TL;DR: AI Readiness Score is a single-page, dependency-free web app that grades any local business 0–100 on AI/automation readiness through 10 multiple-choice questions. Five categories — Lead Response, Booking & Scheduling, Follow-up & Nurture, Reviews & Reputation, Data & Insights — an animated score reveal with letter grades, per-category bars, top-3 personalized recommendations drawn from the weakest categories, a copyable report, and an embeddable badge. One index.html (333 lines, ~18 KB), no network calls, no analytics, no build step. github.com/Shivay00001/ai-readiness-score.
The problem: businesses can't tell where AI actually pays
Local businesses are drowning in AI hype and starving for diagnosis. Every vendor pitches a tool; almost nobody helps a clinic owner, hotelier, or realtor answer the prior question: where in my operation would automation actually move the needle? The businesses that need AI most — thin front desks, missed calls, no follow-up system — are the least equipped to evaluate a pitch.
A 10-question self-serve grader is a better first conversation than a sales call. It takes two minutes, it asks about the business's actual operations instead of AI buzzwords, and it ends with a prioritized fix-list rather than a pricing page. For an automation studio, it's also the perfect lead magnet: useful even if the visitor never buys anything.
What it does
- 10 multiple-choice questions, one at a time, with a progress bar and a working Back button.
- 5 scored categories: Lead Response, Booking & Scheduling, Follow-up & Nurture, Reviews & Reputation, Data & Insights.
- Animated results: count-up score reveal with a ring, letter grade, per-category animated bars.
- Top-3 personalized recommendations drawn from the lowest-scoring categories — the fix-list, not a generic tip sheet.
- "Copy my report" clipboard summary, embeddable badge snippet, and a Retake button.
- Vertical-aware: enter a business name and pick a vertical (clinic / hotel / real estate / other) before starting.
How it works: the scoring model
The scoring is heuristic, transparent, and entirely client-side:
- Each answer maps to 0–3 points.
- Questions roll up into the five categories; categories are weighted equally — the final score is the mean of the five category percentages, scaled 0–100.
- Letter grades follow fixed bands: 90+ AI-Ready, 70+ Almost There, 50+ Leaking Leads, 30+ Manual Mode, below 30 Urgent.
The grade names are doing real work: "Leaking Leads" and "Manual Mode" describe the business's condition, not its technology — which is exactly the framing that makes a non-technical owner keep reading. The top-3 recommendations are generated from the lowest-scoring categories, so two businesses with the same total score can get completely different fix-lists. That personalization is what separates this from a static quiz.
The category design
The five categories aren't random — they map to the highest-ROI automation surface for local businesses:
| Category | What it probes | Why it matters |
|---|---|---|
| Lead Response | How fast new inquiries get answered | Speed-to-lead decides deals; most SMBs are slow |
| Booking & Scheduling | Whether booking happens without human effort | Calendar friction is the #1 lost-revenue leak |
| Follow-up & Nurture | Whether leads get systematic follow-up | Most businesses follow up zero times; money lives in touches 2–5 |
| Reviews & Reputation | Whether reviews are requested and managed | Review velocity drives local search ranking |
| Data & Insights | Whether the business measures anything | You can't automate what you don't measure |
Implementation notes
One HTML file, inline CSS, vanilla JavaScript — no frameworks, no CDNs, no build. The script's function list reads like a clean little state machine: show and renderQ drive the question flow, gradeOf and gradeColor map scores to grades, finish and renderResults build the animated results screen, reportText assembles the clipboard summary, copyText handles clipboard with a fallback. It works over file:// because there is genuinely nothing to fetch — no fonts, no libraries, no API keys.
The privacy posture is a feature, not an accident: no data leaves the browser. No network calls, no analytics, no tracking pixels. For a tool aimed at business owners wary of handing data to vendors, "your answers never leave this page" is a meaningful trust signal — and it's verifiable by reading the file.
