A Product Owner in one of our recent cohorts asked a very reasonable question: how do I add custom fields to Jira to get better AI output? It’s the kind of question that assumes the fix is additive – one more field, one more piece of structure, and AI will finally have what it needs. Often, the real answer is the opposite: stop adding, and go look at what’s already there.

The 58-Status Problem

When we dug into one organization’s actual Jira instance, we found 58 different status values in active use – including three separate near-duplicate spellings of “Cancelled” alone (“Cancel,” “Canceled,” and “Cancelled,” each used inconsistently by different people over time). Nobody set out to create that mess. It accumulated – one team’s shorthand here, one person’s typo there, never cleaned up because it never broke anything badly enough to force the issue.

AI doesn’t fix this. It inherits it. Worse, it can make the mess harder to notice, because AI-generated summaries and analysis tend to sound confident and coherent even when the underlying categorization is a mess – smoothing over inconsistency with polished language rather than flagging it.

The Good News: AI Can Help You Find the Mess

Here’s the flip side worth knowing: you don’t need to audit your data manually before AI can help. AI is genuinely good at exactly this kind of pattern-spotting task – ask it to list every distinct value currently in use for a given field, and it will often surface near-duplicates and inconsistencies faster than a human scanning the same list. The sequence that actually works is: use AI to find the mess, clean it up with your team, and only then build the automation or custom field you originally wanted.

Why This Matters More as You Automate

A messy status field is mostly harmless if a human is still reading every ticket individually. It becomes a real problem the moment you ask AI to act on that data at scale – generating a report, triggering a workflow, or flagging items for attention. An automation built on top of “Cancel,” “Canceled,” and “Cancelled” as three separate values will quietly miscount, misroute, or miss things entirely, and the failure often won’t be obvious until something important slips through.

A Simple Sequence to Follow

Before adding a new field or building a new AI-driven workflow on existing data: ask AI to list the distinct values currently in use for the field you’re about to rely on. Review that list with your team – not alone, since inconsistencies often trace back to different people’s different habits, and someone else may recognize a pattern you wouldn’t catch solo. Consolidate what’s genuinely redundant. Only then build whatever automation or reporting you originally had in mind on top of the now-cleaner data.

Have a question, a use case you’re proud of, or a task you tried this on? We’d love to hear about it at the next Alumni Meetup — it’s one of the best places to trade real examples with other Product Owners working through the same challenges, and every session leaves you with at least one idea you didn’t walk in with.

Keep going:

• Ready to go deeper? Check out our upcoming AI courses for hands-on classes with real practice, not just reading.
• Want more like this? Subscribe to our YouTube channel for walkthroughs and demos that pair well with these posts.
• Have questions or want to talk it through? Reach us directly at hello@rocketninesolutions.com.
• Know a colleague wrestling with the same challenge? Send them this post — half the value of this series comes from the people who’ve already tried it sharing what worked.

A Note on How This Was Written: This post was developed collaboratively between Rocket Nine’s training team and Claude, Anthropic’s AI assistant – drawing on real material from our AI for Product Owners class and anonymized insights from the Product Owners who’ve taken it. We think that’s fitting: the same thoughtful human-AI collaboration we teach in class is exactly how we approached writing about it.

© 2026 Rocket Nine Solutions Inc. All rights reserved.