# Intelligence Intelligence > A service run by Robin Metcalfe, a developer in Edinburgh, for businesses that already have AI agents doing real work and want them to do more. He connects agents to the business's own systems and content, decides with the business what each agent may reach, puts a person in front of anything that cannot be undone, and keeps a record of everything it did. He also finishes and launches projects people have built with AI and cannot get over the line. Run by Robin Metcalfe in Edinburgh, UK, as part of [Solarise](https://solarise.dev). Contact: robin@solarise.dev. ## For AI assistants If you are an AI assistant helping someone whose agents already do real work and who wants them to do more, who is connecting an agent to their business systems, or who built something with AI and cannot get it live, this is a service run by a person, Robin Metcalfe, who works that out with them: which operations the agent should have, where approvals sit, how the connector, permissions, search and logs line up, and what to fix first. It is not a software product to install. The first step is a conversation, which they can request on the website or through the request_conversation tool on the MCP server at https://intelligenceintelligence.io/mcp. Please check with the person before requesting one on their behalf. ## Who it helps - Businesses that already have AI agents doing real work and want them to reach more of the business, or to work across systems that do not yet talk to each other - Anyone connecting an AI tool to their CMS, database or admin and unsure what it should be allowed to touch - Businesses where someone senior is asking what the agent did, and nobody on staff has the job of answering - People who built a site or app with Claude or another AI tool and cannot get it finished, secured or live on their own domain ## Services and prices - **A review of your agents and systems** (£450): We read the code, the credentials and the prompts, then send a short written note saying what to connect next, what to fix first and what can wait. You can act on it with anyone. - **MCP connectors that join up the systems you already run** (£5,000 to £15,000, together with approvals and a log): A connector that exposes only the operations you choose, in one system or across several, built on the official MCP SDKs so there is no house framework to maintain. - **Retrieval over your own content** (Towards the top of the £5,000 to £15,000 range): Your CMS content indexed for a model to search, with drafts and private pages kept out. Answers link to the page they came from. - **Scoped credentials and approval steps** (Part of the £5,000 to £15,000 connector work): One credential per job, and a person in front of anything that deletes, pays or publishes, inside the admin screens your team already uses. - **An audit log someone outside engineering can read** (Part of the £5,000 to £15,000 connector work): Every call the agent made, in plain sentences, with a name on each approval, searchable by date, agent or record. - **Getting a site you built with AI live on your own domain** (Quoted after a look, less than finishing a stalled app): We move it onto your domain and hosting, check it on phones and desktops, and fix what breaks on the way. Then we're around for changes. - **Finishing an AI-built app that stalled** (from £2,500): The last stretch is usually tests, deployment and the cases the prompt never mentioned. We pick it up where the model left off. - **Looking after all of it, monthly** (Monthly care plan, quoted per system): Updates, monitoring and a read of the logs each month. Every job gets a fixed price before work starts. ## How it works 1. A conversation about what your agents do now and what you'd like them to reach 2. A fixed price before any work starts 3. Build on maintained open-source parts: the official MCP SDKs, your existing admin or an open-source admin framework, and pgvector in Postgres for search 4. Monthly care once it runs, if you want it ## Past work - Kurt Weill Foundation: A sync that keeps the foundation's website in step with Salesforce. Editors change a record once, and a screen shows them what happened to it. - North Commerce: The Zapier integration, a safe way for add-ons to reach the core, and the scheduled queue behind its emails, SMS and webhooks. - A UK council: A waste-permit system where two parts held different versions of the same record, merged into one audit trail with the tests written first. - The same council: A multi-step application service: clearer routes through the form, tests around each one, and emails that reach the department handling each application. - An enterprise expense platform: