AI Visibility Analysis: Belgium’s Flagship Companies

A Small Country, a Wide Gap: Belgium’s Best and Worst AI Scores Are Thirty-One Points Apart

Solvay. Umicore. Proximus Group. AB InBev. KBC. Ageas. Bekaert. BNP Paribas Fortis. Colruyt Group. Delhaize. Ten companies that between them define the Belgian economy: a global chemicals group, a materials-technology and battery-recycling leader, the national telecommunications incumbent, the world’s largest brewer, the country’s dominant bancassurance group, one of its largest insurers, a global steel-wire and industrial-technology group, the Belgian retail-banking arm of one of Europe’s biggest banks, and the two supermarket chains that anchor Belgian grocery retail. Between them they employ hundreds of thousands of people, generate tens of billions of euros in annual revenue, and represent the industrial, financial, and commercial backbone of one of the EU’s founding economies.

This is the second installment in this series’ country-level strand, following the Austria analysis published earlier this month. The same AI Visibility Inspector and Ivica Srncevic Framework used across that report, and across the ten industry-vertical installments that preceded it, was applied here.

Where the Austrian dataset produced near-uniform mediocrity, every one of ten companies clustering within a 16-point band and triggering a Structural Decay warning, the Belgian dataset tells a different story. It is a story of genuine separation: a leader at 66, a trailer at 35, and a 31-point spread between them that is nearly double what Austria produced. Belgium does not have a shared national ceiling. It has winners, losers, and a clear structural explanation for why each company landed where it did.

Methodology

Each company’s primary corporate website was evaluated using the AI Visibility Inspector across four structural dimensions:

• Structure – how content is architecturally organized for machine parsing, including H1 clarity and navigational coherence
• Depth – the substantive quality and retrievability of content as AI systems process and extract it
• Schema – the presence of structured data markup that enables confident entity identification and citation
• Freshness – whether content age signals are present and verifiable to AI retrieval systems

The overall AI Retrieval Index score runs from 0 to 100. Scores below 50 indicate significant structural invisibility. Scores between 50 and 74 represent fair to moderate visibility with material gaps. Scores at 75 and above indicate good to strong AI readiness.

A Structural Decay warning is triggered when critical signals are absent or conflicting: a missing H1 tag preventing AI parsers from anchoring a primary topic, multiple competing H1 tags fragmenting intent, or absent date signals leaving content age unverifiable.

The Scores

National average: 52.3 – Grade C/D border, AI Retrieval Index

Zero companies in Grade A. Zero in Grade B. Four in Grade C. Six in Grade D. Belgium’s best performer, Solvay at 66, would land comfortably mid-table in the Austrian comparison. Belgium’s worst performer, AB InBev at 35, is lower than any company recorded in that report.

Five Findings the Belgian Corporate Sector Needs to See

Finding 1: A Widening Spread, Not a Shared Ceiling

The 31-point gap between Solvay (66) and AB InBev (35) is almost double the 16-point spread recorded across Austria’s ten companies. That is the headline structural fact of this dataset. Austria showed a country where every company had made roughly the same partial investment in AI-readable infrastructure. Belgium shows a country where some companies have clearly made deliberate structural choices, Solvay and Umicore both cleared a Structure score of 100 with no Structural Decay warning at all, while others, AB InBev most starkly, have not addressed even the most basic machine-readability requirement of a functioning H1 tag.

This matters commercially. A spread this wide means AI visibility in Belgium is not a national condition to be waited out. It is a competitive variable that individual companies are already actively winning or losing.

Finding 2: Structural Decay Hit 7 of 10 – Split Across Three Distinct Failure Modes

Seven of the ten companies evaluated triggered a Structural Decay warning, lower than Austria’s 100% rate, but still the dominant pattern in the dataset. The failures split three ways:

Missing H1 tag (AI parsers cannot anchor a primary topic): Delhaize, KBC, BNP Paribas Fortis, and AB InBev, four of the country’s largest names in retail and finance.

