Free Resource · Interactive · No Email Required

Machine Customer
Readiness Index
Scorecard

Twenty questions. Ten minutes. A live score that tells you whether an AI assistant acting on behalf of your buyer can find you, verify you, and buy from you — and exactly where it cannot.

This scorecard is the companion to The Agent-Ready Revenue Architecture. The article makes the case; this tool gives you the score. Run it live below, or download the workshop PDF and run it with your leadership team. Either way, the result is built to be read aloud in a board meeting without translation.

Executive Summary

AI assistants are becoming a permanent part of B2B buying. They research vendors, assemble shortlists, verify claims, and — in a growing set of categories — complete purchases. And they evaluate you differently than a human does: they reward what they can retrieve, parse, and verify, and they silently drop what they cannot.

The Machine Customer Readiness Index (MCRI) measures how your company performs against that evaluation. Twenty questions, five dimensions:

Your score places you on a four-level ladder from Invisible to Agent Ready. Most enterprises that score themselves honestly land at Legible: good enough to be recommended by AI, not yet good enough to be bought by it. The gap between your level and the next one is the most useful to-do list this scorecard can give you.

The Short Answer
What the MCRI tells you

The MCRI is a 20-question self-assessment that scores your readiness for AI-assisted buying. Score each question 1–5 for a total out of 100, and read your level:

  • Invisible (20–44) — AI buyers cannot find or read you
  • Legible (45–64) — readable, but not buyable
  • Transactable (65–84) — an AI buyer can purchase inside your guardrails
  • Agent Ready (85–100) — built for the machine customer by design

Your weakest dimension — not your total — is your first wall, and the place to start. The assessment below does the arithmetic for you.

How to Use the Scorecard

Method
Four steps, one working session
Ten minutes alone, or an hour with your leadership team.
Step 1 — Get the right people in the roomBring marketing, RevOps or sales operations, product, and someone who can speak for legal or IT. Scored from inside one function, the number flatters you — and tells you nothing.
Step 2 — Score honestlyUse the scoring guide below, and score what exists today — not what is on the roadmap. The number only helps you if it is true.Rule of evidence: if you cannot point to the page, document, or system that proves the answer, subtract a point.
Step 3 — Watch your result buildAs you answer, the page totals each dimension out of 20 and the overall index out of 100. The moment the twentieth answer lands, you get your level, your strongest capability, and your first wall. Nothing is stored or sent anywhere.
Step 4 — Start at the weakest pointYour lowest-scoring dimension is your first wall — the exact place an automated evaluation of you stops. The guidance below the assessment tells you where the next 90 days should go.
Scoring guide
ScoreWhat it means
1Not true today
2Rarely true
3Partially true
4Mostly true
5Fully true — and we can prove it

The Assessment

Five dimensions, four questions each. Score every question from 1 (not true today) to 5 (fully true — and provable).

