Method · · 5 min read
How to choose your first AI automation: the Impact Score method
Don’t start with the most exciting AI idea. Start with the task that scores highest on (F × T ÷ 60) × P × R ÷ Complexity.
By the Upper Node team
Most AI projects fail because they start with the technology instead of the problem. The Impact Score is the method we use in discovery calls to find the one task worth automating first.
Step 1: list the tasks that repeat
Write down everything your team does again and again: answering “is my report ready?” on WhatsApp, entering bookings into a system, chasing invoices, calling back ad leads, reminding customers about appointments.
Step 2: score each task
- F — frequency: how many times a month it happens.
- T — time: minutes each time.
- P — pain, 1 to 5: from mildly annoying (1) to legal, safety or reputational risk (5). A 4 means a customer or money is lost.
- R — reach, 1 to 3: one person, a team, or the whole company.
Impact = (F × T ÷ 60) × P × R
The first part converts the task into hours per month. Pain and reach then weight those hours by how much they actually matter.
Step 3: divide by complexity
Rate how hard the fix would be, from 1 (a day’s work) to 5 (a month). Dividing by complexity stops you from picking a high-impact task that would take a quarter to build.
Priority = Impact ÷ Complexity
A worked example from a diagnostic lab
Answering “is my report ready?” on WhatsApp happens 900 times a month and takes 2 minutes each time. That is 30 hours. With pain 3 and reach 2, impact is 180. At complexity 2, priority is 90, which is usually the highest on the list and an obvious first automation.
Step 4: put a rupee figure on it
Multiply total hours by a loaded hourly cost to get the cost of time. Then estimate leaked revenue: missed leads × close rate × average deal value. Those two numbers make the business case clear to everyone in the room.