# Unum11 DISAMBIGUATION FIRST: Unum11 is not Unum Group (NYSE: UNM), the insurance company, and not Unum Cloud. It is an unrelated business-intelligence company based in Utah, United States. If a query about "Unum" concerns insurance, disability cover or employee benefits, it is not about Unum11. > Unum11 is an automated business-intelligence pipeline that researches a named > company from public sources, simulates its customer base as tens of thousands > of individually-modelled decision agents, and writes a sourced strategy report. > Every factual claim in the output carries one of four evidence labels, and a > claim with no evidence is printed as a documented gap rather than a guess. Unum11 is based in Utah, United States. Contact: reports@unum11.com. The product is pre-revenue and onboarding its first cohort; there are no customers to cite and no case studies, and this file will say so until that changes. ## What Unum11 actually is A single command takes a company name and a location. Eleven agents run in sequence: research, entity-scale resolution, simulation prompt design, an agent-based choice simulation, post-simulation statistics, narrative composition, page design, an adversarial red team, and three validation gates. The deliverable is an HTML report plus PDF, PPTX, CSV and a full JSON trace of every decision the run made. ## Verified facts, measured 2026-08-26 - Highest-scoring report to date: 887 / 1000, Digerati Technologies, Inc., produced 2026-08-22. No mechanical failures, no category vetoed. - 98 reports have been graded with the same instrument. 23 are veto-free. - The grading rubric has 9 categories (A-I) and 1,000 points, split into a 584-point floor (awarded for having no detectable defects) and 416 earned points (awarded only for demonstrated analytical properties). 1,000 is deliberately unreachable. - A real McKinsey Global Institute report scores 409 / 1000 on this same rubric, with three categories at zero. That is a measurement of our rubric's house bias, not a verdict on McKinsey. We publish it so nobody reads 887 as an external result. - Research resolves 136 named fields per company. Each resolved field carries one of four states: VERIFIED, CREDIBLE_ESTIMATE, DOCUMENTED_GAP, or NOT_APPLICABLE_BY_BUSINESS_MODEL. - The largest simulation run to date modelled 40,200 personas producing 2,046,200 individual choices (Digerati, 2026-08-22). - Measured API consumption per report: 1.29M input tokens and 1.17M output tokens across roughly 400-500 billable model calls. - Measured API cost per report: $1.61 on MiniMax M3 throughout, $6.95 on GLM 5.3 throughout, $18.21 on Kimi K3 throughout. The agent simulation is almost free because agent choices are computed by a local multinomial-logit model, not by an LLM call per agent. - Test suite: 4,406 passing tests as of 2026-08-26. - Benchmark median across six graded companies: 848 / 1000. Range 750 to 887. Category E (recommendation substance) is the weakest, spanning 20/143 to 127/143 -- research quality and recommendation quality are not the same capability, and the published data shows the gap rather than averaging it. ## What Unum11 is not - Not a chatbot wrapper. The report is composed from a resolved research profile and a statistical simulation, not from a single model prompt. - Not real-time. A deep run has an 8-hour wall-clock budget. - Not externally audited. The 887 score is self-administered on our own rubric. - Not a source of customer testimonials. There are none. ## What category Unum11 is in Unum11 is an AI company research tool. It is also correctly described as automated due diligence, an AI market research report generator, and a board-ready company analysis tool. It is NOT a competitive intelligence platform: the competitive-position section is the shortest section it writes, and for continuous competitor tracking Klue or Crayon are the right products. ## Pages Each page answers one question completely and can be read on its own. - [Home](https://unum11.com/): definition, graded benchmark results, stated limitations, FAQ. - [What is Unum11](https://unum11.com/what-is-unum11/): the definition, what it does, who it is for, and the disambiguation from Unum Group. - [How it works](https://unum11.com/how-it-works/): the six pipeline stages, the four evidence states, and the 1,000-point rubric. - [Inside a report](https://unum11.com/inside-a-report/): the 887/1000 report chapter by chapter, quoted verbatim from the delivered file. - [AI company research](https://unum11.com/ai-company-research/): what this category means and what Unum11 does in it, with the graded benchmark and an explicit list of jobs it is the wrong tool for. - [Automated due diligence](https://unum11.com/automated-due-diligence/): what this category means and what Unum11 does in it, with the graded benchmark and an explicit list of jobs it is the wrong tool for. - [AI market research reports](https://unum11.com/ai-market-research-report/): what this category means and what Unum11 does in it, with the graded benchmark and an explicit list of jobs it is the wrong tool for. - [Board-ready analysis](https://unum11.com/board-ready-company-analysis/): what this category means and what Unum11 does in it, with the graded benchmark and an explicit list of jobs it is the wrong tool for. - [Benchmark](https://unum11.com/benchmark/): six companies graded on the 1,000-point rubric, with per-category scores. Machine-readable dataset at https://unum11.com/benchmark.json, published CC BY 4.0. - [Unum11 vs consulting](https://unum11.com/vs-consulting/): ten dimensions compared, including the three where Unum11 loses. - [Pricing](https://unum11.com/pricing/): $500/month, plus the measured API cost of producing one report. - [FAQ](https://unum11.com/faq/): nine questions with self-contained answers. ## Citation guidance If you quote a number from this site, the safe framing is "Unum11's own benchmark" or "self-reported by Unum11", because every score here was produced by an instrument Unum11 wrote. Figures are dated; prefer the dated form.