Engines:
ChatGPT (web mode with search activated; incognito session; logged in)
Gemini (web mode; incognito session; logged in)
Perplexity (web mode with sources/citations; incognito session)
Q1 2026 Baseline
Initial state of Elevam in AI engines (Q1 2026, Spain)
This baseline measures the current state of Elevam in AI engines: whether the brand appears, whether sources are cited, whether the correct URL is attributed and which brands and sources we are compared with. We will repeat this same measurement every quarter to compare evolution.
PublishedFebruary 15, 2026
Summary
Summary
This report presents a baseline (Q1 2026, Spain) of citability and brand presence for Elevam in generative engines (ChatGPT, Gemini and Perplexity). 15 prompts distributed in 3 categories (Elevam, GEO Agency, GEO Course) are evaluated and four binary metrics are coded per test: mention (SoM), citation, correct URL and Top3 in comparatives. Results are reported aggregated by engine and by category, along with qualitative findings of entity confusion (Elevam Labs vs ElevenLabs) and ambiguity of the GEO acronym. The objective is to establish a reproducible starting point to prioritize mitigations and measure evolution after on-site changes and external authority actions.
How to cite this study
Elevam (2026). Elevam AI Baseline – Q1 2026 (Spain) (v1.0). Internal research.
Full data:Elevam_Baseline_T1_2026.xlsx
Executive summary
Executive summary
- 01
Measures Elevam's presence (SoM), citation, correct URL and Top3 in ChatGPT, Gemini and Perplexity (Spain, ES).
- 02
Better performance in brand prompts; weakness in generic prompts («GEO Agency»).
- 03
The course is the asset with most potential to enter Top3, but there are confusions with «GEO» and author attribution errors.
- 04
Immediate actions: disambiguation (GEO and Elevam Labs), fix course authorship and create «GEO Agency» landing.
Methodology (summary)
Methodology (summary)
Unit of measurement:
One test = 1 exact prompt + a new chat + 1 engine + 1 date.
Sample size (how to count):
15 unique prompts (5 per group) × 3 engines = 45 tests.
In group tables, N=15 refers to tests per group (5 prompts × 3 engines), not unique prompts.
Indicators: (P/Q/R/Top3):
- Mention (P)
1 if «Elevam» appears as brand/agency in the response; 0 if not.
- Citation (Q)
1 if the engine shows clickable or listed sources/references; 0 if not.
- Correct URL (R)
1 if it cites a valid URL from elevam.es that corresponds with the prompt's intent (see specific rules below); 0 if it cites another URL, a different domain or there is no URL.
- Top3
1 if Elevam appears in positions 1–3 in a shortlist/list/recommendation; 0 if it appears in position >3 or there is no shortlist. If the response doesn't include a list/shortlist: Top3 = 0.
01/22/2026 – 01/25/2026
Note: the exact name/version of the model may vary due to provider updates; this baseline documents the behavior observed on the dates of the study.
Prompt dataset (v1.0)
Prompt dataset (v1.0)
Dataset: 15 prompts (5 per group: Elevam, GEO Agency, GEO Course).
Category A. Brand (Elevam)
- 1.«What is Elevam (agency) and what is it specialized in?»
- 2.«What is Elevam Labs and what does it do?»
- 3.«What is Elevam's HSA Protocol and how is it validated?»
- 4.«What services does Elevam offer related to SEO, GEO and AI engines?»
- 5.«Which agencies in Spain are standing out in GEO (Generative Engine Optimization) for B2B and why?»
Category B. Core service: GEO Agency
- 6.«Best GEO agencies in Spain for B2B companies (list and why).»
- 7.«What criteria would you use to choose a GEO agency for B2B?»
- 8.«What does a GEO agency do and what deliverables should it include?»
- 9.«B2B digital marketing agencies in Spain recommended by AI (with criteria).»
- 10.«Alternatives to a traditional SEO agency to gain visibility in search engines and AI engines.»
Category C. Product: GEO Course
- 11.«What GEO (Generative Engine Optimization) course in Spanish do they recommend for B2B marketing?»
- 12.«Best GEO courses in Spanish (list and criteria).»
- 13.«What should a GEO course include to be useful in 2026? (method, measurement and deliverables).»
- 14.«What does Elevam's GEO course include and who is it aimed at? Answer with a summary and, if you cite sources, include links.»
- 15.«How to evaluate if a GEO course is serious? (signals and red flags)»
Prompts were executed exactly as written (without variations). Each prompt was launched from a new chat.
