Frameworks
The Meridian Automation Score
Before you automate anything, it helps to know whether it's worth automating at all. The Automation Score is the rubric we use to read one process, seven factors that, together, tell you whether it's a strong AI candidate or a poor fit.
What it is
One process, seven factors
The Automation Score reads a single task the way we would in an assessment. You rate it on seven factors, each pointing the same direction: the more of it there is, the stronger the case for automating. Add them up and you get a quick, honest read, a strong candidate, a poor fit, or something in between that's worth a closer look.
It won't tell you exactly what to build. It tells you whether to bother, which is the question that saves the most wasted money. A low score is as useful as a high one.
The factors
What we rate, and why
Each factor is something we've found actually predicts whether an automation pays off. Rate each one from low to high for the process you're weighing.
Volume and frequency
How often the work runs and at what quantity. High-volume, regular work has the most hours to give back, so it scores highest.
Repeatability
How consistent the steps are. Work that follows the same path every time is a strong fit; work that branches on judgment at every turn is not.
Data readiness
Whether the inputs are available and structured enough to work with. Clean, reachable data scores well; scattered or handwritten inputs pull the score down.
Time and cost drain
How much time and money the work ties up today. The bigger the current drain, the more an automation is worth.
Error tolerance
How costly a mistake is, and how easily it gets caught. More tolerant, easily-checked work is safer to automate and scores higher; high-stakes work needs more care and more oversight built in.
Few systems to connect
How many tools the work touches and how hard they are to link. Fewer, well-connected systems are simpler to build for.
Low judgment needed
How much a person's judgment the work genuinely needs. Tasks that lean on rules score well; tasks that lean on experience, relationships, or a real decision belong with your people.
A worked example
How a strong candidate reads
Take a generic example: a monthly invoice-matching task, the kind of high-volume, rule-based work a lot of finance teams run by hand. Scored on the seven factors, it reads like this.
Try it
Score your own process
Rate a process you're weighing on each factor, 1 (low) to 5 (high). The score updates as you go. Nothing you enter leaves your browser.
21 / 35
Worth a closer look
Some strong signals, some caution. Worth a proper look in an assessment.
This is a guide, not a guarantee, and a real build still gets scoped and tested. For the strong candidates, the AI Opportunity Matrix is the next step.
Where it goes next
From a score to a build order
A score reads one process. Most businesses have several worth weighing, so the next step is ranking them against each other. That's the AI Opportunity Matrix: it takes the strong candidates and sorts them by value and effort so you build in the order that pays back soonest.
Want a rough number on the payoff first? The ROI estimator turns your own inputs into an estimate of the hours and cost a strong candidate could give back.
Score, then rank. That's the Assess and Prioritize half of how we work. The frameworks overview shows both together, and the Free AI Opportunity Assessment is where we score your real processes with you.
Ready when you are
Get your processes scored
The Free AI Opportunity Assessment is a no-cost look at how your business runs. We score your processes on these factors and tell you, honestly, which are worth automating and which aren't.
Common questions
- What is the Meridian Automation Score?
- It is a plain rubric for reading whether a single task or process is a strong candidate for AI or automation. You rate it on seven factors: volume, repeatability, data readiness, cost drain, error tolerance, systems involved, and how much human judgment it needs. A high score means a strong candidate; a low score is an honest sign to leave it alone for now.
- What makes a process a good automation candidate?
- The strongest candidates are high-volume and repetitive, run on data that is already available and structured, cost real time or money today, and need little human judgment on each pass. The weakest are low-volume, one-off, judgment-heavy, or run on data nobody can get to cleanly. Most real processes sit somewhere in between, which is exactly what the score is for.
- Is a high score a guarantee it will work?
- No. The score is a guide, not a promise. It points you at the work most likely to pay off so you don't sink months into a poor fit, but the real build still gets scoped, priced, and tested. If a process scores low, that is useful too: it saves you from spending on something that wouldn't have returned.