cannot build it reliably.
We do not take every engagement.
We qualify on three criteria: there is a real, measurable process problem; the person we are speaking to has authority to approve the spend; and the engagement can start within 60 days.
Find the workflow worth fixing.
Tell us about it on a 20-minute diagnostic call. No pitch.