Our development team keeps putting out more and more releases, faster and faster. Part of our PCI Review process is that I get a morning report every morning with all of the releases from the previous day. It is useful for me, and assures that we do not have rogue releases going out and that each has been thru the appropriate review process. But I realized I could use AI to generate weekly summaries for different purposes that are much broader.
I finally got motivated by a webinar last week where Duy was showing samples of how to use Claude CoWork. It got me motivated to try out this idea. Here is what I did…
Github Connector
We use Github for a LOT. Every Account Manager and Sales Rep and Support person logs Issues. These are typically ideas or wishes they hear from customers. They will typically put in the customer name when they create the issue. They also will add comments to issues where there is a second, third and more customers. This helps us see the priority if a lot of customers are requesting the same thing. We have a Monday meeting where Stephen, Ryan, Allison, Bruce and I review all the new and updated issues every Monday morning. We categorize and prioritize them. We use a tagging mechanism called Milestones to categorize them – Volunteer, EMail, Website, Infrastructure, Financial, Reporting, Tickets, etc.
Developers get assigned these Issues and will do development. The next step is a review, where they create a Pull Request. The development is reviewed by at least two people (and soon a brand new code review agent), changes are made, and then it is “Merged” – meaning the feature is released to customers.
Our development and infrastructure make this easy and we will do 5-25 releases per day with zero downtime.
We have installed a Github Connector that allows my laptop Claude CoWork environment to connect with Github.
Github Merged Pull Request History
So this past Friday, I set up 84 agents to look at the Github history of all the merged Pull Requests. I set it up so there was an agent that looked at each week and collected the following information:
- Number of Releases
- Number of lines of code added and removed
- Related Issues
- Customers that are tied to those releases via the comments in the Issues
History Spreadsheet
The first thing I wanted was a spreadsheet that showed the number of pull requests and the lines of code added/removed per week so I could compare them YOY. After some trial and error, I got it to work!
This helps us see if we are becoming more productive with AI as we continue to invest in Tokens and software to produce better software for our customers.
Customer Impact
The next thing I wanted to do was generate a weekly report that could go to our AM and Sales teams to tell them the customers who were positively impacted by new releases. With 40,000 + events and 50+ releases per week it can be hard to keep track. I now have a weekly report that is generated that includes each customer and the releases that they were tagged on in the closed Issues. Here is an example of part of that weekly report:
The AM team emails each customer, and the customer knows we care!
Weekly Product Update Blog
We blog a lot, but only for major features. Some of our more minor features might be interesting to customers. And since we will now record all features moving forward, it will be simpler to search and the large LLM’s will pick the content up and have better knowledge of RunSignup when people ask questions.
I did the first one this week – https://info.runsignup.com/2026/08/18/runsignup-product-updates-week-of-aug-10-16-2026/.
We had 73 releases and over 16,000 lines of code added and over 7,000 lines of code removed (we are sticklers for continually modernizing our platform). About 35 of them were meaningful and there was a one sentence description created by Claude from reading the Pull Request and related Issue(s).
Our customers will be able to feel confident that the product they are using continues to get better and better every day, even if they do not read about all 35 issues. If they happen to be interested in our new Volunteer V2, they can follow the particulars for a couple of months and just look at the Volunteer section to see what is new.
Accountability and Responsibility
Our company tracks a lot of numbers and processes. We have to so we can scale cost effectively. We have to keep pushing forward to earn our customers loyalty, and to win new customers. Hopefully these are new reports that AI has made possible that hold us accountable and keep us feeling responsible for our customer’s success.