Skip to content
StrikeArc LabsBook an intro call
Menu
What We Do
What We DoHow It WorksPricingCompareStrikeArc ScoutStrikeArc PowerPlays LibraryStrikeArcOS
PowerPlays
All twelve PowerPlaysP01 Put your ad budget behind what worksP03 Win more deals with better designP07 Respond to new leads fasterP08 Get quotes and contracts out fasterP09 Track the competitor moves that matterP11 Turn your expertise into a productP04 Stop overpaying for AIP05 Find the money you're leakingP10 Spot churn, margin, and cash risks soonerP02 Prepare business reviews in less timeP06 Cut the manual work from orders and invoicesP12 Find the answers your team needs
Who We Serve

Any mid-market business can put AI to work with StrikeArc Labs. These are the industries where our expertise runs deepest.

Distribution and wholesaleManufacturingLogistics and transportationHVAC, plumbing and field servicesB2B SaaSEquipment dealers and rentalCommercial servicesAll industries
Field Guides
AI Readiness AssessmentSales AutomationBack-Office AutomationManaged AI ServicesChoosing an AI PartnerTools, templates and papersBlogThe Field Guides home
AboutBook an intro call

ChecklistChoose who does the work

The AI Partner Scorecard: twenty questions to ask before you sign.

Every firm that sells AI work sounds confident on a first call. These twenty questions separate the ones who will deliver a working system from the ones who will deliver a slide deck. Ask them of everyone you are considering, us included, and score the answers side by side.

StrikeArc LabsOctober 2026About 30 minutes per call

The scorecard

Ask the questions, then mark each answer.

Mark an answer Strong when it matches what a strong answer sounds like, Partial when it is vague, and Red flag when it worries you or nobody answers. Open a question to see both, and how we answer it ourselves.

Your entries stay in this browser. Nothing you type is sent to us.

01Who are you comparing?

02The work

What they would do first, and what they would leave alone.

Question 01

Which of our workflows would you start with, and why that one?

What to listen for
A strong answer
One workflow, named from what you told them, with its volume, what delay or errors cost, and who owns it.
A red flag
A platform tour, a list of everything AI could do, or a first project picked before anyone asked about your work.
How we answer it
StrikeArc Scout scores every play in the StrikeArc PowerPlays Library against your answers, and a person walks you through a readout with the top plays, the value at stake and the do-not-build list.
Option A
Option B
Question 02

What would you tell us not to build?

What to listen for
A strong answer
A written do-not-build list with a reason for each item, and what would change their mind.
A red flag
Everything is in scope, and every idea is a good one.
How we answer it
The readout names the plays we would tell you to skip. When nothing in our library fits, Custom Discovery ($10,000) scopes a play built for you.
Option A
Option B
Question 03

What do you need from us, and how much of our team's time?

What to listen for
A strong answer
Named roles, the access and examples they need, and the hours, step by step.
A red flag
A few workshops to start, then a discovery phase with no end date.
How we answer it
About 95 minutes before you decide: a 30-minute intro call, about 20 minutes of questions and a 45-minute readout. After you decide: IT access plus one operator, 2 to 3 hours.
Option A
Option B
Question 04

What will we have at the end that we did not have before?

What to listen for
A strong answer
A working system in the tools your team already uses, running on your data, with a named person responsible for it.
A red flag
A report, a roadmap or a prototype that someone still has to build and run.
How we answer it
A working play in the tools your team already uses. A one-time install is handed over to your team at launch.
Option A
Option B

03Data and access

What they would read, who holds the keys, and what happens to your data.

Question 05

What will you read in our systems, and what will you never read?

What to listen for
A strong answer
A written scope, system by system, agreed before anything connects, starting read-only.
A red flag
Broad admin access on day one, or no written scope at all.
How we answer it
You choose what we see, system by system, and nothing is read until you have seen exactly what would be read and said yes.
Option A
Option B
Question 06

Who holds the access you grant, and how do we take it back?

What to listen for
A strong answer
Access you grant per system and can revoke yourself, in one step.
A red flag
Shared passwords, or access that only they can switch off.
How we answer it
Permissions you grant and can revoke, at any time. Turning a connection off takes one step.
Option A
Option B
Question 07

What happens to our data when we stop working together?

What to listen for
A strong answer
A written retention policy with numbers: what is deleted, when, and how you ask for it.
A red flag
No policy, or data kept indefinitely to improve their product.
How we answer it
Connected records are deleted 90 days after the readout unless you become a customer. Everything we hold about your company is deleted within 7 days of a request.
Option A
Option B
Question 08

Will what you learn in our business be used for other companies?

