Your AI already knows your business. Now it can run your outreach.
Revenue Force exposes a secure MCP endpoint. Point Claude, ChatGPT, or Cursor at your workspace and run outreach in plain language — check a campaign, import a list, review drafts, book a meeting. Same permissions as the console. Nothing sends without your approval.
“
312 first-touch emails, drafted in your voice, queued and ready. Nothing sends until you say so.
Run your outreach by talking to it.
No new dashboard to learn. Ask your own AI, in your own words, and it does the work inside your workspace — then hands the decisions that matter back to you.
Check on campaigns
“How is the SaaS founders campaign doing?” Get live status, replies, and what’s booked — spoken back to you, without opening a dashboard.
Import contacts
“Import this sheet and enroll them.” Your AI maps the fields, dedupes against your book, and queues drafts — you just point it at the file.
Review & approve
“Show me what’s waiting.” Read every queued draft, tweak a line, and approve the batch — right there in the chat, the same queue as the console.
Book meetings
“Reply to this and find a time.” It drafts in your voice, checks your calendar, and books the invite once you approve it.
Bring the AI you already use.
Revenue Force speaks MCP over streamable HTTP at ops.inkris.ca/gw. Connect in a minute, then talk to your workspace.
Claude
In Claude Desktop, add the connector to your config and restart. On claude.ai, add a custom connector and sign in with your Inkris account — no token to copy.
Claude Code
Run one claude mcp add command in your terminal. The connection saves per project and your AI can drive outreach from the CLI.
ChatGPT
On ChatGPT Business or Enterprise, an admin adds the connector once under Settings → Connectors. After that, everyone on the workspace signs in and it appears under Apps.
Cursor
Drop a small JSON block into Cursor’s global or project mcp.json, then enable Inkris under Settings → MCP.
Any MCP client
Anything that speaks MCP over streamable HTTP works. Point it at your gateway URL with your token, or use OAuth discovery — no special integration required.
Your AI gets access. It doesn’t get control.
Pointing your AI at Revenue Force doesn’t loosen a single guardrail. It runs through the exact same permission and approval layer as a person signed into the console — enforced on our server, not in the AI.
How we keep you in controlYour AI sees exactly the tools your role allows — no more than you’d see signed in. Every call re-checks your scopes on the server. It is a window into your workspace, not a skeleton key.
Sending, publishing, launching, and spend always stage an approval and wait for a human — even when the AI holds the scope. The AI drafts and prepares; you keep the final word.
Every connection is yours: your account, your token, your workspace. Disconnect or rotate in one click and access dies immediately. It doesn’t replace your stack — it runs on it.
The questions a careful buyer asks first.
Is my whole workspace suddenly exposed to the internet?
No. The MCP endpoint is authenticated. Your AI connects with a personal token or an OAuth sign-in tied to your Inkris account, and it only ever sees your workspace — never anyone else's. Every call re-checks your permissions on the server, and you can rotate the token or disconnect in one click, which kills access immediately.
What can the AI NOT do without my approval?
Anything that leaves your workspace. Sending a message, publishing, launching a campaign, ad spend, and sensitive finance always stage an approval and wait for a human — even when the AI holds the scope to do it. The AI can draft, prepare, and queue; it can't send on its own. Nothing goes out without your say-so.
Which AI clients actually work?
Claude (Desktop and on claude.ai), Claude Code in your terminal, ChatGPT on Business or Enterprise, and Cursor — plus any client that speaks MCP over streamable HTTP. If it can add an MCP server, it can connect to Revenue Force.
Is my data used to train AI models?
No. Revenue Force does not use your workspace, contacts, or conversations to train any model. MCP simply lets the AI client you already use call your own workspace tools; your content stays in your workspace. Your AI client runs under its own provider's data terms — on the business and enterprise tiers we support, those providers do not train on your content by default.
How do I turn it off?
Disconnect in one click, or rotate your token from the connections page. The moment you do, the AI loses access — mid-task, immediately, no cleanup required. Every integration is yours to end whenever you want.
Point your AI at your pipeline.
Book a revenue audit and we’ll show you exactly how your AI would run outreach inside your workspace — scoped, approved, and under your control the whole way.