Stop letting the loudest voice set the priority.

Welcome to Issue #06.
Every Saturday I ship one field-tested lesson from live BA × AI delivery — the stuff I actually use on client work this week, not repackaged theory.

Two weeks ago a prioritisation call went the way they usually go. The sales director wanted a reporting export, "because a client asked." The head of ops wanted the duplicate-invoice fix "because it's embarrassing." Both said the word critical. Neither brought a number. The loudest, most senior voice won — and three days later the export shipped to a client who opened it once, while the invoice bug kept quietly double-charging people. A backlog ranked by volume and rank is not a plan; it's a record of who was in the room. Here's how I use AI to make the ranking come from evidence instead.

Make priority something you can point at

Three moves that turn "I feel this is important" into a number the whole room can see, argue with, and agree on — usually before the meeting is even over.

1. Force every item into the same shape before you compare them

You can't rank apples against sentences. So I paste the raw backlog in with a structured prompt: "For each item, restate it as who it affects, how many of them, how often it happens, and what breaks if we don't do it. If any of those four is unknown, mark it UNKNOWN — do not estimate." Now every item reads the same way, and the gaps are visible. Half the "critical" items turn out to affect one person once — you can see it the moment they're all in the same format.

2. Score impact and effort separately — never in one breath

The classic mistake is arguing value and cost at the same time, so the cheap-but-pointless thing beats the expensive-but-vital one. I split them: "Rate each item's impact 1–5 using only the reach and consequence I gave you, ignoring effort entirely. Then, in a separate pass, rate effort 1–5 from the technical notes, ignoring impact." Two independent numbers, then impact-over-effort gives me a ranked list. The duplicate-invoice fix — high impact, low effort — floats to the top on its own, no seniority required.

3. Make it name what each choice costs you

A ranking is only convincing if it's honest about the trade. So I ask: "For the top five, write one sentence naming what we're choosing NOT to do by picking this, and who will feel that." Now the export isn't just 'lower' — it's "we're accepting that one client waits a week so we stop double-charging everyone else." People rarely fight a priority once they can see exactly what the alternative costs. That sentence ends more debates than any score does.

What NOT to do

• Don't let AI assign the impact number from thin air. It will happily rate everything a confident 4. Feed it the reach and consequence you gathered; its job is to apply the scale consistently, not to invent the inputs.
• Don't treat the ranked list as the decision. It's the starting position that makes the conversation honest. A stakeholder can still override it — but now they have to say out loud why, against a number, in front of everyone.
• Don't hide the effort scores from the business side. The whole point is a shared view. The moment priority lives only in the BA's spreadsheet, you're back to being the loudest voice — just a quieter one.

Try this week

Take the backlog you're about to walk into a prioritisation meeting with. Run it through the same-shape rewrite, then score impact and effort in two separate passes. Sort by impact over effort and write the one-sentence trade-off for your top three. Walk in with that on the screen instead of in your head — and watch how fast the meeting stops being about who's most senior and starts being about what the numbers say.

See you next Saturday.
— Dibya Mishra, Senior Business Analyst