AI Readiness Assessment & Roadmap | AI Consulting | ATH Infosystems
Book an AI Strategy Call

Know where AI will pay off before you spend on it

In two weeks, ATH scores your AI readiness, ranks your use cases by return and gives you a roadmap that starts with one fixed-price pilot.

Book an AI assessment call Check your readiness
  • 2 weeks
  • Fixed price
  • NIST AI RMF aligned
  • ISO/IEC 42001 aligned
  • Vendor neutral

AI opportunity scorecard

6 use cases ranked
  1. 1Invoice matching and approvalValue 9 · Feasibility 8 · Clean data, low riskPilot first
  2. 2Customer support triageValue 8 · Feasibility 7 · Needs help-desk integrationNext
  3. 3Sales follow-up draftingValue 6 · Feasibility 9 · Quick winNext
  4. 4Contract reviewValue 8 · Feasibility 4 · Contracts spread across systemsFix data first
  5. 5Demand forecastingValue 7 · Feasibility 4 · Two years of history neededLater
  6. 6Fully automated pricingValue 5 · Feasibility 3 · Risk too high without controlsNot yet
Example scorecardScored out of 10 with your team

Most AI spending is not paying off yet

Nearly every company now uses AI somewhere. Very few can point to the profit it produced. The difference is rarely the model. It is choosing the right use case and being ready for it.

95%

of organizations studied saw no measurable business return from their generative AI investments.

Source: MIT Project NANDA, The GenAI Divide (2025)
1 in 3

About one-third of companies have begun scaling AI, although 88% use it in at least one business function.

Source: McKinsey, The state of AI in 2025
5%

of companies get bottom-line value from AI at scale, while 60% report minimal gains.

Source: BCG, The Widening AI Value Gap (2025)
40%+

Gartner predicts over 40% of agentic AI projects will be canceled by the end of 2027 because of rising costs, unclear business value or weak risk controls. An assessment tests all three before you commit a budget. Source: Gartner, June 2025

Check your AI readiness in one minute

Rate your company from 1 (not started) to 5 (fully in place) on the six areas we assess. Your score updates as you move the sliders.

Leadership has agreed AI goals, a budget and an owner.
Key data is accurate, in one place and easy to reach.
Core systems are in the cloud and have APIs AI can connect to.
Teams use AI tools today and are trained to work with them.
There is an AI policy, a risk review and clear rules on data use.
You have a list of AI use cases with owners and expected value.
42/ 100

Your AI readiness score

Your stage

Piloting

You have started. The risk now is pilots that never reach production.

Start here

Data

Pick one use case and clean only the data it needs. Do not wait for a company-wide data project.

Get the full assessment A quick self-check, not the full assessment. Nothing you enter here is stored or sent anywhere.

Six areas, one clear picture

The full assessment goes far deeper than a self-check. We interview your people, look inside your systems and score each area against evidence.

AreaWhat we examineWhat you learn
Strategy and valueBusiness goals, budget, executive sponsorship and how AI success will be measuredWhether AI is tied to outcomes the business cares about
DataQuality, location, ownership, access rights and sensitivity of the data each use case needsWhich use cases your data can support today, and what to fix for the rest
TechnologyCloud platform, core applications, APIs, integration options and current AI toolsWhat AI can connect to now and what would block it
People and skillsCurrent AI use, skills, capacity of process owners and appetite for changeWhere training is needed and who should own each use case
Governance, risk and securityAI policy, approvals, privacy, security controls and the regulations that apply to youThe controls each use case needs before it goes live
Use-case portfolioEvery candidate use case across departments, with volume, cost and pain pointsA ranked list, with a business case for the top few

How we rank your use cases

Every idea is scored on two things: how much it is worth and how hard it is to do. The result is one picture your leadership team can agree on.

Prepare

High value, hard today. Fix the data or integration gap, then build.

  • Contract review
  • Demand forecasting

Pilot first

High value and ready now. One of these becomes your first pilot.

  • Invoice matching
  • Support triage

Park

Low value and hard to do. Review again in a year.

  • Automated pricing

Quick wins

Easy, modest value. Good for building skills and confidence.

  • Sales follow-ups
  • Meeting summaries

Value is scored on

  • Hours and cost saved each year
  • Revenue gained or protected
  • Errors, delays and rework avoided
  • Effect on customers and staff

Feasibility is scored on

  • Whether the data is ready
  • How many systems must be connected
  • Risk and regulatory exposure
  • How much the team’s work has to change

Each top use case also gets an estimated payback period, a running cost per task and a risk tier. Use cases shown are examples.

A roadmap built for where AI is going

AI is moving from assistants that answer questions to agents that complete work. A plan written for last year’s tools will be out of date before it is funded, so we also test how ready you are for what comes next.

Already know what you want to build? See AI agent development.

Agent readiness
Which processes can be handed to AI agents, what tools and permissions they need and where a person must approve.
Model portability
An architecture that lets you switch between OpenAI, Anthropic, Google and open-source models as prices and quality change.
AI cost control
A forecast of cost per task and per month for each use case, so usage bills do not surprise finance.
Build, buy or partner
For each use case, whether an off-the-shelf product, a custom build or a managed service gives the best return.
Workforce plan
How roles change when agents take on routine work, and the training each team needs.
Regulation watch
Which current and upcoming AI rules affect your roadmap, built into the plan from the start.

Grounded in recognized U.S. standards

Our method follows the frameworks U.S. regulators, auditors and enterprise buyers already use, so the results hold up when someone asks how you decided.

