Score · AI Potential Scoring
Every task, process and value stream, plotted by AI value potential and feasibility. One view of where AI pays off, from a single activity to the entire landscape.
AI Value × Feasibility
Most organisations collect AI ideas in a spreadsheet: forty rows, each one somebody’s favourite, none of them comparable. The loud ones get funded. The ones that would move the business stay on row 31.
The matrix replaces the list with a position. Every piece of work gets two scores on the same scale, so the comparison is between steps, not between opinions.
Both scores are derived from the process model: what you already know about each step, not a workshop vote.
Horizontal
How much is at stake if AI takes over or assists this step.
Vertical
How realistic it is to put AI on this step now.
High value · High feasibility
The money is real and the path is clear. These become opportunity cards and go straight into a process branch.
High value · Low feasibility
Worth it, but something is missing: data, a system connection, an explicit rule. The matrix tells you what to fix first.
Low value · High feasibility
Easy, but it won’t show up in the numbers. Fine for a team that wants to learn, wrong as a flagship pilot.
Low value · Low feasibility
Not now. Having this quadrant visible is what stops the loud ideas from getting funded.
The same two scores roll up. Switch the level and the matrix re-plots.
Tasks
Every activity in a process, plotted. This is where opportunities become specific: a step, an owner, a system.
“Check order” — high value, high feasibility, 620 cases a month
Processes
Each process scored by its aggregated steps. Compare quote creation against claims against invoicing on one chart.
Quote creation leads on value, Order to cash on feasibility
Landscape
Value streams and domains on the same axes. The view a leadership team needs to set an AI roadmap, not a wish list.
Lead to cash carries 60% of the total AI value potential
Each position on the matrix opens the scored opportunity behind it: what the agent would do, where, for how much, with what confidence. From there it’s one click to a process branch and the Ship stage.
How the AI Opportunity Finder worksQuote creation · Review TRA · High value, high feasibility
Quotes wait 9.4 d in Proposal sent with no owner activity in 71% of cases. An agent drafts the follow-up from the proposal and the last thread and files it for the owner.
With Agentic Process Mining
Connect your systems and the matrix scores from execution data instead of attributes. Volumes, waiting times and rework come from real cases, and dots shift as the business changes.
About the add-on“Check order” moved from medium to high value after volume rose 38% in Q3.
Re-scored weekly · 2,140 cases · ERP + CRM
Model one process, score it, and see which steps land top right.