Dimension 01 · Questions 1–4
Machine Legibility
Can AI understand your company?
Q1 — Can AI understand your website?If AI read only the words and structure on your most important pages, would it clearly understand what you sell, who it is for, what it costs, and how someone buys it? Design does not matter here. The facts do.
Not true todayFully true
Q2 — Can AI read your product pages correctly?If ChatGPT or another AI tool reviewed your product and pricing pages today, would it understand them without guessing? Score higher if your pages use clear structure, current schema, and plain language.
Not true todayFully true
Q3 — Can AI access your product details?Can an AI system read your full product details without filling out a form, downloading a gated PDF, or asking sales? If the details are hidden, AI cannot use them.
Not true todayFully true
Q4 — Do you describe your products consistently?Do your pages use the same product names, editions, versions, and pricing units everywhere? Inconsistent naming makes AI treat one offer like several different products.
Not true todayFully true
Dimension 02 · Questions 5–8
Verifiable Substance
Can AI verify your claims?
Q5 — Can AI verify your biggest claims?If AI checked your three biggest claims against reviews, directories, documentation, and analyst coverage, would those claims hold up? Claims are stronger when someone other than you can confirm them.
Not true todayFully true
Q6 — Is your proof easy to access?Can AI reach the proof that matters, such as security information, compliance certifications, technical documentation, and integration guides, without a login or form? Proof behind a gate is hard for AI to use.
Not true todayFully true
Q7 — Do outside sources describe you correctly?Do review sites, directories, analysts, and comparison pages describe your company the same way you do? If outside sources misclassify you, AI may recommend you for the wrong thing or not at all.
Not true todayFully true
Q8 — Can your numbers be checked?Can someone verify the numbers you publish, such as ROI, growth, performance, or savings claims, against a source, date, or method? Unsupported numbers are easy for AI to ignore.
Not true todayFully true
Dimension 03 · Questions 9–12
Transaction Readiness
Can AI actually buy from you?
Q9 — Can a customer buy without talking to sales?Can a qualified buyer start or complete a purchase without talking to a person? Score the real buying path, not the ideal one. Every required meeting creates friction for an automated buyer.
Not true todayFully true
Q10 — Can AI calculate your price?Can AI work out what your product costs from published pricing, a calculator, or clear pricing inputs? To an automated buyer, “Contact us” is not a price.
Not true todayFully true
Q11 — Do you know where the buying path breaks?Have you actually tested how far an automated buyer could get through your most common purchase? Score 5 only if you know the exact step where the process stops.
Not true todayFully true
Q12 — Can your systems support the purchase?If software tried to request a quote, accept standard terms, or create an account, could your systems respond? Or would the process immediately route to a person?
Not true todayFully true
Dimension 04 · Questions 13–16
Governance
Can your business safely support AI-driven buying?
Q13 — Are the buying rules written down?What can an automated deal do without human approval, and is that written down? Include discount limits, standard terms, deal-size thresholds, and escalation rules. Score low if the rules only live in someone’s head.
Not true todayFully true
Q14 — Who owns exceptions?If an automated purchase went wrong tomorrow, would everyone know who owns the response and when a human must step in? Fast buying requires clear accountability.
Not true todayFully true
Q15 — Can you stop a bad automated purchase?If AI made a purchase it should not have made, how quickly would your business notice and respond? Weekly reviews are too slow for transactions that happen in minutes.
Not true todayFully true
Q16 — Would Legal be comfortable with this?If AI completed a purchase today, would Legal and Compliance be comfortable with how the contract was accepted, recorded, and audited? If the answer is unclear, the process is not ready.
Not true todayFully true
Dimension 05 · Questions 17–20
Discoverability
Will AI recommend you?
Q17 — Does AI recommend you?When buyers ask AI tools who they should consider, does your company appear as a recommendation, not just a mention? Score from actual testing, not instinct.
Not true todayFully true
Q18 — Are you present where AI looks?Are you accurate and up to date in the places AI pulls from, such as review platforms, comparison sites, directories, documentation, and public profiles? If you are missing there, you may be missing from the deal.
Not true todayFully true
Q19 — Can AI reach your best content?Can AI systems reach the content that helps you win, or do crawler blocks, forms, and locked PDFs get in the way? A locked door often reads like an empty room.
Not true todayFully true
Q20 — Does AI describe you correctly?When AI describes your company, does it get your category, customer, and differentiator right? Being recommended for the wrong reason can still lose the deal.
Not true todayFully true
Your Result

0 of 20 questions answered. Your executive briefing — overall readiness, level, strongest and weakest dimension, first wall, and priorities — will appear here the moment the twentieth answer lands. Nothing you enter is stored or sent anywhere.

The Four Readiness Levels

Your total score places you on the MCRI ladder. The levels are deliberately blunt: each one describes what an AI assistant acting on behalf of your buyer can and cannot do with you today.

The MCRI Ladder
From absent to advantaged
Level 1 — Invisible · 20–44AI buyers cannot find or read you. You are absent from machine-built comparisons — and you do not know it, because no one calls to tell you that you were never considered. Companies here are losing AI-assisted deals today, not in 2028.
Level 2 — Legible · 45–64AI can read you but cannot act. You make the shortlist, your claims mostly hold up — and then the evaluation hits a wall: no computable price, no self-serve path, no way to transact. This is where most enterprises honestly sit today: shortlisted, not bought.
Level 3 — Transactable · 65–84An AI buyer can complete a real purchase inside your guardrails: computable pricing, a genuine self-serve path, written rules for what may happen without a human. You are now competing on fit and terms — not on whether a machine can read you.
Level 4 — Agent Ready · 85–100You treat the AI buyer as a first-class customer by design: readable, verifiable, buyable, governed, and easy to find. Readiness compounds here — every improvement makes you easier to select the next time an agent runs a pass, in a loop your competitors have to build from scratch.

Executive Guidance

How the score works

Each question scores 1–5. Each dimension subtotals out of 20, and the five subtotals sum to an index out of 100.

The assessment above applies one honesty rule automatically: your level is capped by your weakest dimension. If any dimension scores 8 or below, your level drops one rung from what your total suggests. A company with brilliant legibility and no governance is not Transactable; it is an incident that has not happened yet.

The Executive Workshop Scorecard includes the same arithmetic as a printable worksheet for team sessions.

