AI readiness assessment: five dimensions to score before you fund a pilot

Score your data, systems, workflows, people and governance against what a production AI workflow needs. You come away knowing what to build first, what to fix before you build, and what to stop funding.

Readiness scorecardSample result

61/ 100

  • Data quality and governance72%
  • Systems and integrationHigh risk
  • Workflow stabilityStable
  • People and adoption58%
  • Governance and security57%

First fix Let workflows write back to the ERP before any build starts

The five dimensions of AI readiness

AI programs usually fail where the model meets your data, your systems and the people who use it. Each of those gets its own score.

1. Data

Data quality and governance

Do the records a use case needs exist? Do your departments agree on them? Can a workflow reach them through an API? When three teams keep conflicting customer records, no model gets accurate enough to rely on. Often the fix is an enterprise context layer.

2. Systems

Systems and integration

Use cases that look profitable on paper fail on limited ERP access, slow APIs and systems that can be read but not updated. Map each system down to its APIs and data structure, including SAP, Odoo and HubSpot. Then test whether an n8n orchestration layer can write results back.

3. Workflows

Workflow stability

Automation pays off on high-volume work that follows the same rules every time. Automating a broken process only spreads the damage. Invoice processing is a typical fit, with stable rules and a clear before and after. That is why we run AI invoice automation as its own service.

4. People

People and adoption

A system nobody uses fails as surely as one that breaks. Name who owns each outcome and which teams will use the workflow every day. A skills gap found here becomes the brief for corporate AI training built around your own processes.

5. Governance

Governance and security

Score who can access what, which data the AI may use and where a person must approve. Measure them against the NIST AI Risk Management Framework and ISO/IEC 42001. The findings carry into your enterprise AI governance model.

Score it

Turn your answers into a number

The AI Readiness Scorecard asks the ten questions below. It returns a score out of 100 and names the dimension holding you back.

Open the scorecard

AI readiness checklist for your operations and IT leads

Answer each question for the process you would automate first. Every "no" names the dimension to fix before you build.

Data

  1. Can you name the system that holds the master copy of every field the use case reads?
  2. Do your departments agree on the same customer and product records?

Systems

  1. Does each system involved have an API you can read from?
  2. Can a workflow write its results back without anyone typing them in again?

Workflows

  1. Does the process run the same way every week?
  2. Is the volume high enough that hours saved repay the build?

People

  1. Is one named person accountable for the outcome?
  2. Will the team that uses the workflow help design it?

Governance

  1. Have you defined which data the AI may never see?
  2. Does a person approve any action before it reaches a customer?

The questions are Ovidius's own framing of the five dimensions. If your team cannot answer half of them, the AI Audit answers them on your systems. For a first AI agent, the 20-question AI agent readiness checklist goes further. It covers what the agent itself may read, change and do.

Where your company sits on the AI maturity curve

AI maturity runs in five stages. The ten questions above test whether you are ready to move from pilots to live workflows, the step from Active to Operational.

  1. 01
    Stage 1

    Awareness

    AI is discussed, and nothing runs.

  2. 02
    Stage 2

    Active

    Pilots and experiments, with nothing in a live workflow yet.

  3. 03
    Stage 3

    Operational

    AI runs inside live workflows, with an owner for each.

  4. 04
    Stage 4

    Systemic

    AI runs across departments under one set of oversight rules.

  5. 05
    Stage 5

    Transformational

    AI shapes the business model itself.

Four studies, four readiness gaps

Each figure below matches a dimension that a readiness check tests before you spend money on a build.

FigureWhat it measuresDimensionSource
80%+of AI projects fail, twice the rate of IT projects without AIPeople and adoptionRAND Corporation, 2024
88%of AI proofs of concept never reached production: 4 out of 33Systems and integrationLenovo CIO Playbook 2025, with IDC
95%of generative AI pilots show no measurable P&L impactWorkflow stabilityMIT NANDA, The GenAI Divide, 2025
97%of organizations with an AI-related breach lacked proper AI access controlsGovernance and securityIBM, Cost of a Data Breach Report 2025

When a self-check is enough, and when to bring in engineers

The checklist tells you where to look. The AI Audit does the looking, down to each system's APIs and data structure. It ends in a plan your engineers can start on.