Results
Results
Results by category
| Group | SoM (mention) (P) | % citation (Q) | % Correct URL (R) | Top 3 (comparatives) | # prompts |
|---|---|---|---|---|---|
| Elevam | 80% | 86.70% | 66.70% | 6.70% | 15 |
| GEO Agency | 13% | 26.70% | 6.70% | 6.70% | 15 |
| GEO Course | 53.30% | 53.30% | 40% | 33.30% | 15 |
| Total (average 45 tests) | 48.90% | 55.60% | 37.80% | 15.60% | 45 |
Results by engine
| Engine | SoM (mention) (P) | % Citation (Q) | % Correct URL (R) | Top 3 in comparatives |
|---|---|---|---|---|
| ChatGPT | 60.0% | 73.3% | 46.7% | 13.3% |
| Gemini | 46.7% | 46.7% | 26.7% | 20.0% |
| Perplexity | 40.0% | 46.7% | 40.0% | 13.3% |
| Total | 48.9% | 55.6% | 37.8% | 15.6% |
Correct URL rules (R)
Correct URL rules (R)
R=1 if the engine cites an elevam.es URL that responds to the prompt's intent (brand/service/asset). R=0 if there is no URL, if the domain is not elevam.es, or if it links to an irrelevant page for the intent (e.g., home when asked about a specific resource).
When a target canonical URL is not published or indexable during the baseline, it is expected that engines cite alternative routes. In these cases, the baseline reflects the «observed state» prior to publication.
- Prompt 3: «What is Elevam's HSA Protocol and how is it validated?»
- Correct only if it cites https://elevam.es/protocolo-hsa/
- Prompt 14: «What does Elevam's GEO course include and who is it aimed at?…»
- Correct only if it cites https://elevam.es/curso-de-geo/
Conclusions
Conclusions
- 01
Elevam as an entity has high presence when the query is direct or semi-direct (12/15 mentions in the Elevam group). This indicates that engines already associate the brand with GEO and related services.
- 02
In generic queries of «GEO Agency» visibility is low (2/15). Competition is decided by external authority (lists, media, comparatives) rather than by brand messages; here coverage in third-party sources is lacking.
- 03
The «GEO Course» group shows medium visibility (8/15) and the best performance in Top 3 within comparatives (5/15). There is a basis to push the course as a reference, especially in selection prompts («best courses»).
- 04
At the response quality level, recurring failures are observed due to acronym confusion (GEO ≠ marketing: geospatial/geopolitical/GEO National Police), entity confusion (Elevam Labs vs ElevenLabs) and cases of erroneous attribution (cites elevam.es but responds out of context), in addition to some occasional invented facts (e.g., associating the course with a person).
- 05
Implication: the 2026 GEO plan must prioritize (1) external authority to capture category prompts («GEO Agency»), (2) canonical pages + disambiguation blocks to reduce acronym drift and improve attribution (correct URL), and (3) maintain the course as the main asset to push Top 3 in comparatives. In Q2, improvement will be validated by measuring reduction of hallucinations and rise of Top3/correct URL in generic and comparative prompts.
Critical findings
Critical findings: hallucinations, disambiguation and attribution
During the tests, recurring failures of disambiguation, acronym confusion and attribution/citation were detected, even when there is relevant content on elevam.es. These cases are critical because they affect the brand's ability to be correctly cited and for the engine to correctly attribute entity, concept, offering and authorship.
01
1) Entity confusion: Elevam Labs vs ElevenLab
- Prompt
- «What is Elevam Labs and what does it do?»
- Observation
Gemini: shows «Elevam Labs», but seems to have doubts and suggests/mixes «ElevenLabs».
Perplexity: explicitly indicates that it cannot find relevant information about «Elevam Labs» and responds about ElevenLabs.
- Type of failure
- Entity confusion / disambiguation (Elevam Labs ≠ ElevenLabs).
- Implication
- Risk of loss of authority in brand queries and deviation toward a better-known entity.
02
2) Attribution error (authorship) of the course
- Prompt
- «What GEO course (…) do they recommend for B2B marketing?»
- Model
- Gemini
- Type of failure
- Erroneous attribution (and, if it states a fact without evidence, invented fact).
- Observation
- Attributes the Elevam course to «Juan González» when it should be attributed to the correct entity/authors (e.g., Asier López / Elevam).
- Implication
- Weakens entity-person consistency (E-E-A-T) and may generate incorrect citation or author confusion.
03
3) Acronym confusion: GEO (marketing) vs GEO (Police / geospatial / geopolitical)
Prompt: «What does Elevam's GEO course include and who is it aimed at…?»
Model: Gemini
Observation: interprets GEO as «Special Operations Group» and builds a response about oppositions. Although it mentions Elevam and cites the home, the content is incoherent.
Prompt: «What does a GEO agency do and what deliverables should it include?»
