PULSE Method v1.0
PULSE (Position, Endorsement, Sentiment & Entity Performance) is a client-readable, evidence-based methodology for scoring corporate visibility across commercial AI engines.
What is a PULSE Score?
A PULSE Score summarises how visible and how strongly recommended a company is across AI platforms, for a documented set of buyer questions, competitors, location and language. Each individual answer is scored as Position × Recommendation × Sentiment; those results roll up into Platform Scores, a PULSE Benchmark Score and a PULSE Market Score. Every published score states its method version, testing period, platforms, prompt count, repeat runs and reliability rating.
PULSE Method v1.0 Specification
Evaluates presence and prominence rank order (e.g. 1st choice vs secondary mention).
Measures endorsement strength (Strong Recommendation, Neutral, Conditional, or Negative).
Assesses tonal sentiment, factual accuracy, and absence of hallucinated complaints.
The PULSE Score reflects observed AI recommendation performance within a controlled prompt set and testing period. It does not measure every possible answer or guarantee future visibility.
Recommendation Classes
Being mentioned is not the same as being recommended. Each appearance is classified:
- Strong Recommendation: The AI engine explicitly names the entity as a top-recommended solution.
- Neutral Mention: The entity is listed among alternatives without explicit commercial endorsement.
- Conditional Endorsement: Recommended only for niche sub-contexts or under specific caveats.
- Negative / Risk Mention: Listed alongside complaints, regulatory warnings, or hallucinated defects.
Reliability Ratings
- High Reliability: Evaluated across 20+ prompts, 5+ repeat runs per prompt, and 4 platforms over 7+ days.
- Medium Reliability: Evaluated across 10–19 prompts with 3+ repeat runs.
- Indicative Reliability: Initial snapshot evaluation. Never presented as definitive.
Why results vary, and how we control for it
- Between runs: the same prompt can produce different answers, so every prompt is repeated and scored as a pattern.
- Between platforms: each assistant retrieves and weighs sources differently, so each gets its own Platform Score.
- By location and language: answers change with the buyer’s market, so each test fixes both and states them.
- Over time: models and their sources are updated, so scores carry a testing period and are re-measured rather than assumed.
Illustrative Scoring Matrix
Illustrative Example — Not Client Data| Prompt Scope | Position | Endorsement | Sentiment | Result |
|---|---|---|---|---|
| "Best MSP in London" | 1st Choice (1.0) | Strong (1.0) | Positive (1.0) | 1.00 |
| "Compare London MSP vendors" | 2nd Choice (0.8) | Neutral (0.6) | Positive (1.0) | 0.48 |