Cohley is a content marketing platform that helps consumer brands generate high-quality user-generated content, product reviews, and influencer campaigns at scale. Brands use the platform to create briefs, match with vetted creators, manage content production, and activate assets across paid, owned, and retail channels.
Software
51-200
NA
Customer and product data at Cohley lived in silos. It was scattered across Pendo, Snowflake, the CRM, Gong, and Freshdesk. Customer Success teams needed behavioral insights to prepare for client meetings, but only one person could pull them: Michelle Green, the Director of Product, who had become the bottleneck for every data request.
Green connected Pendo's MCP server to Dust, Cohley's internal AI tool, creating a single natural-language interface across all of their data sources. Now anyone on the CS, Sales, or Support team can ask questions like "What features is this customer not using?" or "How long does it take the customer to vet Creators?" and get answers instantly.
CSMs now prepare for client meetings in minutes instead of waiting on product to pull data. One CSM independently built a personalized "year in review" for a client, no dev time required. The team is equipped for smarter retention conversations, more targeted upsell opportunities, and proactive outreach grounded in real behavioral data.
Before Pendo's MCP server existed, there was a familiar problem at Cohley: the data existed, but almost nobody could see the full picture. Customer Success managers needed to understand how their accounts were using the product: which features they relied on, and how to predict churn. . Product and design needed to see which features moved the needle, what was saving customers time, and quickly investigate any issues. But all of that information was scattered across Pendo in individual reports, replays, guide metrics,,, Gong call recordings, and Freshdesk tickets.
Michelle Green, Cohley's Director of Product, was the only person outside of engineering who could pull it all together. She'd taught herself SQL, she knew how to navigate Pendo dashboards and session replays, and that made her the default answer to every question about customer behavior.
"Customer Success had to ask me for all customer information," Michelle says. "All data was very isolated and hard to access."
It worked, but it didn't scale. Michelle couldn't be in every meeting. She couldn't anticipate every question a CSM would need answered before a quarterly business review. And the more the company leaned on her, the more she became a bottleneck rather than a product leader.
The shift came when Michelle connected Pendo's MCP server to Dust, the AI assistant Cohley had already rolled out across its Customer Success, Sales, and Design teams. Dust was already pulling from Snowflake, Planhat, Gong, Freshdesk, and Guru. Adding Pendo meant behavioral product data (time spent on features, login frequency, rage clicks, error clicks, navigation patterns, etc.) now sat alongside everything else in a single conversational interface.
"Customer Success had to ask me for all customer information. All data was very isolated and hard to access. The MCP allows our employees to ask questions in natural language and get back custom charts. This is a game changer for upsells, retention, and ensuring smart questions are asked in every client meeting."
— Michelle Green, Director of Product, Cohley
Connecting to Pendo’s MCP unlocked a whole new way to interact with data for the org. A CSM preparing for a call could type a natural-language question and get back a custom chart showing account-level activity. They could ask which features a customer wasn't using, something that's surprisingly hard to filter for in a traditional dashboard. They could check whether a user had been searching the platform repeatedly without finding what they needed, or whether an account showed signs of curiosity about a feature that might be a candidate for upsell.
"What is hard is knowing which dashboard to look at and how to make sense of the data," Michelle says. "The MCP allows our employees to ask questions in natural language and get back custom charts to their specific preferences."
One of Michelle's favorite examples of fast MCP wins was when a customer success manager used the combined data from Pendo and Cohley's other sources to build a personalized year-in-review presentation for one of her accounts entirely on her own. It showed how many times the client had logged in, how much time they'd spent in the platform, how many assets they'd generated, and how much time they'd saved.
"As opposed to product having to build things, as opposed to using dev time to build things, a customer success manager can work that much faster to get the data into a presentation mode," Michelle says.
That kind of self-service wasn't possible before. It meant one less request routed through product, one less bottleneck, and one more client-facing team member who could tell a compelling, data-backed story in a meeting.
The broader impact is about the quality of conversations Cohley's teams now have with customers. With Pendo data flowing through the MCP, retention and upsell discussions are grounded in actual behavior instead of assumptions. CSMs can walk into a meeting knowing which features an account has ignored, where users have experienced friction, and where there might be room to expand.
Michelle also uses session replays as a daily tool, watching how users navigate the platform and reaching out proactively when she spots someone struggling. She describes one case where she watched a user search repeatedly without finding what they needed. "While data can tell me a little bit of that, watching the session just unlocks so much more," she says.
Cohley’s next challenge is adoption of this empowering tool. Michelle describes herself as the internal evangelist, sending screenshots and reminders to make sure teams know what's available. "The door is open. The data is available. Now we just have to use it."
Michelle's answer is immediate: "We'd be in the dark." Without Pendo, the team would fall back on surveys, emails, and phone calls to understand customer behavior. Pendo lets them skip the guesswork and ask more pointed questions, faster. And with the MCP connection piping that data into the tools teams already use, the insights go from living in a dashboard to showing up right where the work happens.
