AI consulting

AI consulting services that show which jobs are worth automating

A short paid review of how your team works, ending with two or three use cases priced against the hours they save, and the ones to skip.

Book the AI review
Trusted by the world's fastest-growing brands

What AI consulting is

A short review that tells you where AI pays and where it doesn’t

AI consulting services here means a short paid review of how your team works. It ends with two or three use cases priced against the hours they would save, the ones to skip, and what each would cost to build.

AI review · Sample Dental CenterRecommendation5 tasks · 24 hBuild1Appointment reminders by text5 h a week2Intake forms into the system3 h a weekSkipInsurance verificationeach payer differsTreatment plan callsa person should call

Your work, mapped first

We list the tasks that fill your team’s week, how often each happens and how long it takes. The hours come from your own people.

Use cases priced in hours

Each candidate is scored on hours saved, data readiness and risk, then priced. Two or three make the list, with the cost of building each.

The skips, named

Some work should stay with a person, and some isn’t worth its price yet. The review says which, so you don’t spend on them.

Review first or build first

Two ways to start with AI

Both end with something running. The difference is whether the choice of what to build is made before the money is spent or after.

Starting with a buildBuild, then learnA paid AI reviewDecide, then build
First stepPick a tool or a use case and beginMap the work and score each candidate use case
What you can skipFound out during the build, after the spendUse cases that fail on data, risk or hours saved
How it’s pricedOne price for the build that was chosen firstA written scope for the review, then a price per use case
Who it suitsA team that already knows the one job to automateA team with several ideas and no way to rank them
What you hold afterwardThe finished build and its documentationA ranked list, a risk memo and a pass line for each build
Do you need the review?

If you can name the one job, say who does it today and how long it takes, you can go straight to a build from $3,500 per workflow. The review is for the team with a list of ideas, a vendor pitch on the desk, or no idea where to begin. We say which one you are on the first call.

What the sources say

Four facts that shape AI work in 2026

What NIST and the FTC say about AI risk and AI claims, with the source under each.

  1. FrameworkNIST

    The four functions in NIST’s AI Risk Management Framework are govern, map, measure and manage.

    NIST released AI RMF 1.0 on January 26, 2023, for voluntary use. It is a guideline, not a law, and NIST says it is being revised. We borrow its four functions as the headings of the risk memo, so the memo follows a structure a board or an insurer already knows.

  2. Language modelsNIST

    NIST publishes a separate profile for the risks of generative AI.

    NIST AI 600-1, the Generative AI Profile, came out on July 26, 2024. It helps organizations identify risks that are specific to generative AI and proposes actions to manage them. A use case built on a language model is read against it.

  3. AI claimsFTC

    The FTC acts against businesses that overstate what their AI does.

    Its September 25, 2024 sweep, Operation AI Comply, named five cases, including a “robot lawyer” service that paid $193,000. For a buyer, it is a reason to ask any vendor, ours included, what a tool was tested on before it was sold.

  4. AccuracyFTC

    A claim about how accurate an AI tool is needs evidence to back it.

    The FTC said Workado advertised its AI content detector as 98 percent accurate, while independent testing found 53 percent on general-purpose content. Its final order, August 28, 2025, requires competent and reliable evidence for such claims. Our evaluation plan sets the pass line before anything is built.

The scope

What the review produces

Six parts of the work. Each ends in a document you can open, check and forward.

Task map and scores24 h a weekInsuranceverification calls7 hSkipConfirmingappointments6 hBuild firstTreatmentplan calls5 hSkipIntake forms4 hBuildWeeklyreport2 hLater

Process map and use-case scores

We map the tasks that fill your team’s week and score each candidate on hours saved, data readiness and risk. You receive: a ranked list of use cases, with the reason each one sits where it does.

HoursDataRisk
Data readiness1 of 4 readyIntake forms into the systemWhere it livesweb form and paperWho can see itfront desk onlyHow clean it ispaper is handwrittenWhat permits its useHIPAA: ask counselFix firstpaper intake to the web form

Data readiness check

For each use case we look at where the data lives, who can see it, how clean it is and what permission covers using it. You receive: a one-page readiness note per use case, with what to fix first.

Build or buyAppointment remindersOff the shelfConfiguredCustomCost to startlowerhigherWhere your data goesto the vendorstays in your accountsWhat you can changesettings onlyall of it

Build or buy

Where a product already does the job, we say so and name the trade-offs: cost, where your data goes and what you can change. You receive: a build-or-buy note for each use case, including the ones where buying wins.

Off the shelfConfiguredCustom
Risk and policy memoNIST AI RMFImpactLikelihood12341Sees more than it needsMap · intake fields only2Acts without approvalGovern · a person approves3Copies a field wrongMeasure · 57 of 60 to pass4Can’t read a formManage · sent to a person

Risk and policy memo

We write down what the tool may see, what it may never do, who approves its output and what happens when it is wrong. The memo follows the four NIST functions. You receive: a risk memo and a short staff policy you can adopt or edit.

Evaluation planSet before the buildReminders by textIntake formsTest set60 real formsPass line57 of 60or betterUnclear forms go to a person15 of 15

Evaluation plan

Before any build, we agree the test set, the pass line and who reads the results, so accuracy is a number you can read. You receive: an evaluation plan for each use case you approve.

Priced plan2 builds from $7,000Value at $30 an hourBuild priceAppointment reminders by text250 h a year$7,500from $3,500Intake forms into the system150 h a year$4,500from $3,500Weekly reportWaits$3,000from $3,500

A sequenced plan with prices

The use cases that cleared the review are put in order, each priced against its hours saved, with the skips listed under them. You receive: a one-page plan and a fixed price for each workflow you choose to build.

How it runs

Map, score, price, set the pass line

You approve the scope before any work starts. Each stage hands you a map, a list or a plan to keep.

task-map.csv

Task map

Sample · Sample Dental Center
TaskHours a weekSteps
Confirming appointments64
Copying intake forms into the system45
Insurance verification calls76
Treatment plan follow-up calls54
Building the weekly report by hand27
Five tasks, 24 hours a week. The next stage decides which of them AI should touch.

Our stack

The tools we test with

The models and tools a use case gets tested on.

OpenAI
Why OpenAI

OpenAI’s models read, write and summarize language, and also handle images and audio.

How we use it

We run your own sample questions and documents through its models and record how many each one gets right.

ReasoningTool useRange
OpenAIBroad
Open modelsMid
Single vendorNarrow

Questions

AI consulting, answered

Start here

Find the AI use cases worth paying for

Tell us where your team loses time. We’ll send back a written scope for the review, with its price, before you agree to anything.