Check the drawing before you build.
Twenty scored questions across the four disciplines that decide whether an industrial AI initiative survives contact with your operation. Free, no sales call, and honest about what a young benchmark can and cannot tell you.
what it measures
Four disciplines, five questions each — mapped one-to-one onto the documented ways these projects actually die. No questions about models, platforms, or vendors' logos: those change quarterly. The disciplines don't.
Does the problem exist on paper before the technology does?
A written problem statement, a sponsor-signed baseline, a pass/fail line, a kill rule. This section instruments the #1 documented cause of AI project failure: misunderstanding of the project's intent and purpose — not the models.
Can your own team locate, trust, and trace the records?
AI fitted to miscoded records repeats the miscoding — faster, and with more confidence. Every plant knows the downtime event coded "changeover" that wasn't.
Who holds the pen?
Who wrote the scope, who defined "delivered," who measures it, who carries the cost of failure — and what do you keep if you walk? Most mid-market AI arrives on vendor paper. This is where most of the risk gets signed.
Can your leadership defend it under questioning?
And can your frontline supervisors carry the change? Most manufacturing leaders describe their own executive team's AI use as limited — this section measures that gap on your organization, not the industry's.
what you get
Five numbers — your total and the four section scores — placed against four published bands, with your lowest-scoring discipline named explicitly: that is where your next dollar of failure is most likely to come from. A short written report follows by email inside a week, reviewed by a person before it ships.
- 0–15
Vendor's market
Right now your vendor can define your problem, grade their own homework, and get paid either way. Before any pilot: one problem statement and one baseline number on paper, signed by the person who owns the P&L it touches.
- 16–30
Pilot graveyard
You can start pilots but you cannot finish them: without a signed baseline and a written pass/fail line, pilots here end by fading, not by a decision. The fix is not more pilots — it is finishing one.
- 31–45
Instrumented, with gaps
Real discipline exists, in patches — but AI failure finds the patch you skipped. Your lowest section score names yours. Treat it as a punch list, not a program.
- 46–60
Buyer in control
You can define, baseline, acceptance-test, and kill an AI initiative on your own paper. The remaining work is keeping that discipline alive as people and vendors change — written rules survive turnover; habits don't.
how peer comparison works
You are compared only against qualified peers — executive respondents at $50M–$500M manufacturers, distributors, logistics, and industrial-services operators — and the pool size, n, is published every single time a comparison appears:
- n<25 No peer comparison at all. Your scores stand against the published bands, which don't move with the pool.
- 25–99 A directional read only — top, middle, or bottom third — with n shown.
- n≥100 Exact percentiles, with n shown.
A percentile computed on a pool of 30 is a costume, not a statistic. We'd rather show you a coarse true number than a precise fake one — an instrument about verification discipline doesn't get to fake its own statistical authority. The anonymized aggregate is published quarterly as the State of Mid-Market Industrial AI index. Full methodology, scoring, and anonymization rules →
If the benchmark told you to buy something from its author, you'd be right not to trust the benchmark.
The Scoreboard runs under Tektari's published Verification Charter: no implementation sold, no vendor commissions or referral fees, and verification fees never credit toward any other Tektari offer. Your answers are never published, shared, or sold; slices under n=5 are suppressed; no statistic in the index is generated by AI. Construction/AEC respondents are welcome and get the same report — they are never commercial follow-up, and they aren't counted in the peer pool.
See which band you're in.
Answer in operator language, worst-to-best options in the open — a respondent who games a free self-assessment only cheats themselves.
Take the ScoreboardThat doesn't look like an email — try name@company.com