How AllFly turned its backlog from a bottleneck into a background process
AllFly's engineering team hands roughly 80% of backlog items to Pebble the moment they're reported — doubling the volume of issues they clear and ending the dedicated backlog sprints.
AllFly helps companies manage group travel, which means their software needs to keep up with a business that never really slows down. Nate leads engineering for a lean team where developer bandwidth is always at a premium and where the backlog used to be a chronic source of friction.
The problem wasn't priorities. It was throughput. Minor bugs, small quality-of-life improvements, and low-lift production issues were piling up faster than the team could address them. Every few cycles, AllFly would dedicate an entire one- to two-week sprint just to chipping away at the backlog — pulling developers off feature work to do it.
That changed with Pebble.
Backlog: from bottleneck to background process
Now, roughly 80% of backlog items get handed to Pebble the moment they're reported. If it's a clear, low-lift issue — typically anything in the zero-to-three point range — Nate doesn't ask when it can be scheduled. He just sends it to Pebble.
The result: developers stay on feature work. The backlog moves in parallel. And the dedicated "backlog sprints" that used to eat up weeks? Gone.
Internal teams noticed immediately. Sales and ops, who used to navigate a slow back-and-forth about whether something was a priority and when it might get fixed, now hear a simple answer: next release. Every two weeks, AllFly ships a meaningful set of resolved issues — and customers feel it.
We found that a good 80% of backlog items can be handed off to Pebble — sometimes as immediately as the second they get reported, and then it's just a quick PR review.
QA cycles that don't stall over the weekend
AllFly recently started using Pebble in their QA process as well. When something gets kicked back from QA, it goes straight to Pebble for triage and a proposed fix. If QA finds a side effect or a related case, they keep adding to the same card and Pebble responds.
One recent example stuck with Nate: a feature went into QA on a Friday. By Monday morning, Pebble had already triaged a follow-up issue, fixed it, and opened a PR. Nate reviewed it, approved, and merged — before his first coffee.
I came back Monday, and Pebble had already triaged it, fixed it, and opened a PR. I just looked it over, and merged it.
Nate estimates the team has roughly doubled the volume of issues they're able to clear, both in ticket count and in story points. The dedicated backlog sprints are gone. The QA cycle is tighter. And perhaps most notably, the QA team has started tagging Pebble directly when they log issues in Slack — building it into their process without being asked.
The remaining 20% of issues — the complex, high-context work that requires deep knowledge of AllFly's systems — still gets developer attention. But it's now completely manageable, because everything else is already handled.
Everyone's happier, our customers see stuff faster, and our internal teams for sales and ops see stuff going out a lot quicker. It used to be a lot more back and forth about when we'd have time for it.
