Free tool
AI-Ready Task Test
Score any task in your business out of 5 and get a straight answer: automate it, pair it with a human, or leave it alone. The same five checks we run in discovery before recommending anything, and your answers never leave this device.
Score a task
You enter: a task in plain words and five yes or no answers. You get: a verdict per task and a ranked shortlist with your first automation named.
Or start from one we hear every week:
Tasks you score are kept in this browser only, so your list is here when you come back.
Your automation shortlist
Score your first task
Highest score goes first. Ties go to the task that happens after hours.
How the verdicts work
Each yes is one point. No weighting, no hidden levers.
4 or 5 = automate it now
3 = pair it: AI runs the front, a human owns the finish
0 to 2 = keep a human on it
The five checks come from our task-selection method, written up in What Can AI Actually Do for Your Business? A task that scores 3 is not a failure. It usually means the escalation rule is the thing to design, not the automation.
Worked example
One filled-out list, scored
The strongest lists mix both kinds of task. Here is what an honest one looks like.
The ideal way to fill it in
What comes back
The 5/5 lands at #1 on your shortlist with the next step spelled out: write the rules down, test in the shadows, go live with escalation rules.
Press Load a worked example in the tool above and this exact list appears, ready to edit.
Why this works
Instinct picks the annoying task. The test picks the profitable one.
Most owners automate whatever irritated them most recently. The checklist forces the boring questions that actually predict success: repetition, written rules, words and data, coverage gaps, and the cost of a small miss. Notice that none of them mention technology. They are about the shape of the work.
Run a normal week through the list and a pattern shows up fast. Answering, booking, reminders, review requests, and follow-up pass. Pricing a complicated job, calming down an unhappy customer, and the craft you actually sell do not, and that is fine. Nobody is asking AI to hold the clippers. Once you have a winner, the next step is writing down what you already know so a system can use it, and the AI Training Pack builder does exactly that.
Questions we get asked
Is a 4 or 5 always safe to automate?
Safe to start, not safe to abandon. A high score means the task has the right shape: it repeats, the rules can be written down, and a small miss is cheap. You still test in the shadows first and go live with escalation rules, then review the transcripts weekly for the first month.
What do I do with a task that scores 3?
Split it. Almost every 3 is two tasks wearing one name: a repetitive front half and a judgment back half. Lead qualification is the classic case. AI screens and books, you close. The design work is the handoff, and getting that boundary right is most of what we do for clients.
Should I automate several tasks at once?
No. One task done well beats five done halfway, because each one teaches you how your customers actually talk. Pick the top of your list, run it for a month, then come back for the next. The list will still be here.
Where does this test come from?
It is the scoring method we use in our own discovery calls, published in full in our post What Can AI Actually Do for Your Business? We put it here as a working tool because reading about a checklist and running your week through one are different experiences.
Where do my answers go?
Nowhere. The scoring runs in your browser and your task list is kept in this browser’s local storage so it survives a refresh. Nothing you type is sent to us or to anyone else.
Who built this
The first question we ask on every call
Before we recommend anything, we run a client’s week through these five checks. Omar Ramos, our founder, is a systems engineer and program manager with 16+ years of building operational systems, and this is the filter that decides what we automate and what we refuse to.
When your first task involves the phone, the usual builds are missed-call text back and the AI receptionist. If you want the leak sized in dollars before you pick a task, run the Lead Leak Self-Audit first, and when you are ready to train something, the AI Training Pack builder turns your answers into the actual document.
Built and maintained by Project Driver. Last reviewed August 2026.