Short answer: no, an AI prospecting agent from Formula. does not plug into your CRM. It already prospects on its own: it picks its own targets, sends the messages, and you get the replies on your phone.
The MCP server connects the agent itself, not your CRM or your list. From Claude, ChatGPT or Gemini, you edit personas and messages and check KPIs, with no technical setup required.
Why do so many people assume they need to connect their CRM?
Because that's how the tools we've used for the past decade work. A semi-automatic sequencer needs to read your contact list to function, so you connect it to your CRM or import a file.
The word "agent" naturally triggers the same mental model: a robot that would dig through your CRM, write from your account, and would therefore need technical setup before the first message goes out.
That's not how Formula. works. The agent doesn't read your CRM because it doesn't need to: it already has its own targeting logic and its own list of qualified prospects.
What does Formula.'s MCP actually connect?
The MCP server (Model Context Protocol) connects Formula. itself to Claude, ChatGPT or Gemini. Not your CRM, not your prospect list, not your LinkedIn account directly.
Once the MCP is set up, from a simple chat you can:
- edit your prospecting personas (the targeted profile, the message angle);
- adjust your prospecting messages without going back into the Formula. interface;
- check your KPIs (invitations sent, reply rate, meetings booked);
- get a personalized dashboard on your numbers, built automatically.
None of that requires touching your CRM. The agent is already prospecting before you even open Claude or ChatGPT.
| What you'd expect | What actually happens |
|---|---|
| "I connect my CRM to the AI agent" | Formula.'s agent already has its own list of qualified prospects, nothing to connect |
| "The MCP gives access to my CRM data" | The MCP connects the agent to Claude/ChatGPT/Gemini, to control personas, messages and KPIs |
| "A technical integration is required before starting" | The trial starts with zero setup: the agent prospects from day one |
| "The Claude agent prospects for me" | Formula. does the prospecting; Claude is only used to steer the agent from a chat |
Do I still need to keep my CRM?
Yes, if you use one to track clients you've already signed. The agent handles the upstream work: finding the right prospects, starting the conversation, booking the meeting.
What happens after the meeting is booked (quotes, billing, client relationship follow-up) stays in your existing CRM, with no required technical link between the two tools.
That's exactly why a consultant or small business with no CRM at all can start with Formula. the same day. The prerequisite isn't a tool, it's an offer already validated manually: see our breakdown for solo consultants.
Why do semi-automatic tools need a CRM in the first place?
Because they're tools, not agents. An email sequencer or a LinkedIn automation tool needs to know who to write to, so it reads your list, imported from your CRM or a spreadsheet.
Manually sorting lists is already the task that eats the most sales time. According to Salesforce, sales reps spend less than 30% of their time actually selling (Salesforce).
An autonomous agent removes that manual sorting step entirely: it qualifies and targets on its own, without depending on a list you have to maintain. The qualification mechanism is broken down in this article.
Real case: Mike and the checkbox that wasn't there
Mike runs a digital strategy consultancy. Before his Formula. trial, he spent a week comparing tools, convinced he'd first need to "connect his CRM" for the agent to work.
He didn't even have a CRM in place, just a spreadsheet. He almost delayed signing up thinking he'd need to set one up first.
In reality, the agent started prospecting the day he signed up, with nothing to connect. The first replies arrived before he'd even picked a CRM.
Same mechanism on Formula.'s own account: 283 first messages generated 114 replies, a 40% reply rate (data as of July 23, 2026), with zero CRM integration involved. See the full case study.
The MCP guide: running your agent from Claude, ChatGPT or Gemini
Sign up and we'll send you the MCP setup guide, with the exact commands we use internally to adjust personas and messages from a simple chat.
How do you get started with no technical integration?
There's nothing to integrate before the first message goes out. Three steps: describe your offer and ICP, validate a first message template, launch the agent.
The agent starts sending up to 28 LinkedIn invitations a day, in parallel by email on the same qualified prospects, with none of your data needing to be connected anywhere.
The only step that stays human is the swipe: when a prospect replies, the agent drafts the reply, and you approve it with a swipe on your phone.
The agent prospects. You collect the meetings on your calendar.
Ready to see an agent run without asking for a single CRM integration?
Start the 2-week trialNo CRM, no integration, no setup call required to see the first replies come in.
Frequently asked questions about AI prospecting agents and CRMs
Does an AI prospecting agent need to connect to my CRM?
No. Formula.'s agent does not connect to your CRM or your existing prospect list: it already prospects on its own, choosing its own targets and sending messages without any of your data being plugged in.
The MCP server exists only to control the agent (personas, messages, KPIs) from Claude, ChatGPT or Gemini.
What does Formula.'s MCP server actually connect?
The MCP connects Formula. itself, not your CRM. From Claude, ChatGPT or Gemini, you can edit your prospecting personas and messages, check your KPIs, and get a personalized dashboard built automatically, with no technical setup on your side.
Do I still need to keep my CRM if I use an AI prospecting agent?
Yes, if you use one to track signed clients. The agent handles prospecting and delivers replies to your phone; what happens after a meeting is booked stays in your existing CRM, with no required technical link between the two tools.
Why do so many prospecting tools require heavy CRM integration?
Because they are semi-automatic tools: they help you send messages, but you still qualify and manage the list yourself, so they need to read and write to your CRM.
An autonomous agent like Formula. does that work itself, so it doesn't need access to your CRM to run.