Authority and Credibility for AI

For Local Queries, AI Gives Huge Weight to Geographic Signals

You've been active in your city for years, but when a potential client asks the AI for the best plumber in Turin or the best accountant in Brescia, you never show up — and the job goes to someone who may have been around for less time but has taken better care of their local digital presence. For these geographic searches, AI weighs very specific signals: if one is weak or inconsistent, you're automatically excluded. Building that geographic authority takes an afternoon of work — and it can bring you local clients that you're losing every day right now.

Try asking an AI engine “best tax advisor in Cagliari”. Look at who comes up. Then ask yourself: is that professional really the best, or simply the one with the strongest geographic signals?

The answer, in the vast majority of cases, is the second one. I verify this regularly when I analyze the AI visibility of the companies I work with — and what I find completely changes the way you think about local presence.

The local query is a brutal filter

When a query contains a geographic reference — a city, an area, a “near me” — the AI engine activates a filter that changes the rules of the game. It’s no longer enough to be authoritative on the topic. You have to be authoritative on the topic in that specific place. And geographic authority isn’t a vague concept: it’s built with precise signals that the system can read.

The starting point is to understand that generative AI engines don’t work like Google Maps. They don’t have a database of local businesses with pins on a map. What they have is a corpus of text — crawled from the web, retrieved in real time via RAG, structured by knowledge graphs — in which they look for signals of association between an entity and a place.

If your brand appears across ten different sources always associated with the same city, the same geographic area, with the same contact details, the model builds a strong association. If the signals are weak, fragmented or absent, that association doesn’t exist — and when someone asks “the best X in Y”, you’re not even among the candidates.

Why context signals weigh more than you think

Srba et al. (2024) analyzed what determines the perceived credibility of a source, and the result is significant for anyone working on local visibility:

“Context-based signals considering user/source cues like domain reputation and publication metadata contribute most towards human judgement.”

Srba et al., 2024

Context signals — who the source is, where it’s published, what metadata accompanies it — weigh more than the content itself in determining credibility. For geographic authority this is a fundamental principle: the metadata of your online presence (address in the Google Business profile, NAP on directories, LocalBusiness markup on the site) are context signals. They’re exactly the kind of information that retrieval systems use to decide whether a source is relevant to a geolocated query.

It’s not an opinion. It’s the mechanism by which credibility assessment works in information systems — and generative AI engines inherit it.

Common mistake

If the name on your sign is different from the one on your profile, if the address doesn’t match the one on your site, if the category is generic instead of specific, you’re sending the system contradictory information.

Google Business Profile: the densest local signal you have

Of all the geographic signals, the Google Business profile is the most concentrated. In a single listing you have: canonical name, verified address, phone number, business category, hours, reviews with geolocated text, photos, posts. It’s a package of signals that tells the system “this entity exists in this place and does this work”.

Aggarwal et al. (2023) demonstrated that optimization strategies for generative engines produce measurable results:

“Through rigorous evaluation, we demonstrate that GEO can boost visibility by up to 115% in generative engine responses.”

Aggarwal et al., 2023

That 115% increase in visibility is the overall potential of GEO optimization. For local queries, a significant part of that potential runs through the Google Business profile — because it’s the source that concentrates the largest number of verified geographic signals in a single point.

But be careful: an incomplete or inconsistent profile isn’t neutral. It’s a negative signal. If the name on your sign is different from the one on your profile, if the address doesn’t match the one on your site, if the category is generic instead of specific, you’re sending the system contradictory information. And as I explained in the article on brand entity consistency, contradictions lower the confidence in everything that concerns you.

Pro tip

A complete and verified Google Business profile, consistent NAP citations on every directory, fresh reviews with geographic mentions, content that talks about your territory: these signals transform your generic authority into local authority.

NAP consistency: the repetition that builds certainty

NAP stands for Name, Address, Phone. Three trivial pieces of data. But their consistency across all the platforms where you appear is one of the most powerful signals for geographic authority.

