We’ve heard from users that demo maintenance and analytics can be some of the more manual parts of working in Navattic, so we built a way for them to spend less time on upkeep and more time refining and polishing their demos.

The Navattic MCP server gives AI coding agents and tools – such as Claude, Cursor, and ChatGPT – structured access to your Navattic workspace.

Which means they can do everything from browsing your existing demos, seeing the top leads going through your demos, to querying visitor data and mining your analytics for helpful insights.

In this post, we’ll share four high-impact use cases to get your wheels turning, with real examples from early beta users.

Growth and demand gen managers may spend a lot of time in Navattic’s Analyze tab.

There, they can see which interactive demos are attracting visitors, how far those visitors get in each demo (and whether there’s a consistent drop-off point), and how their demos influence pipeline and revenue.

Pulling that data is straightforward – just find the right dashboard, filter, and export to CSV. But it’s still a few clicks every time.

Now, with Navattic’s MCP, they can ask Claude or ChatGPT to generate a report with a prompt, like:

• “Which of our demos got the most engagement last quarter?”
• “Pull a list of visitors from companies with 200+ employees who viewed our [insert feature] demo in the past month.”
• “Which cybersecurity accounts have the highest total engagement across our demos?”
  

See the full demo of the Navattic MCP →

In the background, the MCP is using several tools to source and compile this information:

• list_demo_analytics, which responds with demo views, visitors, engaged sessions, duration, CTA clicks, and CTR.
• list_visitors, which shows visitors who have interacted with demos, with filtering by demo, company, location, device, and custom properties.
• get_visitor, which delivers a detailed profile for a specific visitor, including their 25 most recent sessions, conversion events, and demos viewed.
• list_accounts, which provides a list of company accounts with engagement metrics, with firmographic filtering by industry, employee count, revenue, and more.
• get_account, which gives a detailed profile for a specific company account, including visitors, demos viewed, and total engagement duration.

Example: A/B Testing Demos

One beta user ran a performance report across six new tours they’ve been A/B testing. With the MCP, they surfaced who viewed each and exactly where in the demo they converted.

“Super excited to get access to the MCP and start playing around. It had no problem accessing the 6 new tours and building the report. I asked it to include a list of the Accounts that have viewed the tours.I tried another prompt that needed get_visitor, and then got step-level conversions for that visitor. I am going to continue finessing this report and then automate it quarterly!”

Example: Automating Weekly Updates

Our Head of Growth and Product Marketing, Natalie Marcotullio, uses the MCP to build her updates for Friday all-hands meetings.

“I already had a [Claude Code] Routine that scrapes Slack for the top leads that came in this week (by company size and ICP fit) and summarizes for me. I’ve added to that Routine to look at which of those leads engaged with an interactive demo and call out the demo.”

She also uses the Navattic MCP to pull the five top-performing demos of the week (along with which accounts engaged with them) and includes them in her report. Below is an example of one of those weekly reports:

Your product is always changing, which means your demos have to change with it.

Remembering to update your demos – and setting aside time to actually make those updates – is tough, particularly if you have a large library.

SEs and PMMs may track this in Navattic with Boards and Labels or in a separate spreadsheet or sales enablement platform.

But now, they can use AI to scan their Navattic workspace for demos with the oldest publish dates and prioritize refreshes from there.

For demo refresh prompts, the MCP uses:

• list_projects, which finds all the projects in your workspace
• get_project, which pulls the details, flows, and links for each demo (including last published date)

Example: Refreshing High-Traffic Demos

“If you build a lot of demos, you’re naturally going to run into situations where they become outdated and inaccurate, but still get traffic,” says Jason Oakley, Founder of DemoDash.

“These are perfect candidates for a simple refresh, to make sure that website visitors or prospects aren't getting shown the wrong thing, or the wrong messaging.”

As part of his beta test, he’s hooking up Navattic’s MCP to Claude and asking:

“Which of my interactive demos are consistently getting traffic, but haven't been updated in more than 60 days?”

Claude suggests which ones to update first (high traffic, longer time to update).

He advises combining those results with your judgment – for example, maybe you know that certain demos feature parts of your product that recently underwent UI changes – to nail down which ones need immediate attention.

