WHY: Growth Now Depends on Becoming Legible, Verifiable, and Recommendable to AI Systems
The most important strategic shift happening today is not the appearance of AI summaries at the top of search results, nor the rapid rise of conversational interfaces, nor the arrival of autonomous agents able to plan itineraries or select products. The real shift is more fundamental: AI systems have become the intermediaries that decide which brands enter a customer’s field of view.
To clarify the landscape, we must begin with a precise distinction. When an AI model answers a question—whether in Google’s AI Overviews, in Gemini, in Perplexity, or in ChatGPT—it ingests a wide range of sources: dozens of URLs, structured data from your website, user reviews, PR coverage, Reddit conversations, social media discussions, entity relationships it has learned from previous training, multimodal signals from your images, product catalogues, and even error-corrected context from forums.
The internal reasoning space is wide. Wide enough that your brand may very well be part of the input, even if the user never sees you.
But the output space is narrow. What the user sees—what the model chooses to surface—is a small, highly curated slice of that large internal evaluation. This is the “compression of visibility” that matters for growth. It is not that the AI considers only three florists or five resorts. It may evaluate dozens. But it will commonly show only one synthesised narrative and then a short, highly selective subset of examples—sometimes three, sometimes two, sometimes one, sometimes none at all.
Think of it like a hiring manager reviewing 50 résumés but only interviewing 3 candidates. The manager didn’t compress all 50 into 3—they evaluated 50 and selected 3 based on interpretable signals. The 47 rejected candidates were considered but found wanting or ambiguous.
Your brand may be evaluated but never surfaced. Only surfaced brands belong to the new buying journey.
This compression radically alters competition, because surfacing is not based on classic relevance signals alone. AI systems elevate brands they can interpret cleanly, verify confidently, contextualise precisely, and corroborate across multiple ecosystems—not only on their own websites.
This is where PR, Digital PR, authentic community discussions, expert citations, and coherent omnichannel narratives become central to AI-mediated discovery. Not as “link building” or “buzz,” but as inputs to the machine-readable identity your brand projects across the digital ecosystem.
A resort in Sardinia may have outstanding content on its own website, but if its attributes are ambiguous, if PR says one thing while OTAs describe another, or if social chatter contradicts the official positioning, AI models will see that brand as fuzzy, undecidable, hard to place. And in a compressive environment, what is vague is excluded.
Conversely, a small florist in Barcelona with consistent messaging, a handful of strong media mentions, real conversations on local Reddit threads, and clear product attributes establishes reliable patterns in how models represent that brand—even with fewer resources and modest traffic.
This layer of reasoning is invisible but commercially decisive. A traveller might never see your “Quiet Luxury Costa Rei Resort” page. They may never click a link. Yet the AI’s capacity to position your resort inside its internal category map—shallow-water entry, low evening noise, June-friendly wind patterns, short distance to walking trails, adults-only—determines whether you appear in a shortlist of two or three surfaced alternatives.
This shift creates an uncomfortable truth: Short-term tactics that inflate artificial mentions, synthetic links, or spammy citations risk creating semantic inconsistencies that make your brand harder for AI to interpret confidently. They generate noise in the associations AI forms around your brand, dilute entity clarity, and can introduce patterns of unreliability. In compressed surfaces, anything that reduces interpretability results in omission, and omission is equivalent to market loss.
The brands that thrive will be the ones that can be understood accurately, corroborated across channels, discussed authentically by communities, and recommended confidently by agentic systems.
Which brings us to the operational levers.
HOW: The Strategic Levers for Becoming a Brand AI Systems Can Trust and Recommend
To compete in an economy mediated by AI, CMOs must orchestrate six foundational levers. These levers overlap deeply because AI does not separate “SEO,” “PR,” “content,” and “product messaging” the way organisations do. AI views all these as one integrated representation of the brand.
Technical SEO or The Infrastructure of Machine Accessibility
Before a brand can be recommended, it must first be parsable. Technical SEO is no longer simply about making Googlebot’s life easier or improving page performance. It is about ensuring that every element of your brand — products, services, amenities, constraints, delivery rules, geographies, reviews — is accessible, unambiguous, consistent, and free of contradictions.
A Sardinian travel brand with outdated canonical rules or inconsistent structured data may accidentally tell the model that its rooms sleep four in one context and only two in another, that it is family-friendly in one part of the site and adults-only in another. AI systems are ruthless with contradiction. Content that contradicts itself is deemphasised, no matter how visible, beautiful, or persuasive it may be to humans.
Prioritisation note: Start here. If your technical foundation is broken, nothing else matters. Expect 2-4 months for comprehensive technical remediation on a mid-sized site.
