Legacy Skin-Care Brands Rebuild for Generative Search: How Borghese and RoC Are Rewriting Product Pages, Data and Strategy for AI Discovery

Key Highlights
Introduction
Why generative search changes product discovery
Borghese’s Project PDP: rebuilding product content from the formula up
RoC’s three-pronged AI strategy: generative visibility, content optimization, and community
Where generative models pull information—and how that changes content priorities
Practical steps brands should take now
Why retailer pages and legacy editorial matter more than you might expect
Governance and guardrails: controlling accuracy without stifling speed
How auditing formulations with AI can reshape marketing stories
Organizational implications: how teams should realign
Measuring ROI: what success looks like for GEO efforts
Common pitfalls and how to avoid them
Real-world examples beyond Borghese and RoC
Practical checklist: launching a GEO program in 90 days
The legal and regulatory boundary: what marketers must not forget
Where this leads: a new discipline of product-centered search engineering
FAQ

Key Highlights

Borghese and RoC are overhauling product data, copy and distribution channels to improve visibility on generative AI search; efforts include Project PDP, long-form content and seeding verified information on platforms that feed AI models.
Brands are prioritizing product detail pages (PDPs), retailer content and earned media as high-impact sources for generative engine optimization (GEO), while instituting AI guardrails, human oversight and performance KPIs.
Practical tactics include extensive A/B testing of PDP elements, adding structured Q&A to reflect conversational search queries, cleansing ingredient and scientific data, and building authoritative signals across the open web.

Introduction

Search is shifting from lists of blue links to narrative answers and synthesized recommendations generated by large models. That change is altering how consumers discover personal-care products and forcing established brands to rethink fundamentals that once produced reliable SEO outcomes. At a recent industry forum, executives from two legacy skin-care companies—Borghese and RoC—described a clear redesign of their content, data and measurement strategies to earn visibility in AI-driven search results.

Their approaches converge on a simple premise: generative engines surface information differently than traditional search. The platforms prioritize authoritative, well-structured source material that maps to conversational queries. Brands that want to remain discoverable must control the facts about their products across owned pages, retailer listings and the broader publishing ecosystem. That requires product teams, regulatory affairs, marketing and commerce to collaborate on data hygiene, long-form narratives and distribution tactics—and to build procedures that ensure accuracy and human review of AI-generated or AI-optimized content.

The following report synthesizes what Borghese and RoC revealed publicly, places their strategies in practical context, and offers a roadmap for product-led brands navigating generative engine optimization (GEO).

Why generative search changes product discovery

Search engines that return synthesized answers behave like aggregators and narrators rather than link directories. Rather than sending users to a list of pages where they might find an answer, generative systems select information from multiple sources, weigh authority signals, and produce prose that answers a user’s question directly. For product-oriented searches—“best face mud treatment for dry skin,” “anti-aging serum with retinol and niacinamide,” or “how long does sunscreen protect skin?”—the output may include recommendations, product comparisons, usage instructions, and citations to the sources that informed the answer.

That behavior shifts the optimization challenge in several ways:

Discoverability depends less on keyword-stuffed pages and more on verifiable, well-structured facts that align with conversational questions.
Sources beyond brand dot-com sites—retailer PDPs, product reviews, digital editions of legacy publications—can disproportionately influence what generative models surface.
Answer quality hinges on accuracy. Misinformation or vague claims can be amplified by models and then distributed widely.
The unit of optimization becomes the product entity and its supporting knowledge graph: ingredients, clinical claims, formulations, clinical endpoints and provenance all matter.

Executives at Borghese and RoC framed GEO as an operational issue, not only an SEO problem. When AI substitutes narrative answers for link lists, product teams must redesign how product facts are authored, approved and syndicated.

Borghese’s Project PDP: rebuilding product content from the formula up

Borghese confronted the problem directly when its flagship Fango mud treatment failed to appear in AI search results. “It took us 14 times to find it on search, on AI,” Dawn Hilarczyk, Borghese’s chief operating officer, recounted. The struggle exposed a gap between the brand’s product heritage and how generative systems index and surface product information.

