Strong on Google, invisible in AI: a barometer of 251 Portuguese B2B companies, 2026

The new starting point of supplier search

Supplier research is changing venue, and recent international studies size the shift. A G2 survey (The Answer Economy, 2026) puts at 51% the share of B2B buyers who turn to an AI assistant more often than to Google, up from 29% eleven months earlier. Forrester (2026), in a survey of around 18,000 buyers, reports that 94% used a large language model at some point in the purchase process. A growing share of B2B buying journeys therefore begins in an assistant rather than in a search box.

Third-party conversion data is directional and contested, and this barometer treats it as signal, not established fact. Some studies suggest that traffic arriving from assistants converts several times better than organic traffic; others find small or non-significant differences. The point worth retaining is simpler: the channel exists, it is growing, and strength on Google does not transfer to it on its own.

The gap metric and the overall picture

The gap between the two channels is defined as: gap = Google strength × (100 - AI visibility) ÷ 100, with both signals on a 0 to 100 scale. A high gap means strong on Google and weak in AI, in other words a large amount of presence still to carry over to the new channel. A low gap means the two signals are already close.

On average across the 251 companies, Google strength is 30 of 100 and AI visibility is 18 of 100: presence on Google is 1.7 times presence in AI answers. The averages, however, say less than the coverage figures. 41% of the companies are totally invisible in AI, with zero mentions in the answers tested, and 59% appear in 10% or fewer of the buyer questions run.

This is not a product quality problem; it is a presence problem. The typical company in the sample exists, has a website and ranks on Google, and is still not named when a buyer asks the assistant about its category. Absence from the shortlist means absence from the conversation, regardless of how solid the underlying business is.

Where the mismatch is largest: maximum gap by sector

The gap is not uniform across the 18 sectors, and it peaks where Google strength is high and AI presence is close to zero. The maximum gap measured per sector: lawyers 72, software for clinics 70, property (condominium) management 62, GDPR consultancy 49, industrial waste management 48, HR software 47. These are the segments with the most presence still to carry over into the new channel, and therefore the largest untransferred opportunity.

At the opposite end, some sectors are already comparatively more present in AI and show smaller gaps: waste, human resources, invoicing and fleet renting. The correct reading is not that these sectors are solved; it is that the distance left to close between their Google strength and their AI visibility is smaller.

The pattern in three anonymised examples

Three anonymised cases make the pattern concrete (the sector is given, not the company), with Google strength, AI visibility and the resulting gap: a law firm scores 87 on Google, 17 in AI, gap 72; a clinic software product scores 93 on Google, 25 in AI, gap 70; a condominium management company scores 80 on Google, 25 in AI, gap 60.

The same profile repeats across all three: solid reputation and ranking on Google, residual presence in AI answers. The strength of one channel is not reaching the other, which is precisely what the gap metric is designed to expose.

Why early entrants consolidate

The new channel does not inherit the hierarchy of the old one; it restarts, in large part, from zero. It also has its own dynamic: assistants tend to repeat the sources and names they already recognise, because trust accumulates. A company that establishes itself early as a citable reference on a topic becomes the default answer, and each citation reinforces the next.

The implication is uncomfortable for anyone comfortable on Google. An organic advantage built over years does not protect a position in AI answers, and the interval in which it is cheap to occupy empty space does not stay open forever. The asymmetry between Google and AI is closing, and those who enter early consolidate.

What remains open in Portugal

This barometer sizes a pattern; it does not close it. The next step, a multi-engine diagnostic with 7 assistants, is designed to address the open questions: which sectors concentrate the largest gap when several engines are measured rather than one, and whether the Portuguese buyer really starts in the assistant or still mostly on Google.

Three further questions remain open. What weight Copilot, via Microsoft 365, carries in supplier discovery across the Portuguese business fabric. Whether European Portuguese weighs in the answers, or whether the models respond mostly from sources in English and in Brazilian Portuguese. And how fast the window closes, that is, how quickly the first entrants consolidate their advantage.

How this barometer differs from our other study

This barometer is distinct from destaque.ai's study The State of AI Visibility for Portuguese B2B SaaS 2026, which covered 45 companies, 4 engines and a recommendation criterion. That study measures whether a company is recommended; this barometer crosses a company's Google strength with its AI visibility, defined as mention, across 251 companies and one engine. The numbers are not directly comparable: they are different cuts of the same problem, and each answers a different question about the same underlying shift in buyer behaviour.