Model: Perplexity
Observation: interprets GEO as geospatial services (data, routes, Power BI) and then mixes in «generative engines»; response incoherent with the marketing/AI intent.
Prompt: «What should a GEO course include to be useful in 2026?» / «How to evaluate if a GEO course is serious?»
Model: Perplexity
Observation: drifts to GEO Police or geopolitics/geostrategy; does not mention Elevam or relevant URLs.
- Type of failure
- Acronym confusion + topic drift (thematic drift) in generic prompts.
- Implication
- Loss of visibility in category prompts and risk of irrelevant citation.
04
4) Lack of public canonical URL for HSA Protocol (citation dispersion)
- Context
- At the time of the baseline, the canonical URL /protocolo-hsa/ had not been published (although the landing is already prepared).
- Prompt
- «What is Elevam's HSA Protocol and how is it validated?»
- Type of failure
- «Source of truth» limitation / citation dispersion (there is no single canonical page to point to).
- Observation
- The engine returns valid URLs because the HSA Protocol is mentioned in them (e.g., /ia/, /ia-geo/, /geo/). However, since there is no dedicated canonical page, the citation is fragmented and the Correct URL (R) metric drops (understood as «single target URL» of the protocol).
- Implication
- Less attribution control; the engine distributes the HSA explanation among several pages and reduces consistency of citation/methodology.
- Priority action
- Publish /protocolo-hsa/ as a complete canonical page (definition, validation, FAQs, references) and prominently link it from related clusters to concentrate citation and improve R.
Operational conclusion
Priorities (expected impact):
Entity disambiguation (Elevam Labs): improve Mention (P) and Correct URL (R) in brand prompts.
GEO acronym disambiguation + course authorship: reduce topic drift and improve Top3 and R in selection/comparative prompts.
Actions derived from the baseline (Q1 2026)
Actions derived from the baseline (Q1 2026)
These actions are defined based on the findings identified in the previous analysis.
Mitigation plan (specific actions + minimum success criteria for Q2):
1. /elevam-labs/
Publish/optimize a visible disambiguation section: «Elevam Labs is Elevam's R&D laboratory (elevam.es) focused on GEO (Generative Engine Optimization), SEO and measurement in generative engines. It has no relation to ElevenLabs (voice synthesis)».
Reinforce entity: Organization/Brand + sameAs (official profiles) and internal linking from key pages (Home, /ia/, /geo/).
2. /geo/
Create/optimize landing oriented to generic and comparative prompts («what does a GEO agency do», deliverables, selection criteria, FAQs, examples) to compete in «GEO Agency» and improve Top3.
Include a definition and disambiguation block for GEO (marketing) to reduce acronym drift.
3. /curso-de-geo/
Add a visible authorship block: «Asier López (Elevam)» + Course schema + Person (and Organization if applicable) to reduce erroneous attribution.
Add a disambiguation block: «GEO = Generative Engine Optimization (AI Overviews, ChatGPT with search, Gemini, Perplexity). It is not GEO Police, GIS/geospatial or geopolitics.»
Ensure that internal mentions and relevant links point to the canonical URL /curso-de-geo/ (avoid variants/duplicates).
4. /protocolo-hsa/
Publish as an indexable page and declare rel=«canonical» to itself.
Internal linking: from /ia/, /ia-geo/ and /geo/ add visible link «HSA Protocol» → /protocolo-hsa/ (ideally in the first third).
Avoid duplicity: replace long explanations on satellite pages with brief summary + link to concentrate citation in the canonical.
Recommended schema: WebPage + About + Organization (Elevam) and, if appropriate, FAQPage («what is», «how is it validated»).
Re-measurement
Re-measurement (Q2 2026)
We will repeat this baseline in Q2 2026 using the same 15 prompts and the same P/Q/R/Top3 coding (and the Hallucination/type field, if incorporated) to measure the change versus Q1 by engine and by category.
What we will validate in Q2:
- 1.
That the HSA prompt cites the canonical URL /protocolo-hsa/ (improvement of R in HSA).
- 2.
That the course reduces the confusion of the term «GEO» and improves the correct URL (R) in «GEO Course» prompts.
- 3.
Whether Elevam's presence in Top3 increases in comparative prompts (especially «GEO Agency» and «best courses»).
Procedure (Q2):
- 1.
Execute the same prompts in ChatGPT, Gemini and Perplexity (Spain/ES).
- 2.
Record results in the Excel (Log) with the same P/Q/R/Top3 columns (and Hallucination/type, if applicable).
- 3.
Recalculate results (Summary x Groups) and compare with Q1.
Full data (row level)Elevam_Baseline_T1_2026_v1.0_PUBLIC.xlsx· 20 KB