Project PDP is Borghese’s response: a full-company initiative to rebuild product pages, overhaul copy, and scrub formulations for factual clarity. The aim is to produce the kind of factual building blocks that generative models and downstream platforms will pull into their responses.

Key elements of Project PDP:

Data hygiene and ingredient clarity: Borghese is “scrubbing all of our formulas to get really factual information,” extracting scientific benefits and standardizing ingredient nomenclature so that models recognize the product as a distinct entity with verifiable properties.
Long-form content: The brand is augmenting product pages with extended narratives that explain formulation science, use cases, and evidence—content better suited to answer complex, conversational queries.
Cross-platform seeding: Borghese approved copy for a dedicated Wikipedia page for Fango to seed an authoritative, public source that many knowledge layers and models consult. The brand recognizes that “AI is only as good as the information you pull from it,” and that absence from the conversation leaves the narrative to be defined by others.
Product architecture and metadata: Product content is being reorganized to include structured fields that map to schema and entity-knowledge graphs—ingredient lists, INCI names, clinical study notes, SKU identifiers and precise usage instructions.

Borghese’s discovery during the initiative provides an operational lesson: brands may already possess product benefits that are hidden by poor data or inadequate storytelling. After reanalyzing formulations with AI assistance, Borghese found “our formulas had great benefits that helped increase longevity in the skin,” benefits they had not been telling customers about.

This illustrates two important points. First, product discovery depends on accurate, granular claims. Second, internal R&D and regulatory teams are crucial partners in converting laboratory insights into consumer-facing, model-friendly narratives.

RoC’s three-pronged AI strategy: generative visibility, content optimization, and community

RoC framed their approach as a pragmatic, performance-led program covering three areas: generative search visibility, content optimization and community engagement. Hillary Hutcheson, RoC’s CMO, cited consumer behavior as the immediate driver: “We see that over 40% of consumers are now discovering new products through Gen AI search.”

RoC’s specific tactics:

Prioritize high-value sources: Using AI simulation tools from a platform that models generative search, RoC identified which sources most strongly influence model outputs. They discovered that retailer product pages and digital archives of legacy magazines are significant drivers of generative search traffic.
Optimize PDPs with A/B testing: RoC deploys partners that enable A/B testing of PDP elements—from image treatments to media formats—to determine what variants most reliably map to the questions posed by AI searchers. This includes testing hero imagery, ingredient callouts, benefit-focused headers and video snippets.
Embed Q&A at scale: The brand has integrated more than 400 Q&A entries into product pages to reflect conversational forms of user queries. These Q&A sets map natural language queries (e.g., “Will this product cause irritation with retinol?”) to precise, evidence-backed answers that generative models are more likely to pick up.
Double down on earned media: The surprising influence of magazine archives—Vogue, Allure—prompted RoC to shift investment away from paid placements toward earned stories and coverage that create durable, high-authority references on the open web.

RoC anchors the program in performance metrics. “Everything that we do around AI needs to drive team efficiency and serve the consumer,” Hutcheson emphasized. Tools and tactics should free team capacity rather than create new manual burdens.

Where generative models pull information—and how that changes content priorities

Generative models infer answers by drawing on a corpus of web data, structured knowledge sources and proprietary content. The exact composition varies by model and deployment, but common high-impact sources include:

Brand-owned product pages and help centers
Retailer PDPs and stock-keeping metadata
Editorial articles and archived magazine content
Community Q&A and forum posts with high visibility
Wikipedia and other reference resources
Schema markup and public knowledge graphs with strong entity links

Two practical implications follow:

Presence across multiple, credible outlets increases the chance of being referenced. A brand’s single-page story is less influential than consistent, corroborated signals appearing across retailer listings, editorial mentions and public reference pages.
Source authority depends on durability and verifiability. A dated press release or a nonstandard ingredient list can be deprioritized by ranking algorithms. Publications with editorial controls and persistent archives—magazine back catalogs, Wikipedia entries, and retailer pages with consistent SKUs—carry weight.

