Skip to content
Sprinter Consulting
Start here

AI Workflow Opportunity Map · free · self-scored

Find a useful place to start with AI.

Choose a workflow and answer 15 questions about how your team works. Get a readiness snapshot and a three-step trial plan you can download and use. Start with research, reporting, customer follow-up, or another recurring task.

Start with your team’s everyday work and look for a useful place to improve. Your answers stay in your browser unless you request a review. We count completions and downloads without your answers.

Build my opportunity map

§ 01Why this exists

A useful starting point is closer than you think.

Start with the people doing the work. They know which tasks repeat, where information gets lost, and what a better result would look like. That knowledge is the foundation for useful AI.

This scorecard helps you find an opportunity, check what is ready, and choose one place to begin. A small improvement can give the team confidence, time, and ideas for what to try next.

§ 02Part 1 · The 12 leaks

Where your team could gain time and capacity.

Read this list with your own business in mind and count every leak you recognize. A few specific examples will make your answers more useful. You’ll need your count for the first question below.

  1. 01 The quote/estimate bottleneck

    Quotes take days, depend on whoever is available, and vary by who builds them. Pricing logic lives in one person's head or a fragile spreadsheet.

    COST SHAPE: LOST BIDS, INCONSISTENT MARGIN, KEY-PERSON RISK

    What AI realistically does: drafts the estimate from plans, photos, specs, or history, so a human reviews instead of building from scratch.

  2. 02 Orders and inquiries sitting in email

    New orders, RFQs, and customer questions land in an inbox and wait until someone notices. Response time depends on workload and vacation schedules.

    COST SHAPE: SLOW RESPONSE, LOST ORDERS, CUSTOMERS WHO QUIETLY LEAVE

    What AI realistically does: triage: classify, extract the key fields, route to the right person, draft the first response, flag anything aging.

  3. 03 Re-keying the same data

    The same order, job, or invoice gets typed into the ERP, then a spreadsheet, then the CRM, then the accounting system.

    COST SHAPE: TIME SPENT ON REPEAT ENTRY AND THE ERRORS IT CAN INTRODUCE

    What AI realistically does: reads documents and messages, extracts structured data, pushes it into the systems; humans verify exceptions instead of typing everything.

  4. 04 Tribal SOPs

    How the work actually gets done lives in a veteran's head, old screenshots, and a binder nobody updates. Training a new hire takes months. Losing the veteran is a crisis.

    COST SHAPE: SLOW ONBOARDING, SINGLE-PERSON DEPENDENCY, A DISCOUNT ON YOUR VALUATION IF A BUYER EVER LOOKS

    What AI realistically does: turns recordings, walkthroughs, and message threads into living SOPs, and keeps them current as the work changes.

  5. 05 Follow-up decay

    Leads, quotes, and customer touchpoints go stale because follow-up depends on memory. Useful context is scattered across the CRM and inbox.

    COST SHAPE: USUALLY THE CHEAPEST REVENUE YOU'RE LOSING: PEOPLE WHO ALREADY RAISED THEIR HAND

    What AI realistically does: watches the pipeline, drafts timely follow-ups in your voice, and flags quotes that need attention.

  6. 06 The catalog/pricing swamp

    Hundreds or thousands of SKUs, options, or configurations. Descriptions missing or inconsistent. Price updates are an event. Versions are hard to track.

    COST SHAPE: QUOTING ERRORS, CHANNEL INCONSISTENCY, AN ARMY OF SMALL MANUAL UPDATES

    What AI realistically does: bulk-generates and maintains descriptions, checks consistency across channels, flags pricing anomalies.

  7. 07 Reporting that arrives too late to matter

    The numbers come together days or weeks after the period closes, assembled by hand from three systems. By the time you see the problem, it's history.

    COST SHAPE: DECISIONS MADE BLIND, PROBLEMS CAUGHT LATE, AN OWNER WHO CAN'T STEP AWAY

    What AI realistically does: pulls from the source systems on schedule and delivers the operating picture daily or weekly, with anomalies called out in plain language.

  8. 08 The owner as router

    Every exception, judgment call, and “where does this go?” lands on the owner or COO. The business's real operating system is one person's attention.

    COST SHAPE: THE MOST EXPENSIVE PERSON DOING DISPATCH WORK; GROWTH CAPPED BY ONE CALENDAR

    What AI realistically does: handles the routing rules that are actually rules, drafts responses for the judgment calls, and shrinks the exception queue to the cases that genuinely need you.

  9. 09 Invoicing lag

    Work finishes; the invoice goes out weeks later because assembling it is a chore.

    COST SHAPE: CASH SITTING IN OTHER PEOPLE'S BANK ACCOUNTS, PLUS WRITE-OFFS FROM DISPUTES THAT STALENESS INVITES

    What AI realistically does: turns completion records, time logs, and job notes into a draft invoice the day the work is done.

  10. 10 Document processing drudgery

    POs, invoices, packing slips, permits, inspection forms: printed, scanned, re-typed, filed, occasionally lost.

    COST SHAPE: HOURS PER WEEK PER PERSON, PLUS THE COST OF THE ONE THAT GOT LOST

    What AI realistically does: reads them, extracts what matters, files them findably, and updates the systems.

  11. 11 Compliance and paperwork scramble

    Permits, safety docs, certifications, audit trails: scattered across trucks, trailers, email, and folders. Producing them on demand is a fire drill.

    COST SHAPE: RISK EXPOSURE, FAILED AUDITS, DEALS DELAYED IN DILIGENCE

    What AI realistically does: keeps a current, queryable record and assembles the package when someone asks.