StandardHow we use it
NIST AI Risk Management FrameworkIts four functions, Govern, Map, Measure and Manage, shape how we score governance and rate the risk of each use case
NIST Generative AI Profile (AI 600-1)A checklist of risks specific to generative AI, such as made-up answers and data leakage, applied to every generative use case
ISO/IEC 42001The international standard for AI management systems. We show how far you are from it, which is useful if customers ask for certification later
HIPAA, GLBA, SOX and PCI DSSWhere your industry rules apply, each use case is checked against them before it is ranked
State AI and privacy lawsWe flag use cases that make decisions about people, such as hiring, lending or insurance, where state laws add duties

The assessment identifies gaps against these frameworks. It is not a certification or legal advice. For controls and audit evidence, see AI security and compliance.

Two weeks from kickoff to roadmap

A fixed scope, a fixed price and a small demand on your team’s time.

  1. Discover

    Kickoff with your sponsor, then short interviews with the people who own each process.

  2. Examine

    We review the data, systems and controls behind each candidate use case and score the six areas.

  3. Rank

    We score every use case for value and feasibility and build the business case for the top few.

  4. Decide

    An executive readout: your scorecard, your roadmap and a scoped first pilot ready to approve.

Typical timeline for a mid-sized company or one business unit. Larger scopes are agreed before the start.

What you walk away with

Eight documents that are yours to keep and use, with ATH or without.

  • AI readiness scorecardYour score in each of the six areas, with the evidence behind it
  • Ranked use-case portfolioEvery use case scored for value and feasibility on one page
  • Business case for the top use casesBaseline, expected savings, build and running cost, and payback period
  • Data readiness reportWhat data each use case needs, its condition and what to fix
  • Risk and compliance mapThe risk tier of each use case and the controls it needs
  • Target architectureHow AI connects to your systems, with model and platform recommendations
  • 12 to 18 month roadmapWhat to do, in what order, with owners and decision points
  • Pilot planScope, success measures, timeline and a fixed price for your first pilot

Savings and payback figures are estimates based on the baseline we measure with your team. The pilot is where they are confirmed.

Your roadmap, in three horizons

The plan starts small on purpose. Each stage has to earn the next one.

First 90 days

Prove

Run one pilot on the top-ranked use case and measure it against the baseline.

  • Six-week fixed-price pilot
  • Real data, real users
  • Go or no-go on evidence

Months 3–6

Scale

Move the pilot into production and start the next two or three use cases.

  • Fix the data gaps found
  • Put AI policy and reviews in place
  • Train the teams involved

Months 6–18

Transform

Redesign whole processes around AI agents that work across departments.

  • Agents that hand work to each other
  • Shared AI platform and controls
  • Roadmap reviewed every quarter

Answers for every leader at the table

The readout is written so each executive finds their own question answered.

LeaderTheir questionWhat the assessment gives them
CEO and COOWhere should we use AI first, and what will it change?A ranked list of use cases and a staged plan tied to business goals
CFOWhat will it cost and when does it pay back?A business case per use case, with build cost, running cost and payback period
CIO and CTOCan our systems and data support it?A data readiness report and a target architecture that avoids vendor lock-in
Compliance and riskIs it safe and allowed?A risk tier for each use case and the controls required before launch
Department headsWhat does it mean for my team?The tasks AI would take on, the ones that stay with people and the training needed

Where a good first use case leads

Results from AI projects ATH has delivered after choosing the right starting point.

AI assessment FAQs

Something else? Ask us on a call.

What is an AI readiness assessment?

It is a short, structured review of how prepared your company is to get value from AI. We look at six areas: strategy, data, technology, people, governance and your list of possible use cases. You get a score for each area, a ranked list of AI use cases with estimated return, and a roadmap that tells you what to do first.

How long does the assessment take?

Two weeks for a typical mid-sized company or a single business unit. Larger organizations with many departments or locations may need longer, and we agree the timeline with you before we start.

How much does it cost?

The assessment is fixed price. We agree the scope and the price on a call before any work starts, so there are no hourly bills or surprises.

What do you need from us?

An executive sponsor, short interviews with the people who own the processes we review, and a read-only look at the systems and data involved. We keep the load on your team small and schedule interviews around their work.

What if the assessment shows we are not ready for AI?

Then you have saved the cost of a failed project. The roadmap will list the specific gaps, such as data quality or missing approvals, and the order to fix them in. Most companies have at least one use case that is ready now, even when others need groundwork first.

Do we have to build with ATH afterwards?

No. The scorecard, the ranked use cases and the roadmap belong to you. They are written so your own team or another partner can act on them. If you want us to build the pilot, it is a separate fixed-price project.

We already use Microsoft Copilot or ChatGPT. Is an assessment still useful?

Yes. Giving staff an AI assistant is a good first step, but it rarely changes a whole process. The assessment shows where AI agents and automation connected to your own systems would return the most, and whether your data and controls are ready for that step.

How do you estimate return on investment?

For each top use case we measure today’s baseline with your team: volume, time per task, error rate and cost. We then model the realistic share AI can handle, the cost to build and run it, and the payback period. These are estimates. The pilot is where they are proven with real numbers.

Find your best first AI use case in 30 minutes

A free call with a senior AI consultant. You’ll leave with:

  • Two or three use cases worth assessing in your business
  • A view of what could hold them back
  • The scope and fixed price of an assessment for you