RoC’s use of an AI search simulation platform illustrates the point. By modeling which sources feed generative answers, they discovered non-obvious contributors—older magazine pieces that now live online—and shifted tactics to secure earned mentions rather than only paying for exposure.

Practical steps brands should take now

The initiatives of Borghese and RoC translate into practical actions other product brands can implement immediately. These steps are organized across content, data, distribution and governance.

Content and copy

Build product pages that answer conversational queries. Convert technical benefits to plain-language explanations and pair them with citation links or source notes.
Create scalable Q&A modules that reflect how consumers ask questions in chat and voice search formats. Use customer service transcripts and search logs to seed question lists.
Supplement short PDP copy with long-form content that explains formulation science, testing protocols and usage contexts. Long-form sections increase the likelihood that models extract substantive information.
Maintain a single source of truth for scientific claims and citations. When referencing clinical data, cite study design, endpoints, and limitations.

Data and metadata

Standardize ingredient names using accepted nomenclature (e.g., INCI for cosmetics) and include active concentrations when allowed by regulation.
Add structured data (schema.org/Product, ProductModel, AggregateRating, Offer, etc.) to product pages to create clearer entity signals for crawlers and knowledge layers.
Include SKU, GTIN, and other unique identifiers consistently across brand and retailer pages to make product entities discoverable and unambiguous.
Cleanse legacy archives to remove contradictory or outdated claims that could confuse automated synthesis.

Distribution and earned presence

Audit retailer PDPs and prioritize fixing discrepancies; models often rely on retailer listing content when crafting answers.
Seed authoritative third-party resources—press mentions, brand-authored Wikipedia entries, and editorial features—that can be crawled and cited by knowledge graphs.
Pursue earned media placements in publishers with durable digital archives, not just transient programmatic placements.
Encourage verified product reviews and expert commentary that add corroborated observations for model training corpora.

Measurement and experimentation

Use generative search simulation tools to prioritize targets and test which content variants influence answers and traffic.
Run A/B tests on PDP elements and measure downstream KPIs—click-through from AI answers to cart, conversion, and search-sourced discovery rates.
Track efficiency metrics: time saved in content creation, reduction in consumer inquiry volume after better content, and decreases in claim dispute cycles due to improved data clarity.

Governance and human oversight

Require regulatory review for any AI-generated consumer-facing claims, and preserve a version control system for all product content.
Train teams to be “AI skeptics”: understand where models source data and be prepared to validate model outputs rather than accept them at face value.
Define KPIs that require AI tools to improve efficiency or consumer outcomes rather than generating more content without measurable impact.

These steps are practical, but execution demands cross-functional alignment. Marketing cannot act alone; product development, regulatory and commerce must coordinate to produce factual, model-ready content.

Why retailer pages and legacy editorial matter more than you might expect

One of the most actionable revelations from RoC’s work was the outsized role of retail PDPs and legacy magazine archives in shaping generative search outputs. Retailers host product-level metadata at scale and maintain content tied to SKUs, pricing and availability—data that is useful to models attempting to recommend options for purchase. Similarly, legacy editorial pieces often contain structured comparisons, expert quotes and technical breakdowns that sit behind durable URLs and editorial authority.

Practical consequences:

Brands should prioritize accuracy and completeness on retailer listings. Discrepancies across distribution partners create noise that models may interpret as conflicting evidence.
Earned coverage in reputable editorial outlets is valuable not only for traditional brand equity but also for feeding authoritative narrative elements into model training corpora.
Updating or augmenting retailer copy with richer Q&A and specs can improve how generative engines select and synthesize brand information.

RoC’s decision to build Q&A sets into PDPs acknowledges that retailer pages are high-leverage points. Those pages capture both the transactional user intent (to buy) and informational intent (to learn). As models aim to satisfy both, retailer pages that include robust, verified answers naturally become attractive sources for synthesized responses.

Governance and guardrails: controlling accuracy without stifling speed

Both Borghese and RoC stressed the need for internal guardrails. AI tools simplify certain tasks but also introduce new risks—hallucinated claims, outdated citations, or content that runs afoul of regulatory rules.

