This engagement is for companies that haven't yet committed to an AI stack and want one engagement that produces a working system. It's the right starting point for most teams in the 10–250 person range.
We come in fresh. We map the workflows that actually matter, design the stack to serve them, and ship a working version in three weeks. You leave with tooling your team is using, documentation to maintain it, and a forecast for what it'll cost over the next year.
What this is not
It's not a transformation program. We're not here to remake how your company works. We're here to give you the AI infrastructure that fits how it already works, with light upgrades where they make sense.
It's not bespoke. We use battle-tested tools and patterns. The lift in this engagement is matching them to you, not inventing new ones.
Why three weeks
Two weeks is too short to actually build. Four weeks invites scope creep. Three is the sweet spot for engagements where the deliverable is a working system, not a deck.
Engagement constraint
We run one of these per month, max. There's only so much bandwidth to ship a working stack inside three weeks.
Who this is for
- Founders or operations leaders who know AI matters but haven't picked a stack
- Teams that tried to assemble their own AI tooling and ended up with a Frankenstein
- Companies in 10–250 person range — large enough to need real tooling, small enough to move fast
- Leaders who want one engagement that produces a working system, not a slide deck
What's included
- Discovery: workflows, constraints, regulatory environment, current tools
- Stack design: model picks, tooling picks, integration architecture
- Build: actual setup of accounts, integrations, and core workflows
- Training: one team-wide session plus a recorded onboarding library
- Documentation: written runbook for ongoing operations
- 60-day post-engagement support window
Process
- 01 Discovery (Week 1)
Three days of workshops with the leadership team and the people who'll actually use the stack. Map workflows, surface constraints, identify the 3–5 use cases that matter most.
- 02 Design (Week 1, late)
Stack proposal with specific tools, models, integrations, and a written rationale for each pick. Reviewed and adjusted with you.
- 03 Build (Week 2)
We stand up accounts, configure tooling, build the core workflows, integrate with your existing systems where needed. You're in the loop daily.
- 04 Train (Week 3, early)
One live training session for the whole team, recorded for anyone who can't attend. Plus a written runbook for the people who'll own the stack going forward.
- 05 Handover (Week 3, late)
Documentation review, final tuning, recorded executive walkthrough. Then a 60-day window for follow-up questions and adjustments.
Deliverables
- Working AI stack tailored to your workflows
- Written stack architecture and rationale document
- Team training session (live + recorded)
- Operations runbook for the team that'll own it
- Cost forecast for the next 12 months
- 60-day follow-up support window
FAQ
- Do we need to have a clear AI strategy before we start?
- No. Most clients don't. The first week of the engagement is about turning a vague intent into a concrete strategy, then the next two weeks are about shipping it.
- What if our regulatory environment requires self-hosted models?
- We do this often. The recommended stack will favor open-weights models on your infrastructure, with a clean fallback path to hosted for non-sensitive workloads. We've shipped this for healthcare, legal, and financial-services clients.
- What if our team resists the stack you build?
- We won't ship something they'll resist. The discovery week includes interviews with the people who'll use the tools, not just the people buying them. If we hear strong objections, we adjust.
- What does this cost over time?
- Engagement fee is fixed. Ongoing costs depend on the stack — typically $20–$60 per user per month for hosted SaaS components, plus any self-hosted infrastructure. We'll forecast the full 12-month cost as part of the deliverables.