Dave Killeen: [00:00:00] Hello, hello. Dave Killeen here, Field CPO, Pendo in EMEA. And a very warm welcome back to episode nine of the Vibe PM where every two weeks or so I'll be showing you different ways of dancing with AI so that you can have more fun and more importantly, more impact with your work. Today, I want to talk about MCP, why it's everywhere right now, why it matters, and what it looks like when you put it to work beyond just having a chat back and forth.

Dave Killeen: So, let's jump in. And the reason MCP is so popular is not just that your AI can now read the data, it's what happens when you start combining different data sources together into a single conversation. With your AI. You can pull from most of the tools you typically have access today in your day-to-day job, mash it all together and get insights that no single tool could give you on its own.

Dave Killeen: And as you'll see at the end of this episode, you can then have those insights rooted into other tools, like generating slide decks for exec updates, which I'll show you at the end of this session. And the barrier [00:01:00] to entry is so much lower than traditional API integration monarchy. You can connect an MCP server in minutes and your AI can start using it right away.

Dave Killeen: And the one I use unsurprisingly is the Pendo MCP as it gives my AI direct access to customer product data. Engagement scores, customer feedback segment, filtering, and much, much more. Let me show you what that looks like in practice. Four quick examples, using dummy data in one long chat that results in a crisp and clean generation of an exec friendly deck for an upcoming quarterly.

Dave Killeen: Business review. Let's jump in. So first up, I've got a team review with the customer account called Dharma Initiative, and I want to walk in prepared, so when I pull in the top, I've got a QBR with them in 30 minutes, pull their account activity for the last 30 days, their product management score, who their key users are, and any feedback that they submitted.

Dave Killeen: And pull all this from the demo. Pendo experience sandbox. So if I just click on this here, what you see is it's now looking to pull this from the sandbox, looking up the application itself. Once it's [00:02:00] found the application, then it's gonna then pull on a whole bunch of other tools, then to then do all of the aggregation and do all the querying.

Dave Killeen: Once that comes through, then we see what we have here. And so the product engagement score is something we have here at Pendo that talks about the overall health of a particular account on based on a measure of adoption, stickiness and growth. And concerning here, growth is zero, so there's no new user expansion.

Dave Killeen: And that's my QBR conversation right there. And what's great is the AI's pulling together this insight for you based on what it can glean. This isn't coming from the Pendo MCP server. This is the AI getting access to the data and then using its own judgment on the situation. It's identifying the key champions as pulling out the feedback that we're seeing.

Dave Killeen: And then giving me the talking points for today's conversation with the team. And it's calling out that Jerry and Jessica are champions and that we should be really orientating the conversation around their use cases. So next up is Monday morning, and we need to find out which accounts need my attention this week.

Dave Killeen: So I will ask the AI to show me the [00:03:00] top 10 accounts by activity over the last 30 days, and then separately show me which accounts have the highest frustration signals. So rage, clicks, dead clicks, U-turns, all good proxy indicators for people not having a good experience. And then it pulls out the top 10 accounts by frustration signal for the last 30 days.

Dave Killeen: And what it's basically saying here is that none of these accounts overlap with top activity accounts, which is a healthy sign. Most active users aren't the most frustrated ones, but these ones here, empire Industries, JetLife, and Hyperion would be good proactive outreach candidates before frustration turns into a churn risk for these smaller accounts.

Dave Killeen: There's clearly an onboarding problem with those accounts. And that's just one prompt, my Monday morning radar all done in a matter of minutes. Next up, I want to understand what customers across the entire base are actually saying. And so I ask the ai, what are the main themes across all of our customer feedback cluster into topics for me, and then give me the raw insights with direct quotes.

Dave Killeen: AI then pulls together 30 clusters, rolled up into eight different themes that you see here. I won't go through all of that. In the interest of time, what again, what's [00:04:00] happening here is the AI is forming its own judgment on the data and saying that these sarcasm signals such as congratulations on making it more convoluted and another half baked feature.

Dave Killeen: They're not just complaints, they're trust erosion. According to Claude, these users are very close to being checked out, which is a very different kind of signal to a feature request. That's the kind of synthesis that would take a team a lot longer than what I've just shown you, and it just appeared in my chat window.

Dave Killeen: Alright, so everything I've shown you so far, the meeting prep, the portfolio policy feedback themes, that's the AI reading data, pulling it in, synthesizing it. Now I wanna ask all to take everything we just explored and create for me some slides. So, I'm asking you to take the key insights from the conversation attached, the feedback themes, the accounts, and your attention.

Dave Killeen: All that governs right? Create me a QBR deck that I can present to my exec. That's three slides and title it, Q1 2026 product Health Pendo World. Okay, so now what it's gonna do, and this is quite interesting, what you'll see here, it's gonna now use Anthropics built-in slides generator, and we'll have [00:05:00] a default crappy looking slides in a second.

Dave Killeen: I'll show it to you in a sec, but we'll build it from there. So all that will come through and then I will get then following, which I'll just open up. This is what it's come up with, big corporate feedback, theme analysis, and a summary. But I just wanna make it a bit more Pendo, so let me show you this bit.

Dave Killeen: I'm like, okay, that's great. Can you create me another version of that? Just using the following style guide. So I'll give it a whole bunch of style guides. I do that by just giving a handful of our slides into Claude, into the AI can be Gemini, whatever. And then having it come up with a style guide that it thinks represents what it's just received.

Dave Killeen: And then just like a cookery program, we take it outta the oven, and here we have an updated version. There it is. Three slides, branded and formatted, 30 clusters, 3,500 mentions and three priority actions. And then we have the voice of the customer with the six themes laid out as cards. And then on the third slide we have a priority actions, which are to number one, triage the notes feature, and then fix search, it's a P zero, and then stop that trust version issue before it hits renewals.

Dave Killeen: Crazy, right? I just had a conversation with my [00:06:00] AI. We explored the data, well it did, and then said, Hey, can you make me a deck? And that's it. That's MCP data in synthesis, and then output all in one conversation. So a quick recap. MCP gives your AI direct access to real data. No more copy and pasting between tools.

Dave Killeen: No more being the person stitching everything together manually, like a headless chicken. The real value is in what you do with it. Meeting prep, in seconds, portfolio health from one question, feedback synthesis across hundreds of accounts, and then turning all of that into something you can then present right there in the same conversation.

Dave Killeen: If you want to see what MCP servers are out there, a great place to start is smithery.ai. It's an MCP directory that shows you trending servers, and you can see what others are finding useful and get ideas for your own setup. And if you're using Pendo, be sure to ask our team about the Pendo MCP. It's changed how I work and we work here at Pendo every day.[00:07:00] 

Dave Killeen: So many thanks for joining me on episode nine of The Vibe PM. Please spread the vibes and I'll see you next time. Thank you.