Semantic Search or Making Your Value Proposition Explicit
AI cannot guess what your resort or product is about; it must infer it from your signals. Semantic optimisation means expressing attributes, relationships, contexts, and use-cases clearly enough that the system can locate you precisely inside the patterns it has learned about your category.
In travel, “quiet luxury” is not semantic clarity. But “low nighttime noise levels,” “shallow-water beach entry suitable for toddlers,” “prevailing low-wind conditions in June,” “private cabanas at 40m from shoreline,” “adult-only zones,” and “walking distance to Oasi di Biderosa” are machine-usable attributes that AI can reason with.
AI needs specificity, not slogans.
Prioritisation note: High ROI for low effort. Audit your top 20 product/service pages and add 5-10 explicit, factual attributes to each. This can be done in 4-6 weeks.
Agentic Search or Being Eligible for Machine-Driven Decisions
AI agents — whether in Gemini, Perplexity Agents, or custom GPTs — will increasingly perform combinations, evaluations, and recommendations on behalf of users. Their choices are driven by verifiable facts, structured attributes, constraints, costs, and objectively expressed qualities.
If a resort does not specify distance to hiking trails, the agent cannot place it in “outdoor-friendly honeymoon destinations.” If a florist does not explicitly encode delivery windows, the agent cannot include it in “deliver before 5 pm today.”
Agentic systems eliminate all options that cannot be interpreted with confidence. Ambiguous brands don’t lose because they are inferior; they lose because they are undecidable.
Prioritisation note: Medium-term investment. Begin by mapping the 10-15 decision-making attributes that matter most in your category, then systematically express them in schema markup and structured content.
Multimodal Optimisation or Being Discoverable Through Images and Context
AI is increasingly multimodal, meaning visual signals will progressively matter alongside textual ones. While most commercial queries today remain text-first, the trajectory is clear: more users will search with images, ask “find a place like this,” or expect AI to match visual style and context.
If your images lack structured captions, metadata, scene descriptions, and consistent photography patterns, AI systems cannot map them to real-world attributes. A user might upload a picture of a quiet cove. Two resorts may both be physically close. But only the one whose images contain the right multimodal descriptors will be matched effectively.
Multimodality is not aesthetic. It is strategic.
Prioritisation note: Forward-looking investment. Begin with your top 50 images: add structured alt text, EXIF metadata, and descriptive captions that connect visuals to your semantic taxonomy.
Brand Amplification or PR, Digital PR, and Authentic UGC as AI Trust Signals
Classic PR and Digital PR have reemerged as foundational levers, not for link-building, but as brand verification infrastructure.
AI models recognise authoritative news outlets, interviews, expert quotes, and in-depth analyses as strong corroborative signals. They interpret consistency across media as evidence of stability and reliability. They treat genuine conversations on Reddit, TripAdvisor Q&A, niche travel forums, or local community boards as real-world validation, far superior to manipulated links or fabricated mentions.
Healthy public discourse strengthens the machine-readable identity of your brand. Manipulative or inauthentic signals can introduce inconsistencies that confuse the patterns AI forms around you.
A resort mentioned organically in a niche Sardinia travel forum because users had meaningful experiences generates a level of authenticity impossible to replicate synthetically. A Barcelona florist praised in a local Reddit thread for reliable same-day delivery becomes a natural candidate for surfaced recommendations.
In an age where AI evaluates dozens of sources but surfaces only a handful of options, authenticity becomes the most efficient amplification mechanism.
Prioritisation note: Highest long-term ROI but slowest to materialise. Invest in 4-6 high-quality, authoritative placements per year rather than 50 low-quality mentions. Encourage genuine customer participation in relevant communities.
Omnichannel Semantic Governance or Ensuring All Channels Reinforce a Unified Reality
AI does not compartmentalise your website, your PR, your CRM, your social presence, and your product catalogue. It synthesises them.
If your PR says one thing, your website another, and your user reviews a third, AI interprets this as semantic instability. Brands need a unified semantic layer, aka a shared taxonomy, consistent attributes, and controlled messaging across channels, to ensure machines form a stable, trustworthy representation of who you are.
A brand that speaks with one voice is easier to surface. A brand that speaks with many contradictory voices is easier to omit.
Prioritisation note: This is ongoing governance, not a project. Assign a cross-functional owner (often a senior content strategist or product marketing lead) to maintain an attribute dictionary and enforce consistency quarterly.
WHAT — The Program for Becoming AI-Recommendable
Minimum Viable Approach (3-6 months, limited budget)
If resources are constrained, focus on these three actions in sequence:
- Fix technical contradictions (Month 1-2): Audit and unify your structured data, eliminate attribute conflicts, and ensure schema markup is consistent across all pages.
- Add explicit attributes (Month 2-3): For your top 20 pages, add 5-10 specific, factual, machine-readable attributes that answer “who, what, where, when, how much, how long.”