Guardrail practices that brands should institutionalize:

Approval workflows: Any AI-assisted or AI-generated content should pass through regulatory and legal review. Maintain clear sign-off paths and audit trails.
Source attribution: Track where an AI tool sourced its information. If a model cites a claim, be able to trace that claim back to the originating page or dataset.
Version control and rollback procedures: If AI-optimized content causes confusion, teams should be able to revert to the last verified state and diagnose the issue.
Education and skepticism: Train employees to question models and interpret outputs critically. Ask, “Where does this tool draw its evidence?” and “Is that evidence authoritative for regulated product claims?”
KPI governance: Mandate that AI tooling must reduce manual work or measurably improve consumer outcomes. Avoid proliferating tools that offer novelty but no productivity gain.

Hilarczyk’s admonition—“We need to talk about being AI skeptics and teach our teams to also be skeptics”—captures the cultural work required. Brands must build processes that combine speed and accuracy.

How auditing formulations with AI can reshape marketing stories

AI-assisted analysis of R&D data can produce marketing insights without creating new formulations. Borghese’s experience demonstrates how mining existing data reveals benefits that were present but untold. The brand’s analysis uncovered that certain formulations contained ingredients or structures associated with improved skin longevity—a benefit the marketing team had not emphasized.

This approach offers two advantages:

It leverages owned IP. Rather than inventing new formulations to chase trends, brands can extract latent value from existing products by surfacing scientific rationales in consumer-facing language.
It reduces regulatory friction. Claiming a benefit justified by internal analysis and properly vetted studies is more defensible than speculative marketing.

Brands should consider periodic audits of formulation and safety data. Use natural-language models to summarize clinical protocols, ingredient functions and potential use cases, then route the summaries through regulatory review before publication.

Organizational implications: how teams should realign

The shift toward GEO requires changes in roles and processes at both strategic and operational levels.

Marketing and e-commerce

Own coordinated PDP strategies and A/B testing frameworks across brand and retailer sites.
Maintain a living repository of Q&A items mapped to product attributes and consumer intents.
Manage relationships with editorial partners and publishers to secure enduring earned placements.

R&D and regulatory

Supply standardized scientific summaries and evidence citations for product claims.
Vet AI-produced summaries against raw study data to ensure accuracy.
Provide ingredient and clinical data in machine-readable formats.

Data and engineering

Implement schema.org/Product and other structured-data markups across pages.
Ensure consistency of identifiers (GTIN, SKU) across platforms and feeds.
Develop audit logs that record the provenance of any AI-suggested content.

Customer service and community

Feed common consumer questions and complaint patterns into Q&A content development.
Monitor conversational trends to prioritize new Q&A items and content fixes.

Leadership and governance

Define KPIs that reward efficiency, accuracy and consumer satisfaction.
Approve AI tooling and set guardrails on usage.

This alignment resembles an operational playbook: product truth is produced by R&D, translated by marketing, encoded by data teams, distributed via commerce partners and validated by earned media.

Measuring ROI: what success looks like for GEO efforts

Measuring the impact of GEO work requires both traditional and new metrics. Track both discovery (top-of-funnel) and downstream conversion.

Suggested KPI set:

Generation-driven discovery: proportion of new product discovery events attributed to generative search channels.
PDP engagement lift: changes in time on page, scroll depth, and interactions with Q&A components after content updates or A/B tests.
Conversion rate changes: purchase rates from traffic arriving through AI-sourced recommendations versus traditional search.
Content efficiency: time saved in content creation and update cycles after implementing AI-assisted workflows with human oversight.
Brand signal quality: decrease in conflicting claims across retailer feeds, and increase in authoritative references (e.g., number of reputable editorial or Wikipedia mentions).
Operational risk indicators: number of regulatory issues or consumer complaints attributable to AI-suggested copy.

RoC’s insistence that AI must “drive team efficiency and serve the consumer” underscores the need to avoid vanity metrics. Narrow objectives to measurable outcomes tied to business and legal risk.

