Skip to research results

Public report summary 2026 Edition · Wave 1

AI Energy Choice
in Italy

When an Italian consumer asks ChatGPT, Gemini or Google AI surfaces for information, comparisons and recommendations about electricity and gas suppliers, different brands, alternatives and preferences emerge. This research measures who appears, who enters consideration and who is explicitly recommended.

Wave 1 establishes the 2026 baseline: proprietary research across 2,589 responses, four AI surfaces and 19 observed suppliers, using Telescop methodology based on structured classification and deterministic aggregation to distinguish presence, consideration and recommendation.

AI CHOICE INTELLIGENCE

This is not just AI Visibility. It measures what happens after presence.

VISIBILITY Does the brand appear?

Measures brand exposure in responses.

AI CHOICE What role does the brand play when AI enters the decision?

Distinguishes Exposure, Consideration, Recommendation and First Choice where the surface semantics support it.

WAVE 1 · IL PUNTO ZERO

This edition creates a versioned baseline: today it measures a difference; future waves can measure change from the same starting point.

Premium Research €2.900 final price Buy Wave 1 Compare licenses
Report publication
September 12, 2026
Observation wave
Wave 1 · September 2026 · Europe/Rome
Research ID
TEL-ENERGY-CHOICE-IT-2026-W1
Access
Public summary
Full report · Premium

Presence and recommendation are not the same thing.

The same category produces different competitive hierarchies depending on the AI surface. A highly visible brand is not necessarily the one most often pushed toward choice.

58,865%Leader cross-surfaceExposure media di Sorgenia.
47,268%Leader RecommendationLower bound ChatGPT di Sorgenia.
75,580%Leader GeminiExposure di Edison Energia.
47,197%Leader AI OverviewExposure di Plenitude.

Study base. 2,589 responses overall: 732 ChatGPT, 733 Gemini, 553 Google AI Overview and 571 Google AI Mode. “Brand present” is not equivalent to “brand recommended”, market share or a sale.

The analytical perimeter behind the published results.

Wave 1 separates the Primary Generic panel collection design dal supplier corpus used in the published analyses. They are different analytical perimeters and should not be treated as the same denominator.

245decision moments eleggibili
Primary Generic
3replicates per moment
in the collection design
2.938 / 2.940osservazioni valide / attese
Primary Generic
2.589responses in the supplier corpus
used in the published analysis
Why three replicates? They prevent the benchmark from relying on a single model response. They do not eliminate intrinsic AI variability: they make it observable within the research design.

The signal that presence-only monitoring does not show.

Knowing that a brand appears in AI is only the starting point. What defines the observed competitive profile is where it appears, at which decision stage and with what strength relative to competitors — information that Exposure alone does not provide.

When Exposure and Recommendation diverge, the signal becomes strategically relevant: it can indicate where to focus attention, what to investigate in content and PR, and where observed presence does not match Recommendation strength. Differences across ChatGPT, Gemini and Google surfaces amplify these asymmetries unevenly across suppliers.

For decision makers

The research provides a structured observational baseline to understand where a brand is gaining or losing ground across AI surfaces involved in decision making — and to turn those differences into priorities to validate across budget, content, PR and competitive positioning.

For teams that need to read the market before deciding.

Energy Choice is not designed to produce a generic “AI score”. It is built for teams that need to understand where the competitive hierarchy changes, which signals deserve deeper investigation and where to focus attention, analysis and budget first.

01CMO · Marketing Director

To understand whether the observed competitive hierarchy across AI surfaces is changing and which areas deserve managerial attention first.

02Head of Digital · SEO · GEO

To read differences across surfaces, identify gaps and decide where to deepen content, sources, search and informational coverage.

03Brand · Communication · PR

To observe whether the brand enters the response as simple presence or takes on a stronger role during consideration and choice.

04Strategy · Innovation · AI Lead

To establish a proprietary baseline and measure how the observed competitive profile changes in later waves.

Una categoria. Quattro gerarchie competitive diverse.