- Secure 2-3 authoritative mentions (Month 3-6): Invest in getting covered by the most credible publications in your category: depth over breadth.
This approach delivers 60-70% of the value at 20% of the cost.
Comprehensive Approach (12-18 months, full investment)
Phase One: Technical and Semantic Foundations (Months 1-4)
Begin by clearing the forest floor. Fix crawlability, unify attributes, clean contradictions, and introduce robust structured data. Rebuild key pages to express meaning explicitly. Establish an attribute schema and apply it consistently across all products or services.
Phase Two: Prepare for Agentic Discovery (Months 4-8)
Identify all decision-making attributes that matter in your category, and express them in machine-readable formats. Build an internal brand knowledge graph—a documented map of how your products, services, attributes, and constraints relate to each other. Create content assets that are easy for models to digest into “reasoning components”: guarantees, constraints, services, pricing logic.
Phase Three: Optimise Multimodal Assets (Months 6-10)
Audit your existing imagery and video. Add structured captions. Standardise metadata. Produce new assets designed for machine vision. Connect visuals to your semantic taxonomy so that the multimodal understanding of your brand matches the textual understanding.
Phase Four: Build Trust Signals Through PR + Genuine Community Presence (Months 8-15)
Invest in authoritative, meaningful media coverage; not volume, but relevance and authority. Avoid poor-quality link-building disguised as Digital PR. Encourage and support positive discourse in authentic communities. Ensure your leaders and experts appear in credible contexts.
This work enhances how AI systems represent your brand internally, providing the stability needed to be consistently surfaced.
Phase Five: Govern Meaning Across Channels (Ongoing from Month 12)
Align PR, product, SEO, content, and CRM teams around shared terminology and attributes. Enforce semantic consistency through quarterly audits. Monitor how AI systems interpret your brand and correct misalignments as they emerge.
METRICS — Measuring Success When Attribution is Fragmented or Invisible
The difficulty with AI-mediated journeys is that users often do not click anything before making a decision. The influence of your brand is real but invisible in analytics. We therefore need a new measurement framework.
Upstream (machine legibility):
- Structured data coverage: % of pages with complete, consistent schema markup
- Attribute completeness: % of products/services with 10+ explicit attributes defined
- Cross-channel consistency score: Manual quarterly audits comparing how your brand is described across website, PR, reviews, and social
Midstream (AI visibility):
- Track the overall visibility evolution of your brand and in its relationship to a topic in AI Models using tools like WAIKay.io.
- Mention frequency: Track how often your brand appears in AI-generated answers for your core category queries (sample 20-30 queries monthly across ChatGPT, Perplexity, Google AI Overviews, Gemini)
- Representation accuracy: When mentioned, is the AI’s description of your brand correct?
Competitive presence: Are you included when AI compares options in your category? - Analyse server log file to review and qualify the accesses of AI Bot (training, search…) to your website: what pages they access, which not, are there sections ignored by them?
Warning: Midstream measurement is labour-intensive and requires manual sampling. Tooling for this is still immature, albeit improving. Budget 10-15 hours per month for this analysis.
Downstream (commercial impact):
- Branded search lift: Increase in searches for your brand name (implies AI introduced you to new audiences)
- Direct traffic uplift: Growth in direct visits (implies users saw you in AI, then came directly)
- Conversion rate improvement: Pages with rich semantic optimisation should convert better when users arrive
- Category share growth: Overall market penetration in your category
Traditional attribution models break because they assume human-driven browsing. AI breaks that assumption. The correct mental model is not deterministic attribution but probabilistic influence. The more interpretable, verifiable, consistent, and corroborated your brand is, the more frequently AI systems choose to surface it.
Being surfaced is now the commercial bottleneck.
Final Message for CMOs
Growth in the AI era is not about gaming algorithms or producing more content. It is about building a brand that machines can reliably interpret: cohesive, structured, authentic, and corroborated across many ecosystems.
A brand that speaks with one voice, expresses meaning precisely, appears in trustworthy media, participates authentically in communities, maintains clean technical infrastructure, provides verifiable facts, and aligns its messaging across channels will be consistently surfaced, recommended, and chosen, even before the user begins their conscious journey.
This is not “the future of SEO.” This is the new economics of visibility.
What Happens When Everyone Does This?
If every brand optimises for AI interpretability, the competitive advantage shifts to:
- Depth of authenticity: Genuine community advocacy becomes harder to replicate than technical optimisation
- Speed of governance: Brands that can maintain semantic consistency as they scale will pull ahead
- Quality of product/service: When AI can accurately represent all competitors, the experience becomes the differentiator again, which is how it should be
The paradox is that optimising for AI ultimately returns competition to fundamentals: be excellent, be consistent, be authentic, and make sure machines can understand why.