MCRI scoring bands
LevelTotal scoreWhat it means operationally
Invisible20–44Absent from machine-built comparisons; losing by omission today.
Legible45–64Recommended but not buyable; evaluations stop at the transaction wall.
Transactable65–84An AI buyer can purchase inside your guardrails; you compete on fit and terms.
Agent Ready85–100Built for the machine customer by design; readiness compounds.

How to read your result

Read the dimensions before the total. The total tells the board where you are; the dimension profile tells the operating team what to do. A 62 built on strong Legibility and weak Governance is a different company than a 62 built the other way around — and they should spend the next quarter differently.

Then find your first wall: the lowest-scoring question in your lowest-scoring dimension. That single question is usually worth more than the other nineteen combined, because it marks the exact point where the machine’s evaluation of your company ends.

Treat the score as a quarterly baseline, not a one-time verdict. It is built to sit beside your Share of Model report in the same board deck: Share of Model measures whether AI recommends you; the MCRI measures whether AI could buy from you. Read together, they describe your position on both sides of the Delegation Threshold.

Immediate priorities by level

Where to spend the next 90 days
Your levelFirst priorityWhat to defer
InvisibleMachine Legibility, immediately: fix schema, open up your specifications, and run the stripped-page test on your top revenue pages. Nothing else matters while AI cannot read you.Everything to do with automated transactions and negotiation. You are not yet ready to be read, let alone bought.
LegibleTransaction Readiness: publish your pricing or make it calculable, walk an automated buyer through your most common purchase, and remove one mandatory human step.Selling through your own AI agents. Close the gap between shortlisted and buyable first — it is the least crowded opportunity in your category.
TransactableGovernance, in lockstep with growth: write the rules for what a deal may do without a human, name the person who owns exceptions, and make your controls as fast as your transactions.Expanding the automated buying channel before the guardrails are tested. An unguarded channel is a liability, not a lead.
Agent ReadyMeasurement and compounding: track Share of Transaction, protect how AI describes you, and keep your third-party proof fresh.Complacency. The ladder is re-scored every quarter, and this level is defended, not owned.

Board discussion questions

Governance Agenda
Five questions worth twenty minutes of board time
01What is our MCRI level today, and which dimension is our first wall? If the answer is a guess rather than a score, that is the finding.
02What share of buying in our category is already AI-assisted, and what evidence are we using to answer that question?
03If a buyer’s agent attempted our most common transaction tomorrow, where exactly would it stop — and how many of our competitors would it get further with?
04Who owns agent-readiness across marketing, sales, product, and legal — one name, not four — and what did that owner ship last quarter?
05What would a competitor at Agent Ready take from us first, and how long would it take us to notice?

Recommended next actions

This week Baseline
  • Run the full scorecard with all four functions in the room
  • Record the total, the dimension profile, and the first wall
  • Name the single accountable owner
Days 1–30 Legibility & Substance
  • Fix schema and open up specifications on top revenue pages
  • Get your three biggest claims verified in sources you do not control
  • Correct the worst third-party description of your company
Days 31–60 Transaction Readiness
  • Walk an automated buyer through your most common purchase
  • Publish your pricing, or make it calculable
  • Remove one mandatory human step from the buying path
Days 61–90 Governance & Re-score
  • Write the rules for machine-speed deals and their escalation points
  • Review automated purchasing with legal
  • Re-run the MCRI and report movement to the board

Where This Sits in the ERM Framework Ecosystem

The MCRI is the diagnostic layer of The Agent-Ready Revenue Architecture, and it inherits from the frameworks that precede it. The AI Visibility Architecture earns the recommendation that the Agent-Ready Stack converts into an order, and the Discoverability dimension is the bridge between the two. The Recommendation Ladder and Share of Model measure how you perform below the Delegation Threshold; the MCRI measures whether you are ready above it.

On the buying-group side, Buying Group Mapping defines the five human archetypes the agent now joins as the Sixth Seat.

And if the scorecard shows your team knows what to fix but cannot get it prioritized, that is not an AI problem. That is the Marketing Execution Gap — and the Enterprise Marketing Operating System is how readiness work survives contact with the quarter.

Run the scorecard, keep the number, and re-score quarterly. Then continue into the rest of the Agent-Ready arc: the architecture that explains the score, and the frameworks that move it. For help running a facilitated baseline, work with E.R.M. Advisory directly.

← Back to the Agent-Ready Revenue Architecture

Enjoyed this? Get The Operator.

Revenue leadership and operator frameworks. Three times a week. No fluff.

Subscribe Free →