
Business Process Automation With MCP: A Practical Guide
What MCP changes for business process automation: which workflows to start with, the tools that support it, and the guardrails that keep AI-run processes safe.
· 16 min read
AI assistants on your docs
Your team and your customers ask the same questions every week. We build assistants that answer from your own documents, show where each answer came from, and hand off to a person when they're not sure.

Sound familiar?
What we build
Answers from handbooks, SOPs and policies, inside Slack, Teams or a web app.
Answers from your help docs and product pages, with a hand-off to your team.
Each answer links to the passage it came from, so people can check it.
When the docs don't cover a question, it says so instead of guessing.
How we work
Five quick questions about what you need and why.
A short call, then a written scope and a fixed quote before any work starts.
Working previews every week, on your real data where possible.
We ship it, watch how it's used and fix what gets in the way.
Built by the studio
A DeFi analytics platform we built with a RAG-backed chat that answers from live data and news.
From the blog

What MCP changes for business process automation: which workflows to start with, the tools that support it, and the guardrails that keep AI-run processes safe.
· 16 min read
Business API plans from providers like OpenAI and Anthropic don't train on your data by default. We agree where your documents are stored during scoping.
It says it can't find the answer in your docs and offers to pass the question to a person. It doesn't guess.
PDFs, Google Docs, Notion, help-centre articles, web pages and most file formats. We can keep it in sync as the documents change.