You can spend 30,000 euros in a year on content, website and AI-related PR and reach December with ChatGPT still not citing you. It's not a matter of figures: it's a matter of where the money ends up. Spreading the budget across the right levers — the ones AI engines actually use to decide who to cite — turns a scattered expense into an investment that produces measurable results.
AI visibility budget: 10k, 30k, 100k a year? Without a prioritization framework, you spend badly. I’ll show you how to allocate across SEO, PR, Entity and Platform so that every euro works to get you surfaced in AI answers.
The symptom is always the same: you reach the end of the year, you’ve spent 30 thousand euros across SEO agency, press office, website fixes, events. Your Google impressions may be holding up. But you open ChatGPT, you ask “best boutique hotel Treviso hills Asolo” and your name isn’t there. It cites two competitors who spend less than you.
It’s not bad luck. It’s that you allocated the budget as if it were 2018, when the problem was ranking on Google. Today the problem is surfacing in AI answers, which is a different game and rewards different levers.
Why, without a framework, you end up spending on the least useful thing
When an entrepreneur asks me “Roberto, I have 30 thousand euros a year, where do I put them?”, the temptation is to answer with whatever’s trendy at the moment: a prompt engineering course, a company podcast, a website redesign. They can all make sense, but in what order, and with what weight?
In my previous articles I showed you the individual mechanisms: how AI recognizes named entities, how it weighs implicit mentions of your brand, how it uses backlinks as an authority proxy. Each mechanism is a potential spending line. The problem is that you can’t play them all with the same intensity.
You need a framework that lines up the actions by estimated impact, cost and time to return. It’s an exercise in applied common sense, not a magic formula: I present it as an explicit deduction from the way these systems weigh signals, not as data from a paper.
The three dimensions that decide the spending order
Every action you want to budget is assessed on three binary/ternary axes, no continuous metrics for a financial analyst.
Impact on AI visibility: high, medium, low. Fixing the Organization schema on the homepage has high impact because it touches the first level of the model’s understanding. Opening a company TikTok channel has low impact, today, on generative AI engines.
Annual cost: under 3k, between 3k and 15k, above 15k. I’m not talking about agency cost in the abstract, I’m talking about euros that leave the bank account for that single line item.
Time before seeing signals: fast (within 60 days), medium (3-6 months), slow (over 6 months). The signals are measurable things: appearing in Perplexity for brand queries, being cited in ChatGPT answers on category queries, an increase in tracked implicit mentions.
The prioritization rule I use is simple: first everything that is high impact + low cost + fast time. Then high impact + medium cost + medium time. Last high impact + high cost + slow time, because there you’re making a long bet and you need to have already sorted out the rest.
Replicating the previous year’s budget without reallocating.
Real case: a boutique hotel in Asolo, 30k a year
Let me tell you about a case I worked on because it helps more than a thousand abstract diagrams. Charming boutique hotel in the Treviso hills, Asolo area, 14 rooms, mid-to-high positioning, target Italian and European couples looking for a long weekend. Marketing budget dedicated to digital visibility: 30 thousand euros a year, already spent in the past across generalist SEO, photographer, social media management.
When we started, on ChatGPT the query “romantic boutique hotel Treviso hills weekend” didn’t cite it. On Perplexity neither. Two competing properties in the Conegliano area and one in Valdobbiadene showed up. The ownership asked me: “where do we cut, where do we add?”.
We applied the framework. I’ll summarize it line item by line item, so you can see how to reason.
Priority 1 — high impact, low cost, fast (target 6k a year)
Optimization of schema markup Hotel + LocalBusiness + FAQPage on the homepage and on the 14 room pages, with GPS coordinates, price range, certifications. One-off cost of 1,500 euros, impact time 30-45 days. Update of the property’s Wikidata entry with relations to “Asolo”, “Prosecco Hills”, “medieval village”. Zero cost, time 60 days. Cleanup of the Google Business Profile with detailed services, FAQ, weekly posts. Cost 2k a year for management. The rest went into small technical work on the site.
Priority 2 — high impact, medium cost, medium time (target 14k a year)
Three in-depth articles a month on the blog, editorial angle (itineraries in the hills, pairings with Asolo’s Prosecco DOCG, the history of the fortress), signed by the owner with a well-structured bio and author entity. Editorial cost 12k a year. Plus 2k for targeted PR on 4-5 mid-to-high quality Italian vertical travel outlets, to obtain implicit mentions of the property name in the context “romantic weekend Veneto”.
Priority 3 — high impact, high cost, slow time (target 10k a year)
Participation in and sponsorship of two local food-and-wine themed events with the owner as a speaker on “charming hospitality in the Veneto hills”. This serves to build event authority and to name the brand in contexts where it ends up in transcripts, programs, press releases. Impact time: 9-12 months. It’s the line item I’d give up first if the budget were 15k instead of 30k.
After 8 months, on an indicative test (not a scientific study, a sample of 25 category queries on ChatGPT, Perplexity, Gemini) the property was cited in 11 answers out of 25, against 0 at the start. Not magic: disciplined application of the framework.
Open the Rich Results Test on your homepage: if no Organization structured data comes out, you have a priority 1 action sitting undone that you should do before anything else.
How to reason if the budget is 10k or 100k
If you have 10k a year, you’re practically only in priority 1. Schema, Wikidata, Google Business Profile, two well-made articles a month, and a small management line. No events, no structured PR, no book. Is it little? It’s just enough not to be invisible on brand queries and to start appearing on local category queries.
If you have 30k a year, you replicate the structure of the case above: 6k priority 1, 14k priority 2, 10k priority 3. It’s the bracket where the framework pays off the most, because every euro is still forced to justify itself.
If you have 100k a year, priority 3 expands: ongoing press office, publishing a book or authored ebook (which builds a stable author entity), participation in 6-8 events a year, a recurring video format. The risk in this bracket is the opposite: you spend a lot on high-cost items and you forget that without priority 1 in order the rest holds up less. I’ve seen companies with a 100k budget and broken schema markup on the homepage.
The mistakes I see most often when people allocate without a framework
Spending 70% on priority 3 because it’s more impressive. Events, books, podcasts are appealing. But if the technical base isn’t there, AI struggles to connect that noise to your specific brand.
Not quantifying the impact time. You sign an 18k-a-year contract with a PR agency and expect signals in 60 days. It doesn’t work like that, and after 4 months everything gets cut because “we’re not seeing anything”.
Replicating the previous year’s budget without reallocating. If last year you put 15k into classic SEO and the AI results stay at zero, it makes no sense to confirm the line. It must be shifted, at least in part, onto new levers.
Confusing impact on Google with impact on ChatGPT. They’re two overlapping but not identical games. There are actions useful for both (schema, authority, mentions) and actions more specific to one or the other.
The mini-audit you can do yourself in 30 minutes
Take your digital marketing spend over the last 12 months, line item by line item. For each item answer three questions: impact on AI visibility high/medium/low, annual cost bracket, time to return fast/medium/slow.
If more than 50% of the budget is in low-impact or slow-time items, you have an allocation problem, not an overall budget problem. Open the Rich Results Test on your homepage: if no Organization structured data comes out, you have a priority 1 action sitting undone that you should do before anything else.
Then compare yourself with the 3-5 competitors that AI cites when you query your category queries. Look at what they do that you don’t: that’s the short list to reason about for the next budget. Real analysis requires professional tools and dedicated time, but this first step you can take today.