Common pitfalls and how to avoid them

Brands experimenting with GEO encounter recurring hazards. Recognizing them early reduces downstream friction.

Pitfall: Relying solely on AI to generate claims

Remediation: Always route AI-generated claims through legal and regulatory review. Maintain documentation of source data.

Pitfall: Siloed updates across channels

Remediation: Centralize product truth in a single, version-controlled repository and propagate updates systematically to retailer feeds.

Pitfall: Prioritizing paid placements over earned credibility

Remediation: Balance short-term paid promotions with long-term earned and reference-building tactics. Earned content often has superior persistence in model corpora.

Pitfall: Overoptimizing for explicit keywords rather than conversational queries

Remediation: Use consumer dialogue data, service transcripts and voice search examples to shape Q&A and narrative content.

Pitfall: Missing unique identifiers and structured metadata

Remediation: Ensure GTIN/SKU consistency and complete schema markup so knowledge layers can disambiguate entities.

Pitfall: Treating generative search as an add-on to SEO

Remediation: Treat GEO as a cross-functional capability with technical, content, legal and commercial ownership.

The common thread is integration: product truth must be coherent across sources and verifiable, not a collection of disparate messages.

Real-world examples beyond Borghese and RoC

While the examples below are illustrative rather than exhaustive, they show common strategic patterns that brands are adopting.

Example 1: A legacy hair-care brand audits its product formulations and finds a long-used botanical extract with clinically documented hydration benefits. The brand publishes a long-form article detailing the extract’s mechanism and links it to ingredient specifics on PDPs and retailer pages. Within months, generative queries referencing “botanical extract for dry scalp” begin returning the brand’s page as an authoritative source.

Example 2: A mass-market sunscreen manufacturer builds 200 Q&A items into its PDPs, addressing consumer questions about SPF measurement, water resistance and layering with cosmetics. They partner with a retailer to ensure the same Q&A set appears on retailer pages. Generative engine simulations show improved likelihood of being cited in conversational sunscreen queries.

Example 3: A boutique brand commissions a neutral, high-quality editorial review in a respected digital magazine. The article includes clinical quotes and a product history. Because the piece resides on a site with long-term archival policies, generative models later cite it when asked about product lineage and efficacy.

Each example demonstrates the same point: authoritative, corroborated content placed in durable channels increases the odds that generative systems will surface a brand’s story.

Practical checklist: launching a GEO program in 90 days

Below is a concise but actionable 90-day checklist to jumpstart generative engine optimization. The roadmap presumes existing product data and a willingness to coordinate across functions.

Days 1–14: Discovery and prioritization

Inventory top SKUs and identify products with strategic importance.
Run generative search simulations for those SKUs to identify which sources feed current model outputs.
Compile a list of the top 20 external sources (retailer PDPs, editorial, Wikipedia pages) that influence answers.

Days 15–45: Data cleanup and content seeding

Standardize ingredient names and create machine-readable product dossiers for each SKU.
Build or augment Q&A sets (start with 20–50 priority questions per SKU).
Add schema.org/Product markup, SKU/GTIN, and standardized offers to brand pages.
Draft and publish long-form product narratives validated by regulatory.

Days 46–75: Distribution and experimentation

Coordinate updates with top retail partners to align PDP copy and Q&A.
Seed authoritative content on third-party pages: pitch earned editorial, update Wikipedia where appropriate and permissible, and secure expert commentary.
Begin A/B tests on PDP images, hero headers and Q&A phrasing. Monitor engagement metrics.

Days 76–90: Governance and measurement

Establish approval workflows for AI-assisted copy.
Set KPIs: generative discovery %, PDP engagement lift, conversion lift, and content efficiency measures.
Review initial outcomes and plan scale-up priorities for the next quarter.

The checklist compresses core activities into achievable sprints while embedding governance and measurement.

The legal and regulatory boundary: what marketers must not forget

Cosmetic and personal-care claims are subject to country-specific rules. Misstatements—especially those that imply drug-like therapeutic benefits—can trigger regulatory scrutiny. AI-generated recommendations increase the risk because models may rephrase or infer claims beyond what was authored.