AI AuditThis checklist
EvidenceA detailed map of your systems, their APIs and how their data is structuredYour team's own answers
OutputReadiness scorecard, ranked use cases, ROI models for your top 3 to 5 and a 90-day roadmapThe dimension to fix first
Time4 to 8 weeks, in eight phasesA meeting with your operations and IT leads
Cost$10,000 to $100,000, credited toward your first buildYour team's time
AfterwardsThe same team scopes and builds your first workflowYou decide the next step

What passing all five looks like in production

Pinkcube's order data was already reachable through the HubSpot and AfterShip APIs. Its chat agent went from kickoff to live in seven weeks.

PinkcubeB2B e-commerce, chat agent
1,800
Website chats a month, answered by hand before launch

One HubSpot chat widget now answers customers at every hour of the week

The agent runs on n8n with live order data from HubSpot and AfterShip. A second AI checks each reply before it goes out.

500chats a week, approximately
95%+order lookups that succeed
12production systems we run for Pinkcube
Read the Pinkcube case

AI readiness assessment FAQ

Ask us something else

What is an AI readiness assessment?

It measures whether your organization is ready to run AI in production. It scores five areas: data, systems and integration, workflow stability, people and adoption, and governance and security. It ends with a ranked list of use cases and the gaps each one has to close first.

What does an AI readiness framework cover?

It covers the five dimensions on this page, checked against the NIST AI Risk Management Framework and ISO/IEC 42001. In the AI Audit, the framework runs across eight phases, from your strategy to the final roadmap.

What are the five stages of AI maturity?

The five stages are Awareness, Active, Operational, Systemic and Transformational. The assessment places you on that curve and names what it takes to reach the next stage.

Can we run the assessment ourselves?

Yes, as a first pass. The checklist above and the AI Readiness Scorecard show which dimension holds you back. To measure it on your own systems, with ROI models and a roadmap, you need the AI Audit.

Do you need access to our live data?

No. We sign an NDA and a data governance agreement before we access any system. We then review your architecture, metadata, what your APIs can do and sample data structures. We work without live customer records or your proprietary source code.

How do we turn a low score into a plan?

Start with the lowest-scoring dimension, because that is your first fix. To measure all five on your systems and leave with a ranked plan, book the AI Audit. It takes 4 to 8 weeks and costs $10,000 to $100,000, and the fee is credited toward your first build.

After the checklist

Measure

AI Audit

Eight phases across strategy, operations, data and systems, ending in a ranked 90-day roadmap.

See the AI Audit
Plan

Enterprise AI roadmap

The AI Audit's final output, which sorts its findings into quick wins, strategic bets and the ideas to stop funding.

See a sample roadmap
Prove

AI pilot

Your top use case as a working prototype on your own data, in 2 to 6 weeks.

See the AI pilot

Guides on AI readiness

Twenty questions to answer before the first agent, five maturity levels, and the people side most plans leave out.

Checklist

AI Readiness Checklist: 20 Questions Before Your First Agent

An AI readiness checklist of 20 questions on data, systems, workflow, people and governance. It comes with a scoring tool and four checks to pass before your first agent.

Read the guide
Guide

AI Maturity Assessment: Where Your Company Sits on Five Levels

Run an AI maturity assessment across data, systems, workflows, people and governance, see which of five levels you sit at, and what moves you up one level.

Read the guide
Guide

Organizational AI Readiness: The People Side Nobody Plans For

Organizational AI readiness: map what changes in each role, name an owner, train on real tasks, answer job fears plainly and run a feedback loop at go-live.

Read the guide

Where this fits

To score your own readiness first, take the AI readiness scorecard.

Proof

Systems we built

  • Pinkcube: chat coverage from 45 to 168 hours a week
  • IES Limited: a funnel drop spotted in 30 minutes instead of 20 days

Put your first workflow on the board

Answer a few questions about your team and what you want automated, then a senior partner scopes it with you on the call.