“We were able to go and get a quick fix in there with engineering, and no support tickets or cases were raised. [Support] didn’t even know about it. Hopefully, customers didn’t know about it. But LogRocket was able to surface that up, and we were able to fix it before it became a problem.”
– Elliot, Senior Product Manager, Software Industry
Elliot is a senior product manager in the software industry. He and his team were frustrated by data scattered across internal databases, the data warehouse, the CRM, and various other applications.
In response, he built Otto: an internal product intelligence platform that helps the team catch bugs and track feature adoption automatically. Otto connects to the LogRocket MCP, which makes LogRocket’s user experience data (sessions, issues, metrics, and AI-powered insights) queryable by any AI agent or internal tool.
He shared a little bit about how Otto came to be, how it works, and how it’s helping his team be more proactive in fixing bugs and user issues.
LogRocket combines session replay, product analytics, and error tracking, empowering software teams to create the ideal product experience. The LogRocket MCP connects your agents directly to LogRocket’s Galileo AI, which is already watching session replays, listening to customer calls, reading support tickets, and tracking product changes.
New to the LogRocket MCP? Check out our launch post for a deeper dive.
We have all these different data sources: data warehouses, internal databases, CRMs, homegrown tools, applications, etc. Trying to find any signals is like finding a needle in a needle stack.
There was no way to coalesce all of it together to drive insights or look at how the business is doing as a whole.
I built Otto for that reason.
LogRocket is like the mesh fabric that gives it shape on top, at least for the interface.
It lets us split users into different segments and cohorts depending on a bunch of factors. Finding signals within all of that noise is incredibly useful.
The MCP helps us be proactive instead of reactive to some of these issues.
Every Monday morning, I have a script that runs and hits the LogRocket MCP.
I have a few different natural language prompts in a file that it reads. We have five or six different data sources coming in.
When I come in Monday morning, LogRocket has already run its script. It goes and looks at all these different signals and data sources, and tries to find context and insights we can surface, whether that’s a bug, a support issue, or tracking feature adoption to see if folks are abandoning a flow or dropping off for some reason.
This past week, I came in Monday morning and went to check Otto. We found there was a higher-than-normal rage click count: very anomalous, in a place where we weren’t expecting to see something like that.
I also have an automation set up so it creates a Jira ticket and puts it on a specific board where I can review it and decide whether it needs engineering, or it’s a support thing, or it’s just a false positive.
This one was not a false positive. I looked into the Jira ticket, and the LogRocket MCP also provides session recordings directly in the Jira ticket, so I don’t have to dig around looking for them, and it describes the issue. When I watched the session replay, we found there was a technical error happening intermittently, depending on some factors of the technology the user was using.
So before customers were even really affected (it was Monday morning; we were just looking into it, no support tickets had been raised yet), we were able to make the fix quickly and be proactive.
One thing we’re looking at constantly is trial conversions and how different cohorts convert.
We noticed our trial conversion was trending in the wrong direction for a few months. I came in Monday morning and looked at Otto, and we had a clear description of what was happening, with numbers and the sessions to actually go watch, so it’s not just a wall of text.
We found folks were abandoning after a certain step. It was clunky and wasn’t loading correctly. LogRocket surfaced that for us; we put in a fix, and now trial numbers are back where they need to be and heading in the right direction.
Finding, identifying, triaging, and fixing something used to take days for me, a support person, or an engineer. Now it takes hours.
It’s hugely beneficial not just to me as a product manager, but to support, engineering, sales ops, you name it. I’m excited to see where we can keep building integrations with this.
The biggest thing for me is that this is another tool in the toolbox, and it runs automatically; I don’t have to think about it. Every Monday morning I come in, and there’s a new tranche of insights and findings to work through.
It’s made a huge difference in the short time we’ve been using it.
LogRocket identifies friction points in the user experience so you can make informed decisions about product and design changes that must happen to hit your goals.
With LogRocket, you can understand the scope of the issues affecting your product and prioritize the changes that need to be made. LogRocket simplifies workflows by allowing Engineering, Product, UX, and Design teams to work from the same data as you, eliminating any confusion about what needs to be done.
Get your teams on the same page — try LogRocket today.
Learn how to read a P&L statement, assess product costs, build stronger business cases and align product roadmaps with financial priorities.
Learn how product managers can design products for both humans and AI agents with reliable APIs, permissions, errors, and workflows.
Learn how much AI fluency product managers need, which skills matter most, and why human judgment remains their competitive advantage.
Add a harm score to product prioritization to catch severe user consequences that RICE and other frameworks can miss.
Learn more →