The Exposure leader changes with the observed surface. This is the first reason why a single “AI visibility” metric is insufficient.

Una sola classifica nasconderebbe il fenomeno.

ChatGPT, Gemini, Google AI Overview and Google AI Mode do not return the same competitive hierarchy. Energy Choice keeps surfaces separate instead of artificially compressing them into one score.

Figura 1

Exposure leader by surface

Share of responses in which the supplier is present · 0–100% scale

ChatGPT · Sorgenia 84.426%
Gemini · Edison Energia 75.580%
Google AI Mode · Sorgenia 57.793%
Google AI Overview · Plenitude 47.197%
0%100%

Values describe observed Exposure within each surface. They are not market share and are not directly a Recommendation metric.

Fonte: Telescop Research · AI Energy Choice Italy 2026.

What it means. There is no single competitive ranking that is valid across all AI systems.

What it changes. Monitoring only one platform can hide competitors that systematically emerge elsewhere.

In practice. Leadership should be interpreted by surface before any cross-surface synthesis.

How uniform is supplier presence across AI systems?

The cross-surface average summarizes Exposure across the four surfaces. It is a descriptive measure, not a weighted index.

Figura 2

Top three suppliers by average cross-surface Exposure

Simple average of Exposure percentages observed across the four surfaces

Sorgenia 58.865%
Edison Energia 56.469%
Plenitude 47.196%
0%100%
Supplier 4 38.413%
Supplier 5 36.575%
Supplier 6 36.125%
Supplier 7 32.952%
PREMIUMResearch Complete ranking of all 19 suppliers

The report includes values, positions and cross-surface comparison for all observed suppliers.

See what the report includes →

The cross-surface average is useful for orientation, but it does not replace surface-level analysis.

Being present does not mean being recommended.

Recommendation identifies an explicit selection or preference toward a specific supplier. In this wave it is published only for ChatGPT, where the Wave 1 decision baseline is Recommendation Engine v1.4.1.5.

Figura 3

Top three suppliers by Recommendation lower bound on ChatGPT

Conservative share of 732 ChatGPT responses · 0–100% scale

Sorgenia 47.268%
Enel Energia 28.689%
Edison Energia 22.814%
0%100%
Supplier 4 20.492%
Supplier 5 17.486%
Supplier 6 15.437%
Supplier 7 15.164%
PREMIUMResearch Complete ranking + upper bound

The report includes all suppliers, lower bound, upper bound and the main gaps between Exposure and Recommendation.

Access the details →

Cases model_review_required are not converted into false negatives. The published value is therefore deliberately conservative.

What it means. Two brands with similar Exposure can have very different decision strength.

What it changes. Visibility alone can overstate or understate the brand’s role in choice.

The research does not decide for you. It changes where you start investigating.

The benchmark becomes operationally useful when an observed difference is turned into a verifiable question without automatically assigning a cause.

Observed signal Question it opens
High Exposure · weaker Recommendation The brand is mentioned, but is it advanced with the same strength when the response enters the decision?
Strong asymmetry across surfaces Where does the competitive hierarchy change, and which surface deserves dedicated analysis?
Competitor emergente in specifici momenti Which family of decision moments concentrates the gap, and which audit should be opened first?
Delta across successive waves Il segnale resta episodico oppure sta diventando un cambiamento persistente rispetto alla baseline?

Guardrail: a gap identifies where to investigate. It does not by itself prove why the gap exists.

Where presence and decision strength separate.

A ranking orders. The map distinguishes the profile. Crossing average cross-surface Exposure with ChatGPT Recommendation reveals different competitive configurations. The complete map with coordinates and clusters is part of the premium report.

Figura 4

Energy Choice Map · anteprima metodologica

Asse X: Exposure media cross-surface · asse Y: Recommendation lower bound ChatGPT

Exposure → Recommendation → High presence
High decision strength
Lower presence
High decision strength
High presence
Lower decision strength
Lower presence
Lower decision strength
PREMIUMResearch 19 suppliers positioned

Coordinates, clusters and brand-by-brand asymmetries are available in the full report.

