The challenge
Before he was a PM, Nick Meisenheimer led Sentry's technical support team. His job was routing customer friction to the right people at the right time, and he learned how badly that scales. "A support agent looks at a hundred cases a week and escalates one or two. Everything else lives in a grey area. If you're relying on humans to annotate, label, and escalate, it's never going to work — or you end up with a company full of people whose job is to label data."
When he moved into product, he inherited the other side of that problem. Sentry's customers talk everywhere: GitHub Issues, Discord, Intercom, Reddit, in-app feedback. Each channel piped into its own Slack feed, but nothing connected them and nothing alerted.
"You basically had to hawk all of those channels, all of the time. Worst case, you hear it from your boss: how come you didn't know about this?"
Quieter moments weren't better. Scoping a new product area meant reconstructing the customer pain from scratch — a full day of searching separate tools, some of which Nick couldn't access directly. Sentry had automated one channel by syncing GitHub Issues into Linear. The PM team was responsible for all of them.
Why Modem
The obvious fixes were more humans or another aggregation tool, and both create a new place to manage. Modem does neither. It connects the channels where feedback arrives to the places the team already works, and closes the loop back in the original channel.
"It's connective tissue. It's not aggregating things into a third space that you resolve somewhere else. It's where I already am in my workflow."
The setup cost is low and the payoff compounds: "Completing the graph is the whole point. Once everything's in there, the automations just work. I don't have to annotate the data; Modem abstracts that away. It just turns into succinct problem statements that could be from anywhere."
See what your customers are already telling you
Connect the channels you have — GitHub, Discord, Intercom, Slack — and Modem builds the graph.
Modem in action: monitoring the Monitors & Alerts launch
When Sentry shipped Monitors and Alerts, Nick set up a Modem automation to summarize the previous 24 hours of customer feedback on the feature. It took one prompt to create and one more to tune. He read the report every morning for a month, and three kinds of signal surfaced:
- Bugs, fixed immediately and prioritized by how much they hurt customers, not just how often they fired
- Confusion and friction, which shaped documentation and onboarding investments
- Long-tail requests, which built the case for what to build next
Sentry's own error monitoring showed where users hit exceptions. Modem showed what those errors were blocking and how customers felt about it. "Getting that in real time helps you understand what to prioritize, because you learn which bugs make people feel bad." That gave engineers a PM who could rank the work with evidence.
"It's another way to ship with confidence. You don't have to schedule a planning session to figure out what to fix next week."
The results
Research that used to take three or four hours is now a ten-minute task. "Just the fact that you're one prompt away from an exploration of user pain always feels like magic. It's made me better at my job."
The bigger long-term win is closing the loop. When an engineer fixes a bug, Modem surfaces the other customers who reported the same thing — often five to ten touchpoints across channels, some filed years earlier. New engineers don't need fifteen years of institutional memory, and PMs get a warm reason to re-engage customers.
"What's a better conversation opener than 'we solved your problem'? You focus on solving the problem, and Modem connects you to the opportunity."
Close the product gap
Connect the channels where your customers already talk, and start every week one prompt away from what matters. Tell us what you're working with and pick a time.
Prefer it as a document? Download the case study PDF.