AI readiness is whether your data, governance, skills, and use-case pipeline can actually run agents in production — not whether you've run a demo or bought a license. Published frameworks disagree sharply on how to measure it: from 3 dimensions to 10 across the models we checked. Below: the tools you can use today, what the frameworks agree on underneath the disagreement, and how to act on your answer. Data as of July 2026.
Three things exist today: a paid one-week audit that scores your use cases, a free self-run assessment, and a free cost estimator for closing the gap. A scored, instant quiz is planned but isn't built yet — don't wait for it to start.
We're building a scored, interactive quiz for this hub — the kind Microsoft, Cisco, and Avanade already run on their own sites (see the sources below). It isn't live yet. Until it is, the assessment, the audit, and the estimator above are the real entry points; nothing on this page depends on the quiz existing.
AI readiness is an organization's preparedness to run AI — specifically agents that take actions, not just chat — in production, measured across data quality, infrastructure, governance, skills, and change management (Genesys). It is a preparedness question, not an activity log.
It is notthe same thing as having run a pilot. Companies pilot constantly without being ready to scale: it's exactly why so many pilots stall before production, a pattern we track in detail separately. It is also not an IT-only question — every framework we checked treats governance, talent, and culture as load-bearing, not just infrastructure. And it is not a purchase: buying a copilot license or an agent platform doesn't make an organization ready any more than buying a gym membership makes someone fit.
Readiness is also adjacent to, but distinct from, a question this whole site is built around: who runs your AI program day to day. Readiness describes the organization's fitness to operate agents; agent leadershipis the specific person and mandate that owns making that fitness real. You can be highly ready and still have no one accountable for using that readiness — which is its own failure mode.
There is no single answer. We checked eight published models and the dimension count ranges from 3 to 10— a spread wide enough that quoting one framework as “the” readiness model would be misleading. Here is what each actually publishes:
| Framework | Count | What it measures |
|---|---|---|
| Avanade + Microsoft (Avanade) | 3 | People, processes, platforms |
| MIT CISR (MIT Sloan, Feb 2025) | 4 stages | Processes, technology, organizational culture |
| Microsoft AI Readiness Wizard (Microsoft) | 5 | Business strategy, tech & data strategy, AI strategy & experience, org & culture, governance & security |
| “5P” framework (Future Processing) | 5 | Purpose, people, process, platform, performance |
| Cisco AI Readiness Index (Cisco, 2025) | 6 | Strategy, infrastructure, data, governance, talent, culture |
| Gartner AI Maturity Assessment (Gartner, Apr 2025) | 7 | Strategy, value, organization, people & culture, governance, engineering, data |
| Accenture + Carnegie Mellon SEI (SEI, Jun 2026) | 8 | Org strategy, workforce & culture, workflow re-engineering, risk & governance, data, engineering, operations, ecosystem |
| Ten Dimension Framework (Digital Education Council) | 10 | Not public without the full report; developed with 27 universities across 17 countries |
Data as of July 2026.
Line them up and a pattern holds regardless of count: strategy, data, governance, and people/culture appear in every single framework above, just worded differently — Cisco's “talent” is Microsoft's “organization & culture” is Gartner's “people & culture.” Where they diverge is the extras: Gartner and Accenture both split out engineering execution as its own axis; Cisco and Gartner both separate infrastructure/data into two; only Accenture names vendor and partner posture (“ecosystem”) as a distinct dimension.
We didn't pick five because five is a round number. Five is what's left when you intersect all eight models above and drop what each disagrees on: strategy & use-case portfolio, data & infrastructure, governance & risk, people, culture & skills, and technical execution.That's the shape of the thing; the scored, dimension-by-dimension version of it is what the assessment actually runs.
Match ambition to readiness, not to what's exciting. Low-readiness organizations should run small, contained pilots that build the operating muscle; only high-readiness organizations should bet on production-scale use cases. Readiness itself predicts whether a pilot ever ships.
