Every sales engagement platform now claims to support Model Context Protocol (MCP). But having an MCP doesn't automatically mean your AI can do meaningful work.

If you're evaluating AI-powered sales platforms, including lemlist, this guide explains what actually matters and why the depth of an MCP can dramatically change what your AI can do.

If you're short on time, here's the key takeaway:

Over the last few months, nearly every sales engagement platform has announced support for Model Context Protocol (MCP).

Apollo has one.

Outreach has one. Salesloft has one. lemlist has one.

At first glance, that sounds like feature parity.

But saying "We have an MCP" is quickly becoming the new "We're AI-powered."

It's a checkbox.

Not a useful buying criterion.

If you're choosing a sales platform because you want Claude, ChatGPT, Cursor, or another AI assistant actually to run outbound work, you need to look much more deeply.

Model Context Protocol (MCP) is a standard that lets AI models securely interact with software.

Instead of simply answering questions, AI can use tools inside an application.

That means your AI assistant can potentially:

But here's the important part:

Every software vendor decides which tools are available through that connection.

That's why two products can both claim "MCP support" while delivering completely different AI experiences.

Imagine asking your AI assistant "Remove every lead who bounced last week."

Sounds like one simple request.

But, behind the scenes, AI needs to chain together multiple capabilities:

Search campaign leads -> Filter by bounce status -> Delete matching leads.

Miss one tool…and the workflow breaks.

There isn't a better prompt to fix this, or a more powerful model that magically invents missing capabilities.

The limitation isn't Claude or ChatGPT, but the MCP.

This is why we believe an MCP is only as powerful as the tools it exposes.

Because tools define what AI can actually do inside a product.

One of the biggest differences between sales platforms isn't the protocol itself. It's the type of tools they expose.

Many current MCP implementations are largely read-only.

That means AI can:

✅ List campaigns
✅ Fetch leads
✅ Retrieve reports
✅ Show analytics

Useful? Absolutely.

But that's closer to a reporting dashboard than an autonomous sales agent.

An action-oriented MCP goes much further.

AI can:

As one industry expert commented on the original discussion:

At the time of writing, publicly available MCP documentation shows roughly:

The numbers alone don't tell the whole story.

Some commenters argued that exposing fewer, higher-level tools can reduce context and improve efficiency for LLMs. That's a valid engineering discussion.

But buyers shouldn't optimize for tool count.

They should optimize for workflow completion.

The real question is:

Here's an easy way to evaluate any sales MCP.

Ask your AI: "Remove everyone who bounced last week, replace them with verified VP Marketing contacts from French SaaS companies, then relaunch the campaign."

That single prompt requires AI to:

If the AI stops halfway through...

...you've learned something important.

The limitation wasn't the AI model.

It was the capabilities exposed through the MCP.

When building the lemlist MCP, our goal wasn't simply to expose API endpoints.

It was to expose the capabilities required to complete real outbound jobs.

Today, the lemlist MCP includes around 100 tools spanning:

This allows AI clients like Claude, Cursor, and ChatGPT to do far more than retrieve information, they can help build, manage, optimize, and execute outbound workflows.

Sensitive actions such as launching campaigns, deleting records, or spending credits still require explicit user confirmation.

The objective isn't to maximize the number of tools. It's to expose enough of the platform for AI to complete meaningful work.

One of the most interesting discussions sparked by the original LinkedIn post wasn't about technology.

It was about trust.

Anyone can create an impressive AI demo, claim to support MCP, or add "AI Agent" to a product page.

But buyers are increasingly asking a different question:

Several industry leaders described this emerging gap as everything from "MCP washing" to the next trust crisis in B2B software, where marketing claims outpace real product capabilities.

As AI becomes table stakes, execution, not messaging, will become the real differentiator.

If you're evaluating lemlist alongside Apollo, Outreach, Salesloft, or another platform, don't stop at "Does it support MCP?"

Ask instead:

Those questions will tell you far more than any feature badge.

Because in the end, your sales team doesn't buy an MCP.

They buy the ability to say: "Build me a campaign targeting Series A fintech founders in Germany."

…and have AI do the work.

If you want to experience end-to-end AI-powered outbound for yourself, connect the lemlist MCP to Claude, ChatGPT, or Cursor and see what you can build in minutes.