See the full report contents →

The preview shows the map logic but does not publish coordinates or positions of individual suppliers.

Public insight. A2A Energia shows stronger ChatGPT decision strength than its average cross-surface position would suggest. Enel Energia instead shows much stronger presence on ChatGPT than on Gemini.

Dal benchmark alla decisione da approfondire.

The research does not replace strategy and does not automatically assign a cause to a gap. It narrows the field: where to investigate, what to compare and which area to explore first.

Il passaggio chiave Observed signal → managerial question → investigation area. This is where the benchmark becomes a decision-enablement tool.
01 AUDIT COMPETITIVO

Which competitors should we watch first?

Identify operators that emerge on surfaces or decision moments that the usual competitive set may not capture.

02 CONTENT · SEO · GEO

Where should we focus verification?

Understand where the brand appears, where it loses strength and which surfaces or decision families deserve deeper analysis.

03 BRAND & REPUTATION

Does presence translate into decision strength?

Separate simple Exposure from consideration and recommendation signals where surface semantics allow it.

04 MEDIA & ADVERTISING

Where should a gap enter the evaluation?

Use the benchmark as one signal to cross with demand, audience and performance before deciding how to allocate media pressure.

05 WAVE & BASELINE

Is the signal changing over time?

Keep Wave 1 as the zero point for reading deltas in future waves, not just the current snapshot.

06 MANAGEMENT

Which question should we bring to the table?

Turn an observed signal into a priority to validate before investing time, budget or analytical capacity.

The benchmark narrows the field before investment. It does not invent the cause: it makes the next question more precise.

Five self-contained findings, with scope and source.

These excerpts summarize only results already public in Wave 1. They are written so they can be cited without reconstructing meaning from different parts of the page.

Scope
Telescop Research analyzed 2,589 responses related to the Italian electricity and gas market across ChatGPT, Gemini, Google AI Overview and Google AI Mode in the September 2026 Wave 1.
Cross-surface finding
Sorgenia records the highest publicly shown average cross-surface Exposure in Wave 1, at 58.865%.
Recommendation finding
On ChatGPT, Sorgenia records a Recommendation lower bound of 47.268%, the highest among the suppliers publicly shown in the ranking.
Surface finding
Exposure leaders differ across the four observed surfaces: the research therefore does not identify a single competitive hierarchy valid across all AI systems.
Method note
Recommendation is published in Wave 1 only for ChatGPT; for Gemini and Google surfaces the research uses Exposure without automatically transferring the same decision semantics.
Public research assets Metadata, already-public data and citation formats.
PREMIUMResearch

The public page orients. Premium opens the competitive benchmark.

Method and core findings remain publicly readable. The full report adds rankings, coordinates, comparisons and brand-by-brand profiles needed to move from reading the phenomenon to analyzing your market.

PUBLIC

Capire il segnale

  • evidenze aggregate e leader pubblici
  • logic of metrics and surfaces
  • Choice Map preview
  • methodology and reading caveats
CORPORATE

Portarlo nei team

  • tutto il contenuto Premium
  • workbook XLSX aggregato
  • Executive Deck e chart pack
  • multi-team license within the same legal entity
Materiali inclusi AI Energy Choice · Italy 2026 · Wave 1
01Complete ranking of all 19 suppliers by cross-surface Exposure.
02ChatGPT Recommendation: lower bound, upper bound and complete ranking.
03Complete Energy Choice Map with clusters, coordinates and asymmetries.
04Detailed comparison across ChatGPT, Gemini, Google AI Overview and Google AI Mode.
05Brand-by-brand profiles with strongest and weakest surface.
06Methodology & Data Notes IT + EN, definizioni operative, caveat e criteri di lettura.

Scegli il diritto d’uso adatto alla tua organizzazione.