Fragmented intent from multiple H1 tags: Ageas, flagged for three competing H1 tags on its homepage, one of Belgium’s largest insurers presenting AI systems with three simultaneous claims about what the page is primarily about.

Absent date signals: Bekaert and Colruyt Group, both of which otherwise posted the strongest Structure and Depth scores in the entire dataset, undone by the inability to verify content currency.

Three companies, Solvay, Umicore, and Proximus Group, triggered no Structural Decay warning whatsoever. That is a meaningful minority. Across the ten industry-vertical installments that preceded this series’ country-level reports, only Volkswagen achieved that distinction in the entire Automobile dataset. Three companies doing it in a single ten-company national sample is a genuinely positive signal, and it proves the fix is available to every other Belgian company in this report.

Finding 3: Freshness Is Polarized, Not Collapsed

Austria’s Freshness average collapsed to 1.6, a near-total, shared national failure. Belgium’s Freshness average sits at 21.9, more than ten times higher, but that number conceals a sharp split rather than describing a uniform improvement. KBC (63), Solvay (53), Umicore (53), and Proximus Group (46) all carry real, verifiable date signals. The remaining six companies, Colruyt Group, Delhaize, KBC’s own banking peer BNP Paribas Fortis, Ageas, and AB InBev, scored a flat zero.

This is a single path slits into two distinct branches, not a national blind spot, and that distinction matters. It suggests date signal implementation is not an unknown practice in Belgium; it is a known, low-cost fix that four communications or development teams have already made and six have not yet prioritized. Given how directly Freshness signals affect whether AI systems trust a company’s own site over a third-party source, that is six companies leaving a low-cost, high-impact fix untouched.

Finding 4: Schema Is the Deepest Hole in the Dataset – Even the Leader Falls Short

The average Schema score across the ten companies is 27.6, lower than Austria’s 37.0, and lower than nearly every sector average recorded across the ten industry verticals synthesized in this series’ 100-company dataset, where Schema averages clustered between 27 and 40. Belgium sits at the bottom of that range.

The highest single Schema score in this dataset belongs to Delhaize at 50, and Delhaize still lands in Grade D at 51 overall, undone by a missing H1 tag and a Freshness score of zero. That result echoes a pattern this series has documented repeatedly: a strong score on one dimension does not offset structural collapse on the others.

At the opposite end, BNP Paribas Fortis records a Schema score of 10, the lowest in the dataset, for a retail-banking institution operating under some of the heaviest financial disclosure and regulatory requirements in Europe. It is a near-identical finding to Vienna Insurance Group’s Austrian result, the lowest Schema score in that report also belonged to a major regulated financial institution, a sector built on structured, licensable, machine-verifiable facts that, twice now in this series, has the weakest structured-data layer of any company evaluated.

Finding 5: Two Grocery Giants, Two Different Failure Modes, the Same Grade

Colruyt Group (54) and Delhaize (51) anchor Belgian grocery retail, and both land in Grade D, but they get there by nearly opposite routes. Colruyt Group has a perfect Structure score of 100 and a strong Depth score of 80, and is undone entirely by a Freshness score of zero and an absent-date-signals warning. Delhaize has a comparatively weak Structure score of 55, missing its H1 tag entirely, but posts the dataset’s highest Schema score at 50.

Neither company’s specific weakness is a resourcing problem, both clearly have functioning content and development teams capable of strong scores elsewhere. It is a sequencing problem: each optimized a different dimension and left another completely unaddressed, and the AI Retrieval Index treats partial investment and no investment as functionally the same outcome once one critical signal is missing.

What AI Actually Sees

The entity interpretation outputs make the practical stakes concrete.

AB InBev, missing an H1 tag entirely, returns no confident primary topic to an AI parser. For a company whose beer brands are recognized on every continent, this means an AI system asked to identify what AB InBev’s own corporate site is fundamentally about has no structural declaration to work from, and has to reconstruct identity from secondary text, the least reliable path to entity recognition in this framework.

