I build agent workflows for GTM teams at funded startups. End-to-end — process audit, agent build on Anthropic's managed agents, MCP connectors into HubSpot / Gong / Notion / Slack, validation architecture for compliance-adjacent workflows, adoption coaching, ongoing maintenance.
I also do shorter audits when teams aren't sure where to start, and pure technical builds (custom agents, MCP integration, RAG pipelines, agent orchestration and architecture, evals) when the team has the strategy and just needs hands.
How an engagement runs
01 · Audit & foundation (2–3 weeks). Role-by-role workflow interviews — what BDRs / AEs / SolEngs actually do versus the playbook. Knowledge architecture mapping. Stack connectivity check (APIs, MCP). Risk tiering by workflow. Output: a prioritised build roadmap and a written governance policy.
02 · Build (6–8 weeks). RAG pipeline over internal docs / call transcripts / Notion / CRM notes. Role-specific Claude workspaces with system prompts, context loading, tool connections. BDR / AE / SolEng tooling. Meeting intelligence with structured CRM push. All MCP-connected to the existing stack — no replacement products.
03 · Embed & retainer (ongoing). Role-specific adoption sessions. Prompt library maintenance tied to product and ICP changes. New workflows as gaps surface. Monthly metrics: time saved per rep, output quality, pipeline velocity.
Scope
Scoped after a call. I'd rather quote the project you've actually got than the one a pricing page implies.