Italian + English · protected digital delivery

Corporate Intelligence

€4.900 final price

  • Multi-team use within the same legal entity
  • Workbook XLSX aggregato
  • Executive Deck + chart pack
  • Executive Briefing 60 minuti · Wave 1
Buy Corporate

Agency / Advisory

€4.900 final price

  • Uso professionale client-facing
  • Singoli insight e grafici in presentazioni private
  • Attribuzione Telescop richiesta
  • Nessuna redistribuzione o white label
Buy Agency

Final prices. No VAT or other tax is added at checkout. Premium Research can add up to three named users at a final €400 each; from seven users or for multi-team use, Corporate Intelligence applies.

Procurement, PO o condizioni contrattuali particolari? Checkout is the standard path. For enterprise or administrative requirements, contact Telescop Research.
research@telescop.it

Disegno prima del risultato.

The pipeline separates simple presence from decision strength and keeps the semantics of individual surfaces distinct.

  1. Wave 1 panelDecision moments and suppliers defined before synthesis.
  2. Raccolta multi-surfaceChatGPT, Gemini, AI Overview e AI Mode.
  3. Claim intelligenceText blocks → entity linking → claim extraction → validation.
  4. MetricheExposure, Consideration, Recommendation e First Choice.
Final corpus
2,589 responses: ChatGPT 732; Gemini 733; Google AI Overview 553; Google AI Mode 571.
Exposure
The supplier appears in the response.
Consideration
The supplier is advanced as a concrete, evaluable option.
Recommendation
The AI expresses an explicit selection or preference toward a specific supplier.
First Choice
Richiede una posizione esplicita top/first/winner.
Decision baseline
ChatGPT Recommendation Engine v1.4.1.5, the versioned Wave 1 baseline.
Lower bound
Cases not resolved with sufficient confidence remain model_review_required and are not converted into false negatives.
Composite score
Not used. The cross-surface average remains a descriptive summary.

Transparency. Recommendation is published only for ChatGPT in this wave. For Gemini and Google surfaces, Exposure is used without automatically transferring ChatGPT’s decision semantics.

AI choice is not a single ranking.

The most useful result is not to declare one absolute winner, but to show that presence, recommendation and surface produce different competitive readings.

For an energy supplier, this means the question is not only “how visible are we in AI?”, but “on which surface, at which stage of choice, and in what role?”

Purchase, licensing and use of Wave 1.

What is included in Premium Research?

Premium Research includes the complete report in Italian and English, Executive Summary, complete rankings, Energy Choice Map and brand-by-brand analysis for one legal entity and up to three named users.

What is the difference between Premium, Corporate and Agency / Advisory?

Premium is designed for internal use by up to three named users. Corporate extends use to multiple teams and includes the XLSX workbook, Executive Deck and chart pack. Agency / Advisory enables professional client-facing use of individual insights and charts in private presentations with Telescop attribution.

Is the displayed price final?

Yes. €2,900 for Premium and €4,900 for Corporate or Agency / Advisory are final prices: no VAT or other tax is added at checkout. Tax treatment is determined from billing details.

How can I pay?

Checkout supports cards and wallets when available, or SEPA bank transfer through Stripe. Il pagamento avviene in un ambiente sicuro gestito da Stripe.

When do I receive the research?

Card purchases are released after server-side payment confirmation. Bank-transfer delivery starts only after Stripe confirms that funds have settled.

Can I share or cite the data?

Rights depend on the purchased license. Premium is for internal use; Corporate enables multi-team use within the same legal entity; Agency / Advisory allows client-facing use of individual insights and charts with attribution. Redistribution of the report or database and white-label use are not included.

Is the research available in English?

Yes. Wave 1 is available in both Italian and English.

How to use this research.

Web report

Telescop Research. (2026). Telescop AI Energy Choice - Italy 2026 · Wave 1. Research ID: TEL-ENERGY-CHOICE-IT-2026-W1. https://telescop.it/research/ai-energy-choice-italia/

The public page is an editorial summary. Complete rankings, full tables, the Energy Choice Map and the complete PDF are part of the premium report.

Telescop ResearchResearch ID · TEL-ENERGY-CHOICE-IT-2026-W1