Example: Building a Digital Asset Management System

Another beta user is planning to build an entire DAM (digital asset management) system, stitching together data from Jira, Figma, and the Navattic MCP to flag which demos need a refresh when a new feature ships.

“We use Navattic primarily for enablement, and our library of demos is gigantic,” they explain.

“Whenever our product is updated, we need to update a ton of content, even besides just our demos. It’s very daunting keeping track of what content touches what parts of our product and to determine what exactly needs to be updated.”

To overcome this problem, they’re using AI to analyze videos, images, courses, and interactive demos to summarize what happens in each one, what it looks like, and which UI elements it features.

That way, when a product update is slated to happen, an agent can:

1. Grab relevant information from Jira, Figma mockups, screenshots, and docs.
2. Scan through the database to identify what content needs to be updated.
3. Give suggestions on how and where to update it.

Because the Navattic MCP provides the last published date, agents can triage the demo backlog, ensuring the team updates the oldest demos first.

“I tested the MCP out last week, and it’s going to work great for the agentic system I am looking to build.”

Some demos just knock it out of the park, keeping visitors engaged and prompting them to book a call with sales.

Those are the demos you want every other demo to emulate.

But new demo scripts usually get written from scratch or from memory, without a systematic way to learn from what’s already working well.

Now, demo builders can use the MCP to prompt their LLM of choice to pull top-performing demo flows and use their structure as a template.

For demo scripting prompts, the MCP uses:

• list_projects, which finds all the projects in your workspace
• get_project, which pulls the details, flows, and links for each demo

Example: Automatically Drafting Internal Demo Builds

At Navattic’s last offsite, Natalie worked with a few team members on a Claude Code Routine.

This one turns a Linear project into an optimized interactive demo script, using the structure from top-performing demos and weaving in best practices from our State of the Interactive Product Demo report.

To request an internal demo build, an engineer can now:

• Add a tag to an existing Project in Linear (our product planning tool).
• That will automatically create a new Linear ticket for Navattic’s internal demo builders.
• Claude will pull from that Linear ticket and write a first draft demo script based on the ticket context and the style, tone, and format of Navattic’s top demos.
• Claude will paste the script into the Linear ticket.

Navattic demos capture high-quality intent signals your AEs can use to tailor their outreach – if they remember to look at it.

With Navattic’s MCP, you can pull visitor activity from target accounts directly in Claude or another LLM.

For intent-related prompts, the MCP uses:

• list_visitors, which shows visitors who have interacted with demos, with filtering by demo, company, location, device, and custom properties.
• list_accounts, which provides a list of company accounts with engagement metrics, with firmographic filtering by industry, employee count, revenue, and more.
• get_account, which gives a detailed profile for a specific company account, including visitors, demos viewed, and total engagement duration.

Example: AE Alerts in Slack

Nick Bennett, a 6x Navattic user, is using Navattic’s MCP to set up a Slack alert for AEs.

“An AI agent pulls visitor activity from target accounts. Filters for high-intent behavior. Pushes an alert to Slack with context before the signal goes cold.”

For his workflow, he’s counting “high-intent” behavior as:

• Viewed pricing or integration flow
• Returned to the same demo twice in 7 days
• Multiple people from the same company viewing different flows
• Spent 3+ minutes in a product walkthrough

The alerts themselves include the visitor’s name, role, and company, the demos they viewed, how many steps they got through in each demo, and any CTAs they clicked.

This gives reps a sense of the use cases the prospect is focused on so they can reach out with the right message as soon as a lead expresses interest.

Example: Automatic List Building

Saad Khan, Director of GTM Engineering at trumpet, is accelerating pipeline with the Navattic MCP.

He’s using Claude to:

1. Determine who is visiting an interactive tour, and what pages they viewed
2. Push visitors from target accounts to a Skill that enriches those contacts with Apollo data
3. Drop those contacts into Clay for a live enrichment score, OR push them directly to Apollo’s sequencer to take action on that day

“That’s where I see the most joy with MCP,” he shares.

“Combining and stacking multiple different data sources and signal sources to generate ‘warm’ pipeline. Engagement is some level of awareness, and awareness is some level of intent.”

Got your own use case in mind? See a demo of the MCP and start building.