How to run it
# Option 1: just double-click index.html
# Option 2: serve it
cd ai-readiness-score
python3 -m http.server 8000
# open http://localhost:8000Demo steps: enter a business name, pick a vertical (clinic / hotel / real estate / other), start the assessment, answer the 10 questions (Back works), watch the score count up, review the category bars and top-3 fixes, try Copy my report and Copy badge code, then hit Retake and compare.
Honest status and roadmap
Status: working prototype. The full flow runs — questions, scoring, animated results, recommendations, report copy, badge embed, retake. What it is not: a scientific instrument. The score is a heuristic self-assessment, and the README says so. Ten questions can't audit an operation; what they can do is surface the obvious leaks and prioritize them, which is the job.
Roadmap (sensible next steps):
- Vertical-specific question sets — a clinic's readiness surface differs from a hotel's.
- Industry benchmarks: "businesses like yours typically score X".
- Email-gated full report — turning the grader into a proper lead magnet with a detailed PDF breakdown.
- CRM/webhook handoff so scores flow into a sales pipeline.
- Multi-language support for non-English markets.
Who should look at this repo
- Agencies and automation studios — this is a ready-made lead magnet; fork it, brand it, embed the badge on client sites.
- Frontend developers — a clean example of a dependency-free interactive quiz with animated results in a single file.
- Business owners — take the 2-minute assessment before your next conversation with any AI vendor.
A worked example
Imagine a dental clinic taking the assessment. It answers honestly: new inquiries go to voicemail after hours (Lead Response: weak), bookings require a phone call during business hours (Booking: weak), nobody follows up on unsent treatment plans (Follow-up: weak), the front desk sometimes asks happy patients for Google reviews (Reviews: middling), and the owner has no idea how many calls were missed last month (Data: weak).
The result: a score in the 30s — Manual Mode — with top-3 recommendations pointing at after-hours lead capture, online booking, and systematic follow-up. That's a prioritized automation roadmap in two minutes, and every recommendation maps to a service an automation studio actually sells. Now imagine the same clinic retaking the quiz six months after implementing those three fixes: the score jumps into the 70s, the badge on their site says "Almost There," and the studio has a before/after story. The grader isn't just diagnostic — it's a progress tracker.
Design decisions worth noting
- Equal category weights. The final score is the mean of the five category percentages. That's a defensible default for a general-purpose grader — it avoids pretending the author knows whether booking matters more than reviews for your business, and it keeps the math explainable in one sentence.
- Vertical picker up front. Asking for the vertical before question one lets future versions swap in vertical-specific question sets without changing the flow — the architecture anticipates the roadmap.
- Everything client-side. No backend means zero hosting cost, zero latency, and the privacy story above. The tradeoff — no lead capture — is exactly what the roadmap's "email-gated full report" item addresses, and it's the right order: prove value first, ask for the email second.
- Badge snippet. A pasteable HTML badge turns every graded business into distribution. It's the kind of growth mechanic that costs ten lines of code.
A final practical note: the grader is most useful when taken twice — once honestly as things are today, and once as a target-setting exercise ("what would we need to change to hit 70?"). The gap between the two is a project plan. And because the whole thing is a single file with no tracking, there's nothing stopping a business owner from running it for every location they operate and comparing scores side by side.
Frequently asked questions
Does any of my data leave the browser?
No. There are no network calls, no analytics, and no external CDNs. Everything runs locally in the page, and it works over file://.
Is the score scientific?
No — it is a heuristic self-assessment. Ten questions can't audit an operation, but they reliably surface the highest-leverage gaps and prioritize them, which is the point.
Can I use this as a lead magnet for my agency?
The embeddable badge snippet is designed for exactly that. A gated full-report version is on the roadmap.
Which businesses is it for?
Any local business. The vertical picker (clinic / hotel / real estate / other) tunes the framing of the assessment.
The bottom line
AI Readiness Score understands something most AI tools don't: the sale starts with diagnosis, not technology. In one dependency-free file it delivers a genuinely useful two-minute assessment with personalized, prioritized fixes — and it does it without tracking the user at all. Fork it, brand it, and put it on your agency site; or just take the quiz and fix your lowest category this week.