The principle is the same as brand entity consistency, applied to the geographic dimension. If ten different directories report the same name, the same address and the same phone number, the system registers a very high overlap — and that overlap translates into confidence. If instead each directory has a different variant (“Via Roma 15” on one, “V. Roma 15/A” on another, and the landline on one but the mobile on another), you’re fragmenting the signal.

And this is where the structural bias of AI engines toward third-party sources comes into play. Chen et al. (2025) document it without ambiguity:

“AI Search exhibit a systematic and overwhelming bias towards Earned media — third-party, authoritative sources.”

Chen et al., 2025

Local directories, the industry publications that mention you, chambers of commerce, the vertical portals of your area — they’re all third-party sources. They’re exactly the kind of earned media that AI engines favor. If your NAP is consistent across these sources, every mention reinforces your geographic authority. If it’s inconsistent, every mention introduces noise.

Local mentions and geolocated reviews

There’s a level that goes beyond structured data: the free text that talks about you in a geographic context.

A review that says “the best accountant in Cagliari, has been following us for years with expertise” contains two signals: a positive judgment and an explicit geographic association. For the retrieval system, that sentence is a passage that links your entity to a place and a category — and it’s generated by a third party, which makes it earned media.

Mentions in local publications work the same way. An article in a local newspaper that cites you as “the consultancy firm on Via Roma” creates an entity-place association on an authoritative source. Ten similar articles on different sources build a pattern that the model can’t ignore.

The mechanism is no different from the one I described in the article on the brand-category association: cross-source repetition solidifies the association. The difference is that here the association isn’t just brand-category, but brand-category-place. A triangle that, if solid, makes you practically unbeatable for local queries in your sector.

The starting check: where you are today

You can get an idea of your situation in less than half an hour.

First step: search for your brand name on an AI engine with a local query — “best [your service] in [your city]”. Do you show up? If so, is the information correct? If you don’t show up, who’s in your place and why?

Second step: search for your brand on Google and open the first ten directories where you appear. Compare name, address and phone number. Every discrepancy is a point of fragmentation in the geographic signal.

Third step: read your Google reviews. How many explicitly mention your city or area? How many are from the last six months? Old reviews without geographic references contribute little to current geographic authority.

These checks give you a direction. But mapping all the sources where your NAP appears — including the ones you don’t monitor — requires specific tools and skills. It’s the difference between knowing that the problem exists and knowing exactly where to intervene.

The link to the brand’s overall authority

Geographic authority doesn’t live in isolation. It’s a layer that overlaps with all the other brand authority signals — and amplifies or weakens them depending on consistency.

If you’ve built strong authority as a founder but your brand has no clear geographic signals, you lose all the local queries. If you’re present on local directories but your brand is fragmented across multiple identities, geographic authority disperses among entities that the model can’t connect. If you have very strong local signals but competitors have worked on displacement better than you, they’ll occupy those slots even in geolocated responses.

It all hangs together. But for those who operate in a local market — and in Italy most businesses operate on a territorial basis — geographic authority is the multiplier. A complete and verified Google Business profile, consistent NAP citations on every directory, fresh reviews with geographic mentions, content that talks about your territory: these signals transform your generic authority into local authority. And for local queries, it’s local authority that wins.

Chapter 2 · Authority and Credibility for AI

Continue with the deep dives

40 deep dives across the 5 sections of the chapter.

2.1 Authority Signals 8 deep dives
2.2 Brand Authority 8 deep dives
2.3 Sources & Citations 7 deep dives
2.4 Technical Credibility 8 deep dives
2.5 Trust & Reputation 9 deep dives
The author
Roberto Serra at the Senate of the Republic Senate of the Republic · Palazzo Giustiniani Conference “The power of artificial intelligence”
Roberto Serra Roberto Serra

SEO consultant for over 15 years, founder of the Serra SEO Agency (RAANK). He helps multinationals and SMEs stay visible where search is moving: ChatGPT, Perplexity, Gemini and Google's AI Overviews.

As featured in
ANSA Il Sole 24 Ore Le Iene Università di Cagliari La Repubblica
How visible is your brand to AI? Analyze your brand