Outsource bug triage to a managed team

Outsourced bug triage means a trained team works your incoming bug queue, so engineers open the tracker and find issues that are reproduced, labeled, deduplicated and routed. Deepen AI does this in your tracker, on your builds, to your severity scheme.

What the work is, and who buys it

Buyers are product and engineering teams whose bug intake has outgrown the time anyone has for it. Reports arrive from support tickets, beta testers, app store reviews, internal testing and crash tools. The symptoms are familiar: duplicates, "cannot reproduce" back-and-forth, every report marked urgent, an untriaged column nobody owns, and engineers reading reports during sprint planning.

This is not test automation, and we do not write or fix code. It is the judgment work in the middle. If you only need written test cases run before a release, that is QA and test execution, a standard task listed on the services page. Triage needs product knowledge and reproduction skill, which is why it sits in the specialist tier.

What our team does

• Reproduce incoming bugs on the right build and environment, and record exact steps, expected result and actual result
• Attach logs, screenshots and screen recordings to the issue
• Label severity, area and component to your scheme
• Find and merge duplicates, linking the original reports
• Route issues to the owning team, using your ownership map
• Chase missing information from reporters through your help desk or the tracker
• Re-test fixes on staging when an engineer marks an issue resolved
• Summarize recurring issues for your weekly engineering review

The team works in Jira, Linear, GitHub Issues or your own tracker, with your help desk, crash reporting tool and test builds, on accounts you create.

How it runs

1. You share the tracker, scheme and examples. Send your severity definitions, component and ownership map, 30 to 50 issues your team triaged well, and access to builds and test accounts.
2. We agree scope and train. Our ops team reviews the work, confirms the platforms and builds the team can test on, and sends the rate in your pilot plan. Triagers train on your examples, and we build a gold set of past issues with known-correct severity, component and duplicate links for you to confirm.
3. We deliver, with a named team lead and QA sampling. Triagers work the live queue. The team lead settles disputed severity calls, keeps a decision list, and posts open questions for your contact.
4. You get a weekly report. Issues triaged, reproduced and not reproduced, duplicates merged, severity mix, QA sample results and recurring themes.

Illustrative example: a support ticket to a triaged issue

Illustrative only. The app, ticket and issue numbers are invented.

• Input, support ticket #8812: "App crashes every time I upload a photo from my phone."
• Output, title: iOS: app crashes on photo upload when the file is HEIC and over 10 MB
• Output, reproduction: build 4.18.2 from your beta channel, on an iPhone 13 test device. Steps: open Profile, tap Change photo, choose a 12 MB HEIC file. Actual: app closes. Expected: upload with progress bar.
• Output, not reproduced with: a JPEG under 10 MB, or the same file on Android build 4.18.1
• Output, labels: severity S2 under your scheme (core flow broken, workaround exists) · component Media/Upload
• Output, duplicate: matches MOB-3307 from a beta tester. Merged, and ticket #8812 linked.
• Output, routed to: Mobile Platform team, per your ownership map
• Who did what: one triager reproduced, labeled and routed the issue. A QA reviewer re-checked the severity and duplicate link as part of the weekly sample and agreed.

How we keep triage reliable

Good triage means the right severity, real duplicates and steps an engineer can follow without asking. The gold set of your past issues measures each triager on all three before live work. QA reviewers re-check a sample of triaged issues against your scheme, including the reproduction steps. They do not re-check every issue. When an engineer sends an issue back, the team lead replies with a fix or a reason, and repeated misses become new lines in the decision list. The weekly report shows how many issues were sampled and the acceptance rate on that sample. See how we check quality.

Pricing scope

Bug and issue triage is quoted per project, above the standard rate. The rate depends on your platforms, devices, tracker setup and how much product knowledge the work needs. Describe the work in the pilot form, and our ops team sends the rate with your pilot plan. Like all our work, it starts with a two-week prepaid pilot at the quoted rate, with at least 20 hours per person per week. The team lead and QA sampling are included. After the pilot, we invoice weekly in arrears. See the pricing page.

When to use the enterprise team instead

If your issues come from autonomous vehicles or robots, and triage means reviewing sensor recordings to work out what the system perceived and why it acted, that work belongs with the Deepen AI enterprise team. It is scoped and priced per project. See enterprise AI data services.

Send your severity scheme and a sample of recent reports. We email you a pilot plan with a quote.

Request a pilot plan