Google Drive integration with lemlist through MCP, with prompts to copy

A rep is about to launch a competitive campaign. The battlecard that should shape every line of it sits in a Drive folder three clicks away, open in another tab, read by nobody who writes the sequence. The case study that would prove the claim lives in the same folder. Nothing connects the two, so the campaign goes out generic, and the research gathers dust.

This article shows you how to fix that with a google drive lemlist integration built through MCP, plus the exact prompts to copy into Claude, ChatGPT, or Cursor. You’ll see how an agent reads a battlecard, a deck, or a case study straight from Drive and turns it into working campaign steps inside lemlist. No exporting, no manual rewriting, no copy-paste between tools. Just the prompts, what the agent does with them, and what to check before you let any of it touch a full list.

Why connect lemlist and Google Drive through MCP

Your team already has the raw material for great outreach sitting in Google Drive. Case studies, competitive battlecards, call notes, pitch decks. The problem is that none of it reaches your sequences unless someone manually copies, rewrites, and pastes it into each campaign.

An AI agent with access to both the Google Drive MCP and the lemlist MCP can read those documents and turn them into personalized copy inside your campaigns in a single conversation. It pulls the source material from Drive, writes or rewrites the messaging, and pushes it into lemlist without you switching tabs or reformatting a thing.

An MCP server is a standardized way for an AI agent to connect to an external tool and take actions inside it. Each server exposes a set of operations the agent can call, like reading a file or creating a campaign step. When two MCP servers are connected in the same client, the agent can chain actions across both tools in one prompt. You can read more about how this works for sales teams in the MCP for sales guide.

What you need before you start

Three things to have in place:

• •

  The lemlist MCP server connected in Claude, ChatGPT, or Cursor. Follow the steps in how to set up the lemlist MCP if you haven’t done this yet.

• •

  The Google Drive MCP server connected in the same AI client, so the agent can access both tools in one conversation.

• •

  Permissions scoped to what each use case actually needs. Grant read access to the specific Drive folders the agent will reference, and limit lemlist permissions to the campaigns you want it to touch.

Start with read-only prompts. Let the agent pull data and draft copy for your review before you give it permission to write directly into either tool.

1. Add proof points by industry

Most teams have a handful of case studies but send the same generic proof point to every prospect regardless of vertical. A fintech buyer and a logistics VP get the same line about “helping companies grow,” and neither one finds it relevant.

The agent opens your case study folder in Drive, reads each document, and extracts the key result per industry. It then writes a short proof-point line for each vertical and inserts them as a Liquid conditional block in the first step of your campaign. That Liquid block serves each lead the proof point that matches their industry, so the email feels specific without you writing dozens of variants by hand.

This kind of dynamic personalization matters. lemlist internal data shows that campaigns using Liquid syntax reach a 5.89% reply rate, compared to 3.57% for campaigns using only basic variables. Add spintax for cold email on top of Liquid and that number climbs to 7.94%.

Before you run this on a full list, preview the Liquid block on five leads from different industries to confirm the right proof point renders for each one.

2. Write a sequence from a battlecard

Your marketing team spent weeks building a battlecard against a competitor. It covers positioning gaps, feature comparisons, and objection handling. Meanwhile, your competitive displacement campaign uses none of it because nobody had time to translate the battlecard into sequence copy.

The agent opens the battlecard, identifies the strongest differentiators and the objections your team hears most, and drafts a three-step sequence built around those points. It proposes the sequence for your review first. Once you approve the messaging, it creates the campaign in lemlist with the steps, timing, and channel mix ready to go.

Each step can span email, LinkedIn, or calls within the same campaign, so the sequence matches how your reps actually sell against that competitor. You can also layer in persona-based messaging to adjust the angle depending on whether the lead is a user, a manager, or an executive.

Before you launch to a full list, send the sequence to 10 leads and check that the competitive claims still hold and that the tone fits your brand.

3. Keep messaging in sync with the latest deck

Your pitch deck gets updated every quarter. New pricing, new positioning, a new customer logo on slide four. But the sequences that went live last month still reference the old claims, and nobody catches it until a prospect replies with confusion.

The agent reads the current version of your sales deck in Drive, then pulls the copy from every step of your active campaigns in lemlist. It compares the two and returns a list of steps where the messaging conflicts with the deck, along with what changed. You get a clear audit without opening each campaign manually.

This is an audit prompt, so the agent reads but does not rewrite anything on its own. Review the list it returns, decide which steps to update, and then ask it to rewrite those specific steps if the suggested changes look right.

Before applying edits across all campaigns, update one campaign first and confirm the new copy reads well in context.

Guardrails when an agent works in lemlist and Google Drive

Giving an agent access to two tools means you need clear boundaries. Follow these:

• •

  Review every campaign before launch. The agent can create and update campaigns, but a human should approve the final version before any lead receives a message.

• •

  Test on 10 leads first. Run every new sequence or edit on a small batch and check replies before scaling to your full list.

• •

  Keep unsubscribes and exclusions on. lemlist enforces these automatically, but verify they are active on every campaign the agent creates.

• •

  Limit who on the team can connect the MCP servers. Treat MCP access like admin access. Only the people who understand the workflows should have it.

• •

  Log what the agent changed. After each session, ask the agent to summarize every action it took in lemlist and every file it read in Drive. Save that summary so you have an audit trail.

These guardrails apply to every prompt in this article. The agent is fast, but speed without review is how you send a battlecard claim to the wrong segment or overwrite a step that was performing well.

Run it in lemlist

The lemlist MCP server works in Claude, ChatGPT, and Cursor. Once connected, the agent can do more than just edit campaign copy.

It can search the People database with 650M+ contacts and 65M+ companies to find leads that match your ICP. It can enrich those leads with waterfall enrichment across 14+ providers to get verified emails and phone numbers. It can create campaigns and sequences across email, LinkedIn, calls, and WhatsApp. It can read your unified inbox and campaign stats to surface what is working. And it can act on intent signals to trigger outreach when a prospect is most likely to engage.

That means the same conversation where you pull a battlecard from Drive and build a sequence can also find the right accounts, enrich them, and add them to the campaign. You can explore more workflows like this in the lemlist MCP use cases guide.

Final takeaway

Your Google Drive already holds the context your outreach needs. Connecting the Google Drive MCP and the lemlist MCP in the same AI client lets an agent read that context and act on it inside your campaigns, without exports, reformatting, or tab switching. Start with the read-only prompts in this article, test on a small batch, and scale once the output looks right.

Build your first signal-based sequence in lemlist.