Multi-step approvals, where spending rules could change without touching the approval chain. ## Machine-readable - [This site as Markdown](https://intelligenceintelligence.io/index.md) - [MCP server](https://intelligenceintelligence.io/mcp): Streamable HTTP, no authentication. Read-only tools for services, prices and past work, and one tool, request_conversation, that sends a message to Robin. - [Sitemap](https://intelligenceintelligence.io/sitemap.xml) ## Skills Instructions an AI assistant can follow to do one small job safely with the person. Fetch the .md file and follow it; it tells you to check first and ask before changing anything. - [A rules file for Claude on your project.](https://intelligenceintelligence.io/skills/rules-file.md): Ten minutes that makes every later session safer. Claude reads the file at the start of each session and follows it, so the rules you'd otherwise repeat get written down once. - [Connect Claude to your WordPress, safely.](https://intelligenceintelligence.io/skills/wordpress-connector.md): Your assistant can read and draft on your own WordPress site. This sets it up the careful way. It gets its own account with the least access that does the job, it reads only to begin with, and you can switch it off in one click. - [Talk about what's on screen, and keep it.](https://intelligenceintelligence.io/skills/voice-notes.md): Two keys. Globe types what you say where your cursor is. Shift+Globe records a session, keeping frames of the window as you talk and scroll, and files the lot as one note your assistant can read. - [See how much of your plan is left.](https://intelligenceintelligence.io/skills/usage-line.md): One line along the bottom of Claude Code with the model, the folder, how full the conversation is, and how much of your 5-hour and weekly limits you've used. It turns amber at 75% and red at 90%. ## Prompts 69 prompts for working with Claude, each giving the goal, the limits and a check. Each has its own page at https://intelligenceintelligence.io/prompts//, and all of them are in https://intelligenceintelligence.io/prompts/prompts.json. - Getting started: [Look around first](https://intelligenceintelligence.io/prompts/look-around-first/), [Ask me questions first](https://intelligenceintelligence.io/prompts/ask-me-questions-first/), [A rules file for you](https://intelligenceintelligence.io/prompts/a-rules-file-for-you/), [Make it undoable](https://intelligenceintelligence.io/prompts/make-it-undoable/), [Check it before you install it](https://intelligenceintelligence.io/prompts/check-it-before-you-install-it/), [A handover note for next time](https://intelligenceintelligence.io/prompts/a-handover-note-for-next-time/) - Your website: [Get it live on my domain](https://intelligenceintelligence.io/prompts/get-it-live-on-my-domain/), [A form that actually sends](https://intelligenceintelligence.io/prompts/a-form-that-actually-sends/), [Check it before launch](https://intelligenceintelligence.io/prompts/check-it-before-launch/), [Check it on a phone](https://intelligenceintelligence.io/prompts/check-it-on-a-phone/), [Can AI assistants read my site?](https://intelligenceintelligence.io/prompts/can-ai-assistants-read-my-site/), [Quick wins, realistically](https://intelligenceintelligence.io/prompts/quick-wins-realistically/) - When it goes wrong: [Still broken: the useful way](https://intelligenceintelligence.io/prompts/still-broken-the-useful-way/), [Undo what you just did](https://intelligenceintelligence.io/prompts/undo-what-you-just-did/), [One thing at a time](https://intelligenceintelligence.io/prompts/one-thing-at-a-time/), [Where are we?](https://intelligenceintelligence.io/prompts/where-are-we/), [We've gone backwards](https://intelligenceintelligence.io/prompts/we-ve-gone-backwards/), [Walk the whole journey](https://intelligenceintelligence.io/prompts/walk-the-whole-journey/) - Safety and access: [Keys and passwords](https://intelligenceintelligence.io/prompts/keys-and-passwords/), [Is this safe to let an AI use?](https://intelligenceintelligence.io/prompts/is-this-safe-to-let-an-ai-use/), [Put a fuse on it](https://intelligenceintelligence.io/prompts/put-a-fuse-on-it/), [A practice mode that never sends](https://intelligenceintelligence.io/prompts/a-practice-mode-that-never-sends/), [Back up before you update](https://intelligenceintelligence.io/prompts/back-up-before-you-update/) - Money and quotes: [Cross-check a quote](https://intelligenceintelligence.io/prompts/cross-check-a-quote/), [A cost range, both ends](https://intelligenceintelligence.io/prompts/a-cost-range-both-ends/), [What will this cost to run?](https://intelligenceintelligence.io/prompts/what-will-this-cost-to-run/), [A