The LSAI method: how we assess specialists
The Local Search & AI Index (LSAI) totals 100 points. 40 points come from people: clients we call ourselves and verified reviews. 25 points come from proven results. The AI measurement follows Rand Fishkin’s approach: visibility as a percentage over many repetitions, no positions.
The LSAI replaces the Local SEO Authority Index of bestelocalseospecialist.nl, the founder’s previous site. That index weighted client satisfaction at 40, proven case studies at 25, follow-up client calls at 20 and knowledge sharing and awards at 15. The human factor remains 40 points, but is now split into two measurable parts.
The six pillars
| Pillar | Points | What counts | Source |
|---|---|---|---|
| Follow-up client calls by the editorial team | 20 | Would choose again 6, results according to the client 5, communication and reporting 4, promise versus reality 3, own implementation 2 | Three clients, chosen by lot from at least eight |
| Client opinion in reviews | 20 | Weighted average 12, recency (24 months) 4, at least two platforms 2, response to criticism 2. Fewer than five reviews: maximum 10 | Google, Trustpilot, Trustoo, Clutch |
| Proven client results | 25 | The best three case studies on the evidence ladder from 0 to 5; score = sum divided by 15, times 25 | Case studies, raw measurement data, follow-up call |
| Implementation and measurement honesty | 15 | Six workstreams of 1.5 points each, plus 6 points for honest measurement: percentage instead of position, number of questions and repetitions stated, per AI tool separately, link to branded searches or direct traffic | Quote, sample report |
| AI visibility of the agency’s own name | 10 | Visibility as a percentage 8, used as a source 2 | Own measurement per edition |
| Knowledge sharing, transparency and foundations | 10 | Own tests with a method 3, trade press, speaking and books 2, prices and terms 2, KvK (Dutch Chamber of Commerce), name and profiles 2, fast and accessible website 1 | Own website, KvK, public profiles |
The LSAI totals 100 points. Awards only count if an independent jury grants them.
The evidence ladder for client case studies
Every case study is given a level from 0 to 5. A claim without data is level 0.
- Case study with figures, but without a reference date or measurement design.
- Before and after measurement with dates.
- AI measurement according to our rules: a fixed set of at least 20 questions, each question asked at least 60 times per AI tool, visibility as a percentage, raw data available.
- Plus a business result: more branded searches or branded impressions in Search Console, more direct traffic, more phone calls or enquiries.
- Plus confirmed by the client in the follow-up call, or measured again by us.
Penalty points and exclusion
- Untrue or unverifiable claim on the provider’s own website: minus 5 points per claim, up to minus 15.
- Selling “number 1 in ChatGPT”, “guaranteed mentions” or an AI position as a result: minus 5.
- Publishing its own best agencies list without disclosing its own interest: minus 3.
- Fake reviews, steered reference clients or logos of non-clients: excluded for 12 months.
Statistics and fairness
- Margin of error: a 95% margin for each final score. If the margins overlap, specialists share a place.
- Data label: A (verified), B (publicly complete) or C (limited data).
- Sensitivity check: in every edition we also show the order under the old weighting, so you can see what the new weighting changes.
- Deliberately not counted: turnover, team size, website traffic and paid awards. Those reward size, not quality.
- Agencies and independents separately: an independent is assessed on having proper cover arranged instead of on team size.
Further reading on the method
Read how we call clients chosen by lot, how we measure AI visibility without positions and which rules we apply to interests and review.
Frequently asked questions
What does LSAI stand for?
Local Search & AI Index. It is the successor to the Local SEO Authority Index of bestelocalseospecialist.nl, extended with GEO and an honest AI measurement. See also: AI visibility tally sheet.
Why does the human factor count for 40 points?
Because clients know best whether a specialist does what they promise. We call clients ourselves and include verified reviews. See also: Position in ChatGPT: why a fixed number 1 spot does not exist.
What is the evidence ladder?
A scale from 0 to 5 per client case study. A claim without data scores 0; a case with a fixed measurement design, a business result and confirmation by the client scores 5. See also: Best GEO agency: how to recognise a fair ranking.
Why no points for turnover or team size?
Because those reward size, not quality. A large agency is not automatically a good agency. See also: Best GEO agency.
How do you prevent an agency from only putting forward satisfied clients?
The agency provides at least eight clients and we choose three of them by lot. If a client cannot be reached, we draw again.
What are penalty points given for?
For untrue claims on the agency’s own website, for selling a position in AI answers and for publishing its own best agencies list without disclosing its own interest. Fake reviews lead to exclusion.
What is a data label?
The label A, B or C shows how complete the data is: verified, publicly complete or limited. This shows whether a score rests on follow-up client calls or only on public sources.
Why do specialists sometimes share a place?
Because every measurement has a margin of error. If the margins overlap, the difference cannot be established firmly and they share the place.
Does the method change per edition?
The method is fixed per edition and published in advance. Any change is listed with a date and reason in the corrections log.
Who checks the method?
LocalSEOLab.nl and Stichting LODAK review the method and scoring rules. Both have ties to the founder; this is set out on the independence page.