You're the best in your field, but when a potential client asks ChatGPT who the go-to expert is on that subject, someone else always comes up. It's not a competence problem — it's that the AI doesn't know you're the expert, because no authoritative source has ever said so in a way the models understand. Every client who relies on that answer is a client who calls your competitor. Building the right associations between your name and your expertise, on the right channels, is faster than you think.
When AI has to answer a question about a topic in your industry, does it include you as an expert commentator or does it always cite others? It’s a question worth asking, because the answer explains why some professionals end up cited by ChatGPT, Perplexity and Gemini, and others, equally good, stay invisible.
Let me explain how the mechanics work: when a journalist interviews you and introduces you as an “expert in maritime law” or a “specialist in insurance litigation”, that phrase ends up in an online article. The article gets indexed, ends up in the training datasets of the models and, subsequently, in the retrieval pipelines that feed AI answers. The formula “First Name Last Name, expert in X” is the mention format that teaches the AI your authority more cleanly than a generic backlink.
In this article I’ll show you how to build an expert commentary strategy that brings your name, paired with the right qualification, into the places where AI goes to look for sources.
The mechanism: why a journalist’s mention weighs more than a link
In the world of research on language models there is a very clear line of work on the topic of source credibility. Schuster and colleagues, in a 2026 study on the 13 most widely used models, investigated precisely how AI chooses between conflicting sources.
“Although we consider our trust hierarchy in Section 4 to be representative of real-world scenarios as well, there are particular examples where this may not hold, modulated by the style and topic of the message, as well as the expertise of the source.”
Translated for your business: the declared expertise of the source is one of the signals the models use to decide whom to listen to when there is contradictory information. It’s not just the domain that publishes, it’s also who is presented as authoritative within the piece.
The operational consequence is simple. If a reporter from Sole 24 Ore writes “according to attorney Mario Rossi, an expert in shipping and maritime law”, that signal is qualitatively different from a simple link to the firm’s website. There is an assertion of expertise tied to a name, inside an article from an outlet with a good reputation. It’s exactly the kind of structure the model learns to recognize.
The same authors explain the framework with which they measured the phenomenon:
“We address this gap with a novel framework to investigate how source preferences affect LLM resolution of inter-context knowledge conflicts in English, motivated by interdisciplinary research on credibility.”
The part that matters to you as a business owner is the last sentence: repeating information can flip preferences. Repetition changes the model’s preferences. A comment given just once is an event. The same concept, repeated across multiple outlets by different bylines who cite you as an expert, becomes a pattern. And patterns are what the models absorb.
Why this piece sits upstream of everything else
In the previous articles in this series I insisted heavily on the fact that AI reasons by entities, not by keywords. I told you about how recognizing the author as an entity works and how implicit mentions carry a specific weight within the models.
Expert commentary is the fuel for that mechanism. Every time a journalist qualifies you with “expert in X”, you are feeding the entity “First Name Last Name” with a domain attribute. It’s the cleanest way to build what in other pieces I’ve called the backlink as citation proxy: it’s not the link that counts, it’s the context of authority in which your name is uttered.
The journalist asks for a technical opinion on the Rotterdam Rules and you reply with general commercial law.
The platforms where journalists look for sources
There are three platforms where journalists, including Italian ones, post requests for comment. They work like marketplaces: they ask for an expert source on a topic, you reply with an opinion of 100-300 words, and if the answer is good they cite you.
- Connectively (formerly HARO): historically the largest, with many requests also from international outlets writing about Italian topics (shipping, luxury, food, fashion).
- SourceBottle: more oriented toward lifestyle and B2C, useful if your industry touches consumer topics.
- Qwoted: born in the US market but growing in Europe, good for finance, tech, professional services.
For a law firm specialized in shipping — let’s take a hypothetical case, a firm based in Messina that serves Sicilian shipowners and shipping companies of the eastern Mediterranean — the interesting requests come when cases break out like containers lost at sea, cargo policy disputes, groundings, international sanctions. In those moments the financial outlets look for technical voices.
