Commercial Architecture
Enterprise SKU LadderSprint, Build, then Run for 2-5 Years
Not cancel-anytime task queues. Not open-ended billable hours. A productized wedge, an accelerated production build, and a software-style operating subscription - the exact model regulated enterprises know how to buy.
Sprint
4 Weeks
Fixed Price · Published “from” on proposal
The low-risk door for regulated institutions. Delivers a working agent prototype on sandbox data, architecture blueprint, and governance brief.
Included Capabilities:
- Use-case prioritization & ROI feasibility model
- Multi-agent pipeline architecture blueprint
- Working sandbox prototype deployment
- ISO 27001 compliant governance brief
Build
Weeks, not Months
Fixed or Capped Investment · Private proposal
AI-accelerated production engineering with senior humans on the loop. Priced at a fraction of traditional Swiss systems integrators without compromising security.
Included Capabilities:
- Full production deployment to your cloud (Azure / AWS / Hetzner)
- Enterprise integrations (Core banking, CRM, ERP, API gateways)
- Audra Eval CI/CD quality gate integration
- Complete InfoSec evidence pack for compliance approval
Run
24-60 Months
Annual Operating Subscription · Private proposal
Software-style operating subscription, not billable hours. We take full responsibility for model drift, security patches, monthly feature velocity, and token allowances.
Included Capabilities:
- Continuous Audra Eval quality gates & drift monitoring
- Wrapped token consumption allowance (no surprise pass-through invoices)
- Model & LLM vendor upgrades (OpenAI, Anthropic, open-source)
- Dedicated Named Senior Architect & guaranteed monthly feature velocity
Market Positioning
Why this model wins
See how Windmill Digital’s Sprint → Build → Run model compares to cancel-anytime vibe boutiques and traditional multi-year billing shops.
| Dimension | Vibe-Coding Boutiques | Traditional Swiss SIs | Windmill Digital |
|---|---|---|---|
| Delivery Model | Ad-hoc queue / "Unlimited tasks" | Time & materials / Billable hours | Sprint → Build → 2-5 Year Run |
| Time to Working Sandbox | Days (unverified demos) | 6-12 months of scoping | 4 weeks (Agentic AI Sprint) |
| Security & Compliance | None / Freelancer laptops | Traditional heavy audits | ISO 27001:2022 certified, Swiss data |
| AI Quality Assurance | Subjective / No evals | Manual QA testing | Audra Eval automated CI/CD gates |
| Production Operating Model | Cancel anytime / Abandoned codebase | Retainer for support hours | Continuous SLA, drift monitoring & feature velocity |
| Token & LLM Governance | Customer pays OpenAI bills | Pass-through markup | Wrapped allowance in operating subscription |
Wrapped token allowances, zero pass-through surprises
Enterprise procurement hates unpredictable API invoices. Under our 24-60 month Run subscription, your base token consumption and infrastructure compute are wrapped directly into your annual fee with a generous allowance ceiling.
- Fixed Monthly Cost: Predictable operational budgets without micro-managing token counts.
- Model Arbitrage: We route easy queries to smaller models and complex reasoning to frontier models to maximize efficiency.
- Enterprise Security: Zero data training guarantees across all LLM contracts.
The Monks Evidence Runway
Institutions rarely sign 5-year operating agreements with unknown vendors. That is why our Sprint exists. The 4-week Sprint delivers tangible code on your data within 30 days. That provides the internal evidence runway your procurement and risk teams need to confidently enter a 2-5 year Run partnership.