What to listen for
A strong answer
A plain answer in writing, before any data moves.
A red flag
A hedge, or an answer that only exists in the terms of service.
How we answer it
What we learn in your business stays in your business.
Option A
Option B

04Control

What the system may do on its own, and who approves the rest.

Question 09

Can anything send, pay, post or change a record without a named person approving it?

What to listen for
A strong answer
A list of what the system may do on its own under your rules, and what waits for a named approver.
A red flag
Full autonomy offered as the selling point.
How we answer it
Every gate is a named human. Agents draft everything; they approve nothing.
Option A
Option B
Question 10

Who approves a change to how the system behaves?

What to listen for
A strong answer
A change you can read before it goes live, with a named approver on your side.
A red flag
Updates you find out about afterwards.
Option A
Option B
Question 11

What does the system record, and can we read it?

What to listen for
A strong answer
Every action on the record, readable by your team.
A red flag
A system nobody on your side can inspect.
Option A
Option B
Question 12

What happens when the system is unsure?

What to listen for
A strong answer
It hands the item to a person, with the evidence attached.
A red flag
It makes its best guess and moves on.
Option A
Option B

05After launch

Who runs it, who notices when it breaks, and what you own.

Question 13

Who runs the system after launch, and what does that cost?

What to listen for
A strong answer
A named person or team, and a stated price that includes running it.
A red flag
Your team takes it from here, with no handover plan, or a support contract that appears later.
How we answer it
From launch, your Value & Deployment Specialist handles daily operation, maintenance and updates on the subscription. The subscription is $7,500 a month, with one play installed per quarter and run, optimization and ROI reporting on everything live.
Option A
Option B
Question 14

How will you know it has broken or drifted before we do?

What to listen for
A strong answer
It reports on a schedule, a silent system raises an alert, and quality is checked on your own examples.
A red flag
Call us if something looks wrong.
Option A
Option B
Question 15

What happens when the AI models change or get cheaper?

What to listen for
A strong answer
They re-test on your examples before switching, tell you, and pass the saving through.
A red flag
No plan, or you find out from the bill.
How we answer it
We keep each play's cost down as the models change.
Option A
Option B
Question 16

What do we keep if we stop?

What to listen for
A strong answer
Clear terms on your data, your configuration and your documents, and how the handover works.
A red flag
Everything stays with them, or nobody can say.
Option A
Option B

06Money and proof

How they charge, how you leave, and how a result gets counted.

Question 17

How do you charge, and what is included?

What to listen for
A strong answer
A stated price, what it covers, and when billing starts.
A red flag
Open-ended hours, or no price even for work they have done many times before.
How we answer it
Every price is on our pricing page: StrikeArc Scout is free, the library subscription is $7,500 a month, a one-time install is $20,000 per play, and Custom Discovery is $10,000.
Option A
Option B
Question 18

How long do we commit for, and how do we leave?

What to listen for
A strong answer
A stated term and a stated way out, in the proposal.
A red flag
A long lock-in, or a renewal buried in the contract.
Option A
Option B
Question 19

How will we measure what it returned, and who signs the number?

What to listen for
A strong answer
A baseline taken before the work starts, results in dollars or hours, and a named person who signs each figure.
A red flag
Activity counted as results: messages sent, tasks automated, hours of AI usage.
How we answer it
Every play reports its number in dollars, and a person signs it.
Option A
Option B
Question 20

What results can you show us, and where do the numbers come from?

What to listen for
A strong answer
An honest answer either way: results they can name, or worked examples clearly labelled as illustrative, and an offer to measure yours.
A red flag
Return figures with no company, no method and no label.
How we answer it
The worked examples on this site run on made-up companies and say so on every figure. Your results are measured against a baseline taken before anything is installed.
Option A
Option B

How they compare.

Mark an answer and the totals appear here, group by group.

How to use the scorecard.

  1. Send the questions ahead.A firm that knows its answers will put them in writing. Download the PDF and send it before the first meeting.
  2. Ask the ones they skipped on the call.Listen for specifics: a named workflow, a named person, a number with a unit. Vague answers are Partial at best.
  3. Mark each answer the same day.Strong, Partial or Red flag. Mark a question nobody answered as a Red flag; an answer that never arrives is an answer.
  4. Compare group by group.Totals tell you who answered well overall. A red flag on data and access, or on control, outweighs any total, because those are the answers you live with after launch.

Why these twenty.

Most AI projects that disappoint an owner fail in the same places. The first project was picked because it demoed well. Nobody wrote down what the system may touch. Nobody owned it after launch. The result was counted in activity, and the cost of running it arrived later. Each group of questions covers one of those places.

The five ways to get the work done, from a consultant to an in-house hire, each answer these questions differently. The paper AI without an AI hire walks through what each one charges for and what you keep. Question 19 has a paper of its own: measuring AI in dollars.

Ask us the same questions.