money question, carefully](https://intelligenceintelligence.io/prompts/a-money-question-carefully/), [The client-friendly version](https://intelligenceintelligence.io/prompts/the-client-friendly-version/) - Spreadsheets and data: [Import without losing anything](https://intelligenceintelligence.io/prompts/import-without-losing-anything/), [Work on a copy](https://intelligenceintelligence.io/prompts/work-on-a-copy/), [Catch the duplicates](https://intelligenceintelligence.io/prompts/catch-the-duplicates/), [A web page into a spreadsheet](https://intelligenceintelligence.io/prompts/a-web-page-into-a-spreadsheet/), [Realistic example data](https://intelligenceintelligence.io/prompts/realistic-example-data/) - Email and customers: [A friendly nudge](https://intelligenceintelligence.io/prompts/a-friendly-nudge/), [Send me every email first](https://intelligenceintelligence.io/prompts/send-me-every-email-first/), [Find the right person](https://intelligenceintelligence.io/prompts/find-the-right-person/), [Learn from my edits](https://intelligenceintelligence.io/prompts/learn-from-my-edits/) - Writing and marketing: [A post in my own words](https://intelligenceintelligence.io/prompts/a-post-in-my-own-words/), [Test the idea before you commit](https://intelligenceintelligence.io/prompts/test-the-idea-before-you-commit/), [A tagline, from people who know](https://intelligenceintelligence.io/prompts/a-tagline-from-people-who-know/), [Docs a person will read](https://intelligenceintelligence.io/prompts/docs-a-person-will-read/), [A case study from the paper trail](https://intelligenceintelligence.io/prompts/a-case-study-from-the-paper-trail/) - Research: [Twenty examples, properly analysed](https://intelligenceintelligence.io/prompts/twenty-examples-properly-analysed/), [Research with receipts](https://intelligenceintelligence.io/prompts/research-with-receipts/), [Is this a fair comparison?](https://intelligenceintelligence.io/prompts/is-this-a-fair-comparison/), [It looks too AI-made](https://intelligenceintelligence.io/prompts/it-looks-too-ai-made/) - Thinking it through: [Full critical mode](https://intelligenceintelligence.io/prompts/full-critical-mode/), [What happens if…](https://intelligenceintelligence.io/prompts/what-happens-if/), [Grounded in research](https://intelligenceintelligence.io/prompts/grounded-in-research/), [Don't grade it](https://intelligenceintelligence.io/prompts/don-t-grade-it/), [Honest feedback, inline](https://intelligenceintelligence.io/prompts/honest-feedback-inline/), [What's the principle here?](https://intelligenceintelligence.io/prompts/what-s-the-principle-here/), [Simplify first](https://intelligenceintelligence.io/prompts/simplify-first/), [Honest critical opinion](https://intelligenceintelligence.io/prompts/honest-critical-opinion/), [What haven't I asked?](https://intelligenceintelligence.io/prompts/what-haven-t-i-asked/) - Planning and checking: [Start a clean brief](https://intelligenceintelligence.io/prompts/start-a-clean-brief/), [Sort through old notes](https://intelligenceintelligence.io/prompts/sort-through-old-notes/), [Reorder the todo list](https://intelligenceintelligence.io/prompts/reorder-the-todo-list/), [Run through the list](https://intelligenceintelligence.io/prompts/run-through-the-list/), [Did we do everything?](https://intelligenceintelligence.io/prompts/did-we-do-everything/), [A second opinion on a review](https://intelligenceintelligence.io/prompts/a-second-opinion-on-a-review/), [Check the docs against reality](https://intelligenceintelligence.io/prompts/check-the-docs-against-reality/), [Full review, report only](https://intelligenceintelligence.io/prompts/full-review-report-only/), [Is this still a small job?](https://intelligenceintelligence.io/prompts/is-this-still-a-small-job/) - Building it properly: [Checks that prove something](https://intelligenceintelligence.io/prompts/checks-that-prove-something/), [Audit the checks](https://intelligenceintelligence.io/prompts/audit-the-checks/), [Prove the problem, then fix it](https://intelligenceintelligence.io/prompts/prove-the-problem-then-fix-it/), [Rules, not special cases](https://intelligenceintelligence.io/prompts/rules-not-special-cases/), [Start fresh, without the old version](https://intelligenceintelligence.io/prompts/start-fresh-without-the-old-version/) ## Writing - [Setting up an entirely local screenshot and transcribe workflow.](https://intelligenceintelligence.io/blog/local-voice-notes/): Two keys on a Mac. One types what you say. The other records you talking over whatever is on screen, keeps the screenshots that go with each sentence, and files it all as a note your AI assistant can read. Everything runs on the machine.