Evidence trace
Each layer has a clear job
The trace separates the verified source, the vocational construct it informs, the custom WhatBusiness.ai rule, the resulting match, and the buyer test that provides demand evidence.
Source evidence
Verified published source
Research or official documentation checked and mapped to a specific claim.
Example: the official O*NET Mini Interest Profiler development report documents a 30-item RIASEC instrument for career exploration.
This source informs the vocational-interest layer used by WhatBusiness.ai.
Source evidence
Vocational-interest construct
Six continuous RIASEC dimensions describe patterns of vocational interest.
Vocational interests are represented across Realistic, Investigative, Artistic, Social, Enterprising, and Conventional dimensions.
WhatBusiness.ai retains continuous scores and combinations of leading dimensions.
Interpretation
Published product rule
A custom, inspectable WhatBusiness.ai matching decision.
Interest scores contribute 30% of the base fit score. Skills, resources, goals, and operating preferences contribute the remaining 70%.
These custom weights are published in scoring version 1.2.0. Observed product outcomes guide future calibration.
Internal check
Explainable match output
Three ordered roles with fit evidence, requirements, and conflicts.
The system orders a best overall match, a fastest realistic path to revenue, and a higher-upside alternative.
Each result explains the offer, buyer, existing advantages, missing inputs, smallest credible version, and reasons it may not suit the user.
Interpretation
Buyer test
A real-world test supplies direct evidence about buyer demand.
The match establishes a direction with documented personal fit and feasibility.
The next step is a small buyer test that checks whether real customers will engage with or pay for the proposed offer.
Base fit
Five inputs form the base score
The matcher scores every opportunity in the curated library against the same five dimensions. RIASEC interests remain continuous combinations rather than a single type. Skills are reported as broad areas plus an overall experience level. Operating preferences capture choices such as online or local and B2B or B2C. Personality scoring is outside the assessment scope.
Interests and activity fit
30%
Skills and experience
25%
Resources and access
20%
Goals and urgency
10%
Operating preferences
15%
Published rule: these five weights are custom WhatBusiness.ai product rules in scoring version 1.2.0. Behavioral outcomes—including completion, result engagement, idea selection, rejection, and paid demand—inform future calibration.
Conflict checks
Constraints do not disappear inside a score
Budget, weekly time, revenue timing, physical work, travel, cold selling, regulation, hiring, and home-only requirements are checked separately from base fit. An opportunity that crosses a stated boundary receives a ranking penalty and a plain-language conflict.
Current scoring rule
Adjusted rank = base fit − conflict penalties
- Budget overrun
- 0.30 × proportional overrun
- Weekly-time overrun
- 0.24 × proportional overrun
- Revenue-deadline overrun
- 0.16 × proportional overrun
- Each categorical conflict
- 0.32
These coefficients are published implementation rules in scoring version 1.2.0. Every detected conflict remains visible when a near match fills an open result role.
Ordered output
How we fill the three result roles
We fill these roles in order. Each role uses an unused exact match while one is available. After exact matches are exhausted, any remaining role may use a clearly labelled near match that discloses every conflict and states that it is not feasible as-is. Revenue timing and upside are estimates stored in our opportunity library, not predictions.
- Result role
Best overall match
Selects the unused candidate with the highest adjusted fit.
- Result role
Fastest realistic path to revenue
Prefers a different business family and the lowest stored estimate for time to first revenue among unused candidates.
- Result role
Higher-upside alternative
Prefers a different business family and the strongest combination of stored upside rating and adjusted fit among unused candidates.
We prefer different business families where suitable unused candidates are available, but diversity is not guaranteed.
Structured selection, bounded explanation
Curated opportunities come first
Candidate directions come from a curated, structured opportunity library. Deterministic code applies the scores, conflict checks, and result roles. Generative AI then explains, personalizes, niches, and creates bounded variants using the selected opportunity and supplied profile data.
Deterministic layer
Controls what can be selected and why it ranks.
- Inputs
- Structured assessment answers
- Candidates
- Curated opportunity library
- Output
- Scores, conflicts, ordered roles
Usually stronger when: the product must be inspectable, testable, and repeatable.
Generative layer
Turns selected records into clearer personal guidance.
- Inputs
- Selected record plus profile facts
- Boundary
- Cannot replace or add a candidate
- Output
- Explanation and bounded variants
Usually stronger when: specific language makes a structured result easier to act on.
Product scope
What each layer establishes
RIASEC provides the vocational-interest structure. WhatBusiness.ai publishes the custom weights, mappings, conflict policy, opportunity library, and three-role selection logic used to turn a profile into business directions.
Sources are verified and mapped to the specific vocational constructs and product claims they support.
Matching weights, opportunity mappings, conflict rules, and result roles are custom published WhatBusiness.ai product rules.
The assessment scores interests, reported skill areas and experience, resources, goals, operating preferences, and constraints. Personality scoring is outside its scope.
Each match explains personal fit and feasibility. A buyer test with real customers provides direct evidence about demand.
Attribution
O*NET interest activities
The interest section uses the 30 O*NET Mini Interest Profiler activity statements to produce six continuous RIASEC interest scores. Those scores are then used as one input to a separate WhatBusiness.ai business-opportunity ranking model. WhatBusiness.ai changes the downstream purpose and output: it ranks a curated business-opportunity library rather than returning occupations for career exploration.
Required attribution and modification notice
This product includes information from the O*NET Career Exploration Tools by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). Used under the O*NET Tools Developer License. O*NET® is a trademark of USDOL/ETA. WhatBusiness.ai has modified all or some of this information. USDOL/ETA has not approved, endorsed, or tested these modifications.
WhatBusiness.ai has modified the purpose of the O*NET® Career Exploration Tools. The U.S. Department of Labor, Employment and Training Administration (USDOL/ETA) has not approved, endorsed, or tested these modifications. As such, USDOL/ETA will not be liable to any third party or end-user for any damages arising out of or from the use or misuse of the modified O*NET Career Exploration Tools or any products incorporating or containing the modified O*NET Career Exploration Tools.
Read modifications and disclaimers