Ageas presents the opposite failure. Three competing H1 tags give an AI parser three simultaneous, conflicting claims about the page’s primary topic. Rather than no anchor, it has too many, and the result for retrieval confidence is functionally identical: the system cannot resolve which claim to trust.

KBC is the dataset’s most instructive case. Its Freshness score of 63 is the highest recorded here, real evidence that KBC’s content-management approach can verify currency. But KBC still lands in Grade D at 51, because a missing H1 tag caps its Structure score at 55 and its overall retrieval confidence with it. KBC has solved the cheaper problem and left the more consequential one untouched.

The Belgian Paradox

None of these ten companies are digital startups improvising their content strategy. They are mature institutions, in several cases centuries-old, operating under EU-level financial disclosure regimes, competing globally against peers that have already treated structured data as a competitive differentiator. Solvay and Umicore compete against chemicals and materials-technology companies worldwide that increasingly use Schema markup to win AI-mediated procurement research. BNP Paribas Fortis and KBC operate in a sector, banking, where structured, regulator-verified facts are the entire product. AB InBev is one of the most recognized consumer brands on earth, and its own corporate site currently gives AI systems no structural anchor for what the company is.

The gap documented here is not a capability constraint. Three companies in this same ten-company sample, Solvay, Umicore, and Proximus Group, prove the fix is achievable without heroic investment. It is a sequencing gap: human-facing communication has been prioritized for decades, and the machine-facing layer, increasingly the layer that determines whether AI systems can identify, date, and cite a company’s own claims about itself, has only been addressed by some of Belgium’s flagship companies, not all of them.

The commercial stakes are direct. When a procurement officer asks an AI assistant to compare European steel-wire suppliers, when an investor asks for a summary of Solvay’s current strategy, when a job seeker asks what Umicore actually does, or when a customer asks an AI system to compare Belgian banks, the answer is shaped by whichever source the system can retrieve with structural confidence. Right now, in six of these ten cases, that confidence is compromised by the company’s own website.

Key Takeaways

• No Belgian company in this dataset reached Grade B or above. Solvay leads at 66, still short of the 70-point threshold this series defines as genuine AI readiness.
• The 31-point spread between the highest score (Solvay, 66) and the lowest (AB InBev, 35) is nearly double Austria’s 16-point spread, indicating genuine competitive separation rather than a shared national pattern.
• Structural Decay affected 7 of 10 companies, split across missing H1 tags (Delhaize, KBC, BNP Paribas Fortis, AB InBev), fragmented intent from multiple H1 tags (Ageas), and absent date signals (Bekaert, Colruyt Group).
• Three companies, Solvay, Umicore, and Proximus Group, triggered no Structural Decay warning at all, a rare result in this series and proof the fix is available to every other company in this report.
• Freshness is polarized, not collapsed, averaging 21.9 nationally but ranging from 63 (KBC) to a flat 0 across six companies.
• Schema averaged just 27.6, the weakest dimension in the dataset. BNP Paribas Fortis recorded the lowest score at 10, echoing Vienna Insurance Group’s result in the Austrian report: a heavily regulated financial institution with the weakest structured-data layer in its own national sample.
• AB InBev recorded the lowest overall score at 35 (Grade D), driven by a missing H1 tag and a Structure score of just 35, the weakest Structure score recorded anywhere in this series’ country-level reports to date.

Belgium’s flagship companies split cleanly into two groups: those that have made the specific, low-cost structural decisions, an H1 tag, dateModified JSON-LD, baseline Schema markup, that determine AI retrievability, and those that haven’t touched any of it yet. The first group is proof the fix works. The second group is still deciding whether AI-mediated discovery is worth addressing before their competitors close the gap.

Want to know where your own company stands? If you’d like a free AI visibility check, similar to the ones behind this report, get in touch and I’ll run your site through the same framework and send you the results.

Research Date: July 2026 | Methodology: Ivica Srncevic Framework + AI Visibility Inspector. This research is independent, not sponsored by any organization or legal entity. All company names and logos are used for identification and analysis purposes only.