Tier 3 · Enterprise · From $75K · 8-16 weeks
Claude Code across your engineering org. Claude Enterprise for everyone else. Custom MCP servers for your internal systems. Production apps on the Anthropic API. SSO, governance, cost dashboards, per-role training. The whole engagement.
Why This Exists
Most implementation shops know one of three things. They know Claude Code (the command-line tool engineers use). Or they know Claude.ai (the chat product knowledge workers use). Or they know the Anthropic API (the backbone for production apps). Rarely all three.
Enterprise rollouts need all three. Engineering needs Claude Code with shared MCP servers and skills. Operations and customer-facing teams need Claude Enterprise with SSO and retention controls. And the real cost savings come from custom production apps on the API, with prompt caching and tool use doing the heavy lifting.
We handle all three inside one engagement. One team, one plan, one owner.
Scope of Work
Claude Code rollout (engineering)
Claude Code across your engineering teams. Shared CLAUDE.md standards, a central skills repo in GitHub, settings.json templates, permission allowlists that match your security policy. Onboarded through squad-level working sessions, not top-down training.
Claude Enterprise deployment
The chat product for everyone else. SSO integration, role-based access, retention settings, data controls, admin dashboards. Rolled out by department with role-specific onboarding so the sales team learns different things from the finance team.
Custom MCP servers
Claude needs context on your actual systems. We build MCP servers for your internal APIs: your CRM, your data warehouse, your ticketing system, your internal docs. So Claude can read your real data, not just generic context. This is the single biggest lever for making Claude feel useful at enterprise scale.
Production API apps
One or two production apps built on the Anthropic API. Usually: a customer-facing agent (support, sales, onboarding) or an internal back-office agent (finance reconciliation, HR triage, research). Built with prompt caching (70-90% cost reduction), tool use, the Agent SDK, and batch API where relevant. These apps pay for the engagement.
Governance and observability
A governance doc covering what data can and cannot go to Claude, who signs off on new use cases, and what the escalation path looks like. Plus a cost and usage dashboard so finance knows what the bill is doing and why.
Per-role training tracks
Different people need different things. Engineers learn Claude Code and MCP servers. Knowledge workers learn Claude.ai and document workflows. Executives learn when to ask for an AI solution and when not to. We run the tracks live and record them for reuse.
Deliverables
Across your engineering teams, with shared patterns, a skills repo your team owns, and an onboarding doc for every new hire.
SSO, retention policy, access controls, admin reporting. Provisioned and rolled out by department with per-role training.
Production-grade MCP servers for your internal systems. Source code, documentation, and a runbook. Maintainable by your team.
Shipped. Running. Measured. With prompt caching and tool use configured. Usually the thing that makes the engagement ROI-positive inside the first year.
A short, working document covering use-case approval, data classification, escalation, and retention. Not a 40-page compliance binder nobody reads.
Live visibility into who is using what, how often, and at what cost. Finance stops being surprised by the Anthropic bill. Engineering can target optimisations where they matter.
How We Work
Most engagements land between $75,000 and $250,000 depending on org size, the number of custom MCP servers, and whether production API apps are in scope. Quoted fixed-fee after a paid two-week discovery phase. You know the full cost before you commit to delivery.
After the engagement closes, clients typically roll into a retainer covering ongoing skill development, new MCP server builds, and per-quarter reviews. Optional. Month-to-month. Cancellable at any time.
We work alongside your existing engineering, IT, and security functions. We are not trying to own your AI strategy. We are there to make your team faster and hand the keys over when the engagement ends.
Why Work With Us
Big consulting firms are happy to sell you a two-year engagement. That is their revenue model. Ours is different. We run short, intense, outcome-fixed engagements and leave you with documentation your team can run without us.
We also do not resell Anthropic licences or take any commission on them. You buy the licences directly from Anthropic. That keeps our recommendations honest: we tell you when Claude is the right tool and when it is not.
And we have one more thing a lot of implementation firms do not: we use Claude Code every day to ship client work. You will see us run it live in sessions. No stock screenshots, no slide decks from six months ago.
Start with a 30-minute discovery call. We tell you what a good engagement looks like for your org, whether or not we are the right people to run it.