Key precautions:

Avoid new claims that are not supported by documented studies and regulatory clearance.
Ensure clinical claims include context: study design, endpoints, population and limitations.
Regulatory teams should be part of any content-creation flow that uses AI; maintain audit logs that show the provenance of claims.

Borghese’s and RoC’s insistence on human oversight and regulatory review reflects the reality that brand protection still starts with accurate, defensible claims.

Where this leads: a new discipline of product-centered search engineering

Generative engine optimization requires a hybrid skill set: product expertise, content craft, data engineering and regulatory literacy. The work resembles search engineering but emphasizes product entities and factual clarity over keyword density.

Emerging capabilities that will define leading teams:

Knowledge engineering: codifying product facts into schemas, ontologies and knowledge graphs.
Conversational content design: writing copy that aligns with natural language queries and includes authoritative citations.
Simulation-driven prioritization: using model simulators to identify high-impact channels and content types.
Evidence operations: systems to gather, vet and publish clinical and safety evidence in machine-readable formats.

Teams that build these capabilities will convert product truth into discoverable narrative assets that generative systems are more likely to surface.

FAQ

Q: What exactly is generative engine optimization (GEO)? A: GEO refers to the set of activities brands undertake to ensure that generative search systems—and other AI-driven answer engines—use accurate, authoritative brand and product information when synthesizing answers. It focuses on creating machine-readable facts, long-form explanatory content, scalable Q&A, and distribution strategies that seed authoritative signals across the web.

Q: How is GEO different from traditional SEO? A: Traditional SEO optimizes for keyword-driven page rankings and click-throughs from search engine results pages. GEO optimizes for being included in synthesized answers. That requires authoritative, structured product facts, corroborated signals across retailers and editorial outlets, and content written to match conversational queries. GEO still benefits from technical SEO fundamentals, but it changes the priority of content types and channels.

Q: Will adding more content always improve visibility in generative search? A: Quantity alone is insufficient. AI models weigh authority, verifiability and coherence. High-volume, low-quality content can create noise and confuse models. Focus on accurate, well-structured content, and corroborate claims across multiple credible sources.

Q: Should brands create Wikipedia pages for products? A: A well-sourced Wikipedia entry can increase the likelihood of being referenced by knowledge layers. However, Wikipedia has notability and sourcing standards; entries must be neutral and backed by reliable, third-party sources. Brands should not write promotional pages disguised as encyclopedia entries; instead, they should seek neutral coverage in reputable publications that can serve as independent references.

Q: How important are retailer product pages for GEO? A: Very important. Retailer PDPs often contain SKU-level metadata and are a frequent source for models seeking transactional and product detail information. Ensuring accuracy and completeness on retailer pages can materially influence generative answers.

Q: Can AI tools help detect product benefits in R&D data? A: Yes. Natural-language techniques can summarize clinical documents, safety reports and formulation notes to surface potential benefits. Those summaries must be validated by regulatory and R&D teams before being published as consumer-facing claims.

Q: How should organizations govern AI-assisted content creation? A: Require human review, maintain traceability of sources, set KPIs that emphasize efficiency and consumer outcomes, and include legal/regulatory sign-off steps. Train teams to question model outputs and to verify claims.

Q: What KPIs should brands track for GEO programs? A: Track generative-search discovery rates, PDP engagement metrics, conversion rates for traffic originating from AI-sourced recommendations, content efficiency (time saved), and brand-signal quality (consistency of claims and number of authoritative references).

Q: Are there tools that simulate what generative search will do? A: Platforms exist that model how generative systems might synthesize answers and which sources they rely upon. These tools help prioritize content updates by showing which pages and publications most influence final outputs. Simulation is an effective way to test hypotheses without waiting for real-world model indexing cycles.

Q: What mistakes should be avoided when pursuing GEO? A: Avoid relying solely on AI to generate claims, creating disjointed messages across channels, neglecting retailer page accuracy, and pursuing paid placements at the expense of durable, earned references