Cisco's 2025 Index found that “Pacesetters” — the most-ready 13% of organizations worldwide, a share that has held for three straight years — are 4x more likely to move AI pilots into production and 50% more likely to report measurable value than everyone else (Cisco, 2025). Within that top group, 77% have already moved a use case into production, against just 18% of companies overall — readiness isn't a nice-to-have credential, it's the thing that determines whether the project you're about to fund ever ships.
Most companies are earlier than they think. In MIT CISR's 721-company survey, 28% were still at Stage 1 (experiment and prepare) and 34% at Stage 2 (build pilots and capabilities)— 62% combined — while only 7% had reached the top, “AI future-ready” stage (MIT Sloan, Feb 2025). If that's honestly where you sit, the highest-ROI move usually isn't your most ambitious use case — it's the smallest one that proves the operating model (data access, oversight, a named owner) actually works, before you fund something harder.
Which specific use case to prioritize — voice customer service, invoice processing, lead generation, and dozens more — is a separate question from whether you're ready to run one at all. We break down the actual use-case menu and its economics on the use-case hub; this page is about the readiness gate that decides which ones you can safely attempt yet.
Low-readiness organizations should buy or embed rather than build — a custom build assumes the data access, engineering capacity, and governance a low-readiness org typically doesn't have yet. High-readiness organizations can justify building where the use case is core to their advantage; everyone else is usually better served renting the capability until the foundation catches up.
That calculus is shifting fast, which changes the math even for companies that assumed they'd need to build. Gartner projects that 40% of enterprise applications will ship with built-in task-specific AI agents by the end of 2026, up from under 5% in 2025 (Gartner, Aug 2025). For a growing share of use cases, “build vs. buy” is quietly becoming “it's already inside a tool you already pay for” — worth checking before scoping a custom build, not after.
We don't sell implementation, so we don't have a stake in pushing either answer. The build-vs-buy-vs-hire call is one of the four scored deliverables in the Agent Readiness Audit— made against your specific use cases and your specific readiness level, not a generic rule of thumb.
Our own audit is a fixed-scope, one-week engagement — $4,900, $3,500 for the first five clients in exchange for a testimonial— that produces four things: a use-case map scored by ROI and feasibility, a governance gap assessment, a priced build-vs-buy-vs-hire call for each priority use case, and a 90-day roadmap with a named owner. Full detail is on the audit page.
It runs entirely over calls and shared documents — no production system access required — and it is a diagnostic and a plan, not an implementation. It exists for the company that's past “should we use agents?” and into “which ones, run by whom, and how do we not get burned,” whether or not you already have someone leading AI internally.
The roadmap always names an owner. What that role should cost — full-time or fractional — is a separate, sourced question we answer on Chief AI Officer salary and size for your situation on the cost estimator; if the audit fee later credits against a hire made through us, that's covered on the audit page too.
What's the difference between this page and the AI readiness assessment?
This page is the map: what readiness means, what it's measured on, and which tool fits where you are. The assessment runs the actual test — dimension by dimension, with a score you can act on. Start here if you're not sure yet; start there if you already know you need the scored version.
Is there a free, instant readiness quiz?
Not yet. A scored, interactive version is planned but not built. Until it ships, your working options are the self-run assessment, the paid audit, and the cost estimator — all live today.
Is a readiness score comparable across companies or industries?
No. Published frameworks disagree on how many dimensions to score — from 3 to 10 — so a “7 out of 10” means something different depending on whose test produced it. Track your own score's change over time, not your score against a competitor's.
Do we need clean data before we start an assessment?
No. The assessment is what surfaces data gaps — you don't need to have solved them first. Most low-readiness organizations find data access and quality is exactly the gap the process is supposed to catch.
Do we need to be enterprise-scale to do this?
No. Every tool on this page is built for a company past “should we use agents?” and into “which ones, and are we actually set up to run them” — a founder-led startup and a multi-thousand-person enterprise both qualify.
How often should we reassess our readiness?
We don't have a sourced figure for an ideal cadence, so we won't invent one. In practice, reassess before each major new production bet, not on a fixed calendar — readiness changes with what you're about to attempt, not with the date.