What Foundry ships in 8 weeks. A real Build engagement, start to finish.
Week by week: what actually gets shipped when you book a full Foundry Build. No consulting theater, no decks, just the work.
People ask what "Build" actually means. The $95k tier. What you get.
Short version: three production AI systems, wired together, deployed into your cloud, running for your team when we hand you the keys.
Long version follows. This is the standard Build timeline. We flex it for your situation, but the shape is real.
Week 1 — scoping and access
Day one is a kickoff call. Three hours, not three days. We map your stack, identify the three systems we're shipping, agree on what "done" looks like for each.
By end of week 1:
- Scope doc signed
- Access provisioned to your tools (HubSpot, Slack, your DB, whatever)
- Repo created in your GitHub org
- Three tickets in our tracker with acceptance criteria
- A Slack channel with your team and ours
No kickoff decks. No "discovery phase." You already know your business. We already know how to build. We get to work.
Week 2-3 — the first module
First module ships first. Usually an automation because it's the fastest to value.
Example: a lead-routing pipeline. Inbound form submissions get classified by Claude, enriched with Apollo data, routed to the right owner in HubSpot, summarized in a Slack channel for your team. Every step logged. Failure modes handled.
Typical breakdown:
- Week 2: scaffold, wire integrations, draft the Claude prompt, test on real data
- Week 3: ship to production, monitor, tune prompts and routing logic
End of week 3: your first system is live. Real traffic. Real results. You can point your investors or your team at something concrete.
Week 4-6 — the second module
Now we build the dashboard. This is where most of the "wow" lands for your team, because they can see the automation from week 3 in action.
A typical Foundry dashboard:
- Pulls from your ops systems (CRM, billing, ticketing)
- Runs SQL + Claude to summarize what's happening
- Flags anomalies and suggests next actions
- Custom for your role (founder view, ops view, sales view)
By end of week 6: two systems live. The automation is running, the dashboard is watching.
Week 7-9 — the third module
Third one is usually an agent. Support, sales qualification, internal knowledge, depending on your biggest pain.
Agents take longer than automations because the evaluation loop matters. We run real cases through it. We measure deflection, accuracy, escalation rates. We tune until it passes.
End of week 9: three systems live. Your operating layer is real.
Week 10-11 — polish and handoff
We write the runbook. We train your team on how to iterate. We set up monitoring that pages the right people when something breaks.
Documentation:
- How each system works (not marketing copy, actual docs)
- How to change the prompts
- How to add a new data source
- Common failure modes and what to do
This is the most underrated part of a Foundry Build. Most dev shops ship the code and bail. We make sure your team can own it after we leave.
Week 12 — you own it
Final walkthrough call. Keys handed over. Operate retainer starts if you opted in, otherwise you're on your own with our phone number.
Twelve weeks. Three production systems. One operating layer. Fully yours.
What Build is not
It's not a kickoff deck. It's not a 6-month roadmap. It's not a "discovery phase" where you pay us to read your website. It's not an MVP or a prototype or a proof of concept.
It's three real systems in production in 12 weeks.
Ready
Run the audit to see which three systems to build first for your operation. Or book a call if you want to talk through scope.
Eleven more weeks than Zapier. A hundred times more valuable.