The goal
By Friday, you should have one clear answer to the question every business owner asks eventually: "Where should our next tech dollar go?"
Not a wish list. Not a Pinterest board of AI tools. One answer, backed by a quick score you can defend to yourself or your team.
The method: score your core processes against four types of leverage — Efficiency, Scale, Insight, and Risk Reduction — and let the highest score tell you where to start. This is a scoring exercise, not a purchase decision. You're finding the target before you pick the weapon.
Why scoring beats gut feel
Most tech spending in small businesses happens one of two ways: something breaks and you scramble, or a vendor pitch catches you at a weak moment. Both are reactive.
A leverage audit flips that. You look at what you already do — sales, delivery, billing, support — and ask which lever is weakest: are you slow (efficiency), fragile under growth (scale), blind (insight), or exposed (risk)? Whichever process scores worst across the most lenses is your starting point. That's the process technology will pay off on fastest.
Score your processes in 90 minutes
Put a single 90-minute block on your calendar this week. Not "sometime this month" — a specific Tuesday, 2:00 to 3:30. If you have a small team, invite the two people closest to daily operations. Solo operator? Do it alone, with a notebook or blank doc open. Resist the urge to turn it into a strategy offsite.
List your core processes. Write down five to eight things that actually run your business: sales quoting, onboarding, order fulfillment, invoicing, customer support, scheduling, payroll. Don't overthink the list.
Example, for a small HVAC company: lead intake, scheduling, dispatch, invoicing, follow-up/reviews, inventory reorder.
Score each process on four lenses, 1–5. For each process, ask four questions and jot a quick gut score. Don't research — score from memory, in under a minute per process.
- Efficiency: How much manual, repetitive work does this involve? (1 = drowning in busywork, 5 = lean)
- Scale: If volume doubled tomorrow, would this process hold up? (1 = chaos, 5 = fine)
- Insight: Do you have real numbers on this process, or are you guessing? (1 = pure gut, 5 = tracked and reviewed)
- Risk: Does this process depend on one person's memory, or lack basic safeguards? (1 = one bad day away from disaster, 5 = solid)
Lower scores mean bigger opportunity. A process scoring 2 on Efficiency and 2 on Risk is waving a flag at you. Keep it to a spreadsheet — six rows, four columns, a total. Nothing fancier is needed.
Circle your worst offender. Add up the scores per process. The lowest total is your candidate. If two processes tie, break it by asking: which one costs us the most in hours, errors, or missed revenue right now? That's your winner.
Using the HVAC example: dispatch might score low on Efficiency (a lot of phone tag) and low on Risk (only one dispatcher knows the routing logic). That combination — high pain, single point of failure — usually outranks something like invoicing that's merely annoying but not dangerous.
Write the winning process at the top of a fresh page. This is what you're solving this quarter — not five things, one thing.
Pilot before you buy anything
Before buying software or building a workflow, test your assumption on a tiny scale. This is the step people skip, and it's the one that saves you money.
Pick the smallest possible slice of the problem and fix it manually or with a free tool for one week. If dispatch is your target, don't buy a $200/month dispatch platform on day one. Instead:
- Write down the current dispatch steps exactly as they happen today (even three bullet points count as documentation).
- Try a stripped-down fix for a week — a shared calendar instead of phone calls, or a simple group text protocol.
- Track one number: minutes spent coordinating, or number of missed/late jobs.
If the manual fix meaningfully helps, you've confirmed the process itself needs a rebuild before any tool touches it — automating a broken process just breaks things faster. If the manual fix barely moves the number, the bottleneck may be volume or complexity that only a real tool can solve — now you know to spend money there with confidence.
Either outcome is useful. That's the point of a pilot.
Decide what happens next
Once your one-week pilot is done, you're at a fork. This is exactly the decision covered in Run an Automate/Improve/Ignore Triage This Week — use that framework now to sort your winning process into Automate, Improve, or Ignore.
If it's headed for Automate, don't jump straight to building something custom. Work through Build, Buy, or Defer: Choosing the Right Path for Your Next Tool before you sign up for anything. For most small businesses, buying an existing tool beats building — save custom development for the rare case where the process is genuinely core to your edge.
A few pitfalls to watch for along the way:
- Scoring too many processes at once. Five to eight is the ceiling for a first pass. More than that and the exercise turns into procrastination dressed as diligence.
- Skipping the pilot because you're "pretty sure" you know the answer. Gut feel is exactly what skipped scoring in the first place. A week of real data beats a month of confident assumptions.
- Chasing the lens that sounds most exciting instead of the one that scored worst. Insight tools (dashboards, AI analytics) get a lot of hype right now. If your actual worst score is Risk — no backups, one person holding all the process knowledge — fix that first. It's cheaper, and the downside of ignoring it is much worse.
- Treating the audit as one-and-done. Scores change as you fix things and as your business grows. Rerun this exercise every quarter, ideally right after you close out a fix — which is also when you should lock in the win so it doesn't quietly slide back into the old way of doing things.
What you'll have by Friday
A ranked list of your core processes, a clear worst offender, a week of real pilot data, and a decision — automate, improve, or ignore — that you can explain in one sentence. That's not a small thing. Most owners never get past "we should probably fix that someday." You'll have already started.