Keep a ready “qualification one-liner” that the journalist can copy: “Attorney, partner at [Firm], specialized in maritime law and transport insurance litigation, Messina”.
The first-person test: six shipping firms monitored for six months
Let me tell you what I observed over the last six months on a sample of six Italian law firms specialized in shipping and maritime law, three from the Genoa-La Spezia cluster and three across Naples, Messina and Catania. I monitored them with periodic queries on ChatGPT and Perplexity — things like “best shipping lawyers in Italy”, “law firms expert in maritime law”, “who to consult for container litigation” — plus a monthly reading of mentions in the specialized press.
Of the six firms, two have a structured expert commentary practice: they regularly respond to trade outlets, get interviewed on maritime news cases, publish brief technical opinions whenever something relevant happens. The other four publish excellent legal notes on their own website but never appear in the media.
The result after six months: the two “commentator” firms appear named in AI answers on their domain in 60-70% of the tested queries; the other four appear sporadically or never, even though their site is technically well-maintained and the content is of good quality.
It’s an indicative test, not a controlled study: six firms are few and the starting queries reflect my framing. But the pattern is clear enough that it can’t be ignored. The real analysis, with large samples and queries generated in a neutral way, requires dedicated professional monitoring tools.
The mistakes I see most often
When I talk about expert commentary with Italian professionals, I see four mistakes recur:
- The generic “jack-of-all-trades” answer. The journalist asks for a technical opinion on the Rotterdam Rules and you reply with general commercial law. When the AI absorbs that piece, it won’t have a clean specialization signal.
- No specific qualification in the byline. Saying “attorney” is not enough. Saying “attorney, specialized in maritime litigation and transport insurance law” creates the expertise attribute that the entity in the knowledge graph stores.
- Responding only to Italian requests. Outlets like Lloyd’s List, TradeWinds, Splash247 also cover Italian cases and their articles weigh heavily in English-language training datasets. Excluding them means leaving out a primary channel.
- A comment with no data. An opinion is worth little. An opinion with a number — “in Italy litigation over damaged cargo rose 30% in 2025 according to our observatory’s data” — is a citable hook.
How to set up a sustainable practice
If you’re a firm owner or head of communications, here is a three-step audit before deciding whether to invest seriously in expert commentary:
- Open the firm’s site Google Search Console and look at the brand queries: if almost everything is “firm name + city”, the thematic domain signal is missing. Expert commentary serves exactly to build it.
- Try 5-7 thematic queries from your industry on ChatGPT and Perplexity (in the shipping case: “Italian maritime law experts”, “firms specialized in container litigation”, “who to consult on Russia maritime sanctions”). If you never appear, the problem isn’t the site, it’s the absence of contextualized mentions.
- Compare with the 3-5 competitors that AI cites in your industry. Search their names on Google News: you’ll almost certainly find 10-30 mentions with an expert qualification in the last 12 months. That’s the benchmark you have to match.
Once the starting point is clear, the concrete practice is this:
- Sign up for Connectively and Qwoted. Spend 30 minutes a day, two or three times a week, scrolling through the requests.
- Set yourself a realistic goal: 5-10 responses per month in your narrow field. Don’t respond to everything, respond well where you are truly a specialist.
- Keep a ready “qualification one-liner” that the journalist can copy: “Attorney, partner at [Firm], specialized in maritime law and transport insurance litigation, Messina”.
- After every citation obtained, check that the piece is indexed (search Google for a phrase from your comment in quotes) and keep an internal log.
The thread that leads to AI answers
The expert commentary strategy is not a magic wand. On its own it won’t put you in first position in ChatGPT’s answers, and it’s not enough to compensate for a fragile site or zero social presence. But it is the piece that makes everything else readable as an authoritative entity: without expertise qualifications associated with your name in the media, the AI struggles to build your implicit profile.
In the next articles in this series we go deeper into the operational side: how to set up a newsjacking campaign for AI, how to turn internal research into data PR that the engines cite and how to build a foothold in industry trade media.
The point is always the same: appearing in AI answers is not a shortcut, it’s the sum of authority signals you build over months of orderly work on the right sources.