  12. 12 Knowledge that leaves in the parking lot

    Customer history, vendor quirks, machine settings, “we tried that in 2019 and here's why it failed”: undocumented, unsearchable, and walking out the door with every departure.

    COST SHAPE: REPEATED MISTAKES, LOST RELATIONSHIPS, EXPERIENCE THAT IS HARDER FOR THE NEXT PERSON TO ACCESS

    What AI realistically does: captures it as a by-product of normal work (calls, messages, documents) and makes it answerable: ask the question, get the institutional answer.

§ 03Part 2 · The scorecard

Fifteen questions, five dimensions. Score yourself 0, 1, or 2.

Use the answers that best describe work today. The dimensions matter more than the total. Your result shows what looks ready, what needs attention, and a practical next step.

Choose a starting point. Your map will include a small experiment, a review step, and a way to measure progress.

dimension A · 01–03

Workflow leakage · How much is draining out?

A1 From the leak list above, how many did you recognize in your business?
A2 A new order or lead arrives at 4pm on a Friday. What happens to it?
A3 How much of your team's week is typing information from one place into another?

A higher score here means more recurring friction to investigate. The trial helps you check whether AI can improve it.

dimension B · 04–06

Data & systems readiness · Can a system reach the work?

B1 Where does the information needed to do the work live?
B2 If you needed last month's orders, jobs, or invoices as a list, how long would that take?
B3 Could you give a trusted outside party read-only access to one workflow's systems this month?

dimension C · 07–09

Adoption readiness · Will the change stick?

C1 Is there one person, not the owner, who owns how a given workflow gets done?
C2 When you've changed a process before, what happened?
C3 How does your team feel about AI touching their work?

dimension D · 10–12

Value density · Is the prize big enough?

D1 Pick your worst leak. What does it plausibly cost per month in hours and dollars?
D2 Does the leak touch revenue (quotes, leads, follow-up, orders) or just cost?
D3 Is there a deadline that makes this urgent: a departure, a sale process, an ERP migration, a busy season?

dimension E · 13–15

Risk fit · How much autonomy can the work tolerate?

E1 For your worst leak, what would “AI helping” need to look like to be safe?
E2 If an AI-drafted quote, email, or report had an error, would you catch it before it caused harm?
E3 Is any of the work in your worst leak governed by regulation or contractual compliance?

Your score

0 OF 15 ANSWERED

Answer all 15 questions to get your readiness snapshot and a three-step trial plan. The map stays free to view and download.

§ 04Part 3 · The one-week inspection

Verify your score against reality before you spend a dollar.

Five inspections, about 30 minutes each, no consultant and no software. Take the worst finding and put a monthly dollar figure on it, even rough. That number is what any AI proposal, ours or anyone’s, should be measured against.

  1. Monday

    Time-stamp the intake

    Pick the inbox where orders, leads, or RFQs arrive. For one week, log arrival time vs. first-human-action time on each. The gap is your triage leak, in minutes you can multiply.

  2. Tuesday

    Shadow the re-keying

    Sit with the person who enters data for 30 minutes. Count systems touched, fields typed twice, and workarounds. Ask what they would improve first; their experience is a useful place to start.

  3. Wednesday

    Pull the quote log

    Last 20 quotes or estimates: days from request to delivery, and won or lost. If slow quotes correlate with losses, you've found a revenue leak, not a cost leak. Compare that opportunity with your other priorities.

  4. Thursday

    Run the SOP test

    Pick one critical process. Ask two people who do it to each write down the steps, separately. Compare. The differences are your tribal-knowledge exposure, in writing.

  5. Friday

    Time the truth

    Ask for one operating number you care about (margin by job, output by line, aged receivables). Time how long it takes and how many hands touch it. That's your reporting lag, measured.

§ 05Part 4 · Choosing a useful partner

Five things to look for in an AI partner.

  1. 01

    A diagnostic before a retainer

    Ask how the partner will understand your work, choose a starting point, and measure whether the engagement is useful.

  2. 02

    Workflow redesign, not tool installation

    Ask which workflow improves, who will use it, and how their experience will shape the result.

  3. 03

    A focused first build

    Agree a realistic timeline for a useful pilot, with a named owner and clear success criteria. Systems access and review needs should shape the plan.

  4. 04

    A value ledger

    Hours saved, response times, error rates, margin recovered: written down before the build and reviewed after. Use the results to decide what to improve or try next.

  5. 05

    The right to stop

    Each phase should stand alone. You keep the map, the SOPs, and the systems whether or not you continue.

Written by the person who’d run your audit

Sprinter is run by a builder, not an agency. Tyler’s first business was roofing and construction, where missed follow-ups, inconsistent estimates, and knowledge trapped in one estimator’s head aren’t slides; they’re payroll. He trained as a mechanical engineer (Auburn) and worked on machinery at ExxonMobil, then spent years building software for non-tech operations: cabinet pricing and configuration, wood-products manufacturing workflows, family-office admin. The scorecard you just took is the same diagnostic logic we use in paid work, self-served.

If your score came back “ready to map”

Turn your findings into a practical plan.

We map one expensive workflow with the people who run it, baseline it, test readiness and risk, and decide whether to build, buy, redesign, or stop, with a build-ready 90-day plan. Paid, fixed scope, and the deliverable is yours whether or not we ever build anything together.

See the AI Opportunity Auditsee how we work

You’re welcome to share this page inside your company. If you score it and want a second opinion, send the scores to hi@sprinterconsulting.com. We’ll tell you what we’d inspect first, no charge.