Open any question in the scorecard to see how we answer it, in the words our site already uses. A few we answer on the call instead, because the answer depends on the play and your systems. Bring the scorecard to a 30-minute intro call and ask all twenty. Compare the answers to those of the other firms you are talking to.

Keep going.

PaperPDFAI without an AI hire“Do we hire someone, buy software or bring in a partner?”The five ways a company gets AI work done, what each one charges for, what you keep when it ends and who runs the system afterward.Read the paper PaperPDFMeasuring AI in dollars“Every vendor promises a return. How would we know it is real?”How to tell whether an AI system paid: the baseline you take first, the readout a person signs, and the ledger that runs after launch.Read the paper ToolPDF + SpreadsheetThe Workflow Shortlist“Where do we start with AI, and what do we leave alone?”Score up to five workflows on the five tests and get a ranked shortlist, a do-not-build list and a note to bring to the call.Use the tool

The AI Partner Scorecard

Mark each answer S (strong), P (partial) or R (red flag). Score S as 2, P as 1 and R as 0, out of 40. A red flag on data and access, or on control, outweighs any total.

QuestionCandidate 1Candidate 2Candidate 3
The work
Which of our workflows would you start with, and why that one?Strong: One workflow, named from what you told them, with its volume, what delay or errors cost, and who owns it.Red flag: A platform tour, a list of everything AI could do, or a first project picked before anyone asked about your work.
What would you tell us not to build?Strong: A written do-not-build list with a reason for each item, and what would change their mind.Red flag: Everything is in scope, and every idea is a good one.
What do you need from us, and how much of our team's time?Strong: Named roles, the access and examples they need, and the hours, step by step.Red flag: A few workshops to start, then a discovery phase with no end date.
What will we have at the end that we did not have before?Strong: A working system in the tools your team already uses, running on your data, with a named person responsible for it.Red flag: A report, a roadmap or a prototype that someone still has to build and run.
Data and access
What will you read in our systems, and what will you never read?Strong: A written scope, system by system, agreed before anything connects, starting read-only.Red flag: Broad admin access on day one, or no written scope at all.
Who holds the access you grant, and how do we take it back?Strong: Access you grant per system and can revoke yourself, in one step.Red flag: Shared passwords, or access that only they can switch off.
What happens to our data when we stop working together?Strong: A written retention policy with numbers: what is deleted, when, and how you ask for it.Red flag: No policy, or data kept indefinitely to improve their product.
Will what you learn in our business be used for other companies?Strong: A plain answer in writing, before any data moves.Red flag: A hedge, or an answer that only exists in the terms of service.
Control
Can anything send, pay, post or change a record without a named person approving it?Strong: A list of what the system may do on its own under your rules, and what waits for a named approver.Red flag: Full autonomy offered as the selling point.
Who approves a change to how the system behaves?Strong: A change you can read before it goes live, with a named approver on your side.Red flag: Updates you find out about afterwards.
What does the system record, and can we read it?Strong: Every action on the record, readable by your team.Red flag: A system nobody on your side can inspect.
What happens when the system is unsure?Strong: It hands the item to a person, with the evidence attached.Red flag: It makes its best guess and moves on.
After launch
Who runs the system after launch, and what does that cost?Strong: A named person or team, and a stated price that includes running it.Red flag: Your team takes it from here, with no handover plan, or a support contract that appears later.
How will you know it has broken or drifted before we do?Strong: It reports on a schedule, a silent system raises an alert, and quality is checked on your own examples.Red flag: Call us if something looks wrong.
What happens when the AI models change or get cheaper?Strong: They re-test on your examples before switching, tell you, and pass the saving through.Red flag: No plan, or you find out from the bill.
What do we keep if we stop?Strong: Clear terms on your data, your configuration and your documents, and how the handover works.Red flag: Everything stays with them, or nobody can say.
Money and proof
How do you charge, and what is included?Strong: A stated price, what it covers, and when billing starts.Red flag: Open-ended hours, or no price even for work they have done many times before.
How long do we commit for, and how do we leave?Strong: A stated term and a stated way out, in the proposal.Red flag: A long lock-in, or a renewal buried in the contract.
How will we measure what it returned, and who signs the number?Strong: A baseline taken before the work starts, results in dollars or hours, and a named person who signs each figure.Red flag: Activity counted as results: messages sent, tasks automated, hours of AI usage.
What results can you show us, and where do the numbers come from?Strong: An honest answer either way: results they can name, or worked examples clearly labelled as illustrative, and an offer to measure yours.Red flag: Return figures with no company, no method and no label.
Total, out of 40

How to get started

Ask us all twenty on a 30-minute call.

Book an intro call

Bring the scorecard. We will answer every question, including the ones that are hard to answer.