MEMO · TO leaders who don't yet know where they stand · RE a five-dimension readiness score
Resources AI Readiness Self-Assessment
AI Readiness Self-Assessment
Five questions, one for each dimension of the readiness framework Praxora Lab publishes. Answer honestly and you'll get a scored result, a single prioritised next step, and a branded PDF report you can take into a budget conversation, not a generic write-up.
Why this matters
Most AI failures are organisational, not technical
Adoption of AI is no longer the hard part. Most businesses have already tried a tool, run a pilot, or asked a vendor for a demo. What decides whether that turns into something a board can rely on is not the model. It is whether the five things this assessment scores, data, process, governance, capability and measurement, were in place before the pilot started.
Skipping that check is expensive precisely because it is invisible until it isn't. A gap in data readiness doesn't show up as an error message. It shows up months later as an output nobody fully trusts, a review step nobody owns, or a business case nobody can defend when a board asks what changed. Checking first is the cheaper failure mode to discover.
This assessment exists to surface that gap before it costs a budget cycle. Five questions, one per dimension, score where the business actually stands today, not where a vendor's pitch deck assumes it stands, and point to the single dimension worth fixing first.
Exhibit · Why readiness gets checked first
Adoption is nearly universal. Being ready for it is not.
-
78%
of companies use AI frequently, per McKinsey
-
33%
have scaled it company-wide, per McKinsey
-
5%
of pilots produce lasting profit, per MIT NANDA
-
60%
of AI projects will be abandoned for lack of AI-ready data by 2026, per Gartner
Your result
Your five-dimension shape
Section 02 · What your score means
Section 03 · The business impact
Your immediate action plan · next 14 days
Grounded in
95%
of generative AI pilots fail to produce a measurable profit-and-loss result
MIT NANDA, The GenAI Divide, 2025
60%
of AI projects fail on poor data foundations, and over 40% of agent projects fail on legacy-system incompatibility, both detectable before a dollar is spent
Gartner, 2025
74%
of companies have yet to show tangible value from their AI investment, despite the spend behind it
BCG, Where's the Value in AI?, October 2024
33%
of companies have scaled AI company-wide, against 78% who use it frequently, the gap this framework exists to close
McKinsey, The State of AI, 2024
Section 04 · Score by dimension
Each dimension is self-rated on a fixed 0–3 rubric, not a subjective impression. A score of 0–1 signals a deployment blocker; 2–3 signals the dimension is on track. Each definition below is paired with a named, publicly reported research finding for context, not to describe your organisation specifically.
Definition — Whether the data behind the process you want AI to touch is accessible, documented, current, and owned by someone specific.
Why it matters — Every other dimension depends on this one. If the underlying data cannot be trusted, nothing built on top of it, including governance and measurement, can be trusted either.
Research 60% of AI projects will be abandoned by 2026 for lack of AI-ready data, Gartner projects. Gartner, February 2025.
Definition — How consistently and explicitly the target process is documented, versus how much it still depends on individual judgement.
Why it matters — AI automates what is defined. An undocumented process does not become predictable because AI is layered onto it. It becomes unpredictable faster.
Research 2 in 3 leaders call their own organisation overly complex, and simplifying the process is the bigger lever, McKinsey finds, not layering technology onto it unchanged. McKinsey, The State of Organizations 2026.
Definition — Whether a named reviewer and a defined escalation path exist for catching and correcting AI output before it reaches a customer or a decision.
Why it matters — Governance determines how much damage an error does before someone notices it. It is the control layer that makes fast deployment survivable rather than reckless.
Research 63% of companies using generative AI have no governance structure in place to manage the risks it creates. McKinsey, The State of AI, 2024.
Definition — Whether staff working with the system can independently operate, review, question and correct its outputs, without outside help.
Why it matters — A capable team is the last line of defence against a silent error. Without it, oversight exists on paper only, and a black box nobody understands is a dependency, not a capability.
Research 35% of staff received AI training in the past year, even though 75% of companies are adopting AI. Randstad, AI Skills Gap, November 2024.
Definition — Whether the business outcome the AI system is meant to move is baselined before deployment begins, so any later change can be credibly attributed to it.
Why it matters — Without a baseline set before deployment, no later improvement can be credibly attributed to the AI system at all, and the business case becomes unverifiable after the fact.
Research 74% of companies have yet to show tangible value from their AI investment, despite the spend behind it. BCG, Where's the Value in AI?, October 2024.
Section 05 · What to do in each dimension
Data readiness
Data underpins every other dimension. Before anything else, get one process's data into a single, documented, current source, even if the rest of the estate stays messy for now.
Process definition
An undocumented process does not become predictable because AI is layered onto it. It becomes unpredictable faster. Write the process down before you automate it.
Governance structure
Name a reviewer and a defined escalation path before anything goes live, not after something goes wrong.
Team capability
A tool nobody on staff can operate, review or correct is a dependency, not a capability. Build the internal skill before you scale the deployment.
Measurement
Without a baseline set before deployment, no later improvement can be credibly attributed to the AI system at all. Set it now, before the next pilot starts.
Want the team to follow up on this result? Leave your email (optional, not required for anything above).
Reference
This assessment is adapted for Praxora Lab from the original AI Readiness Self-Assessment at terencekok.com/resources › . The five-dimension framework it scores is set out in full in Is your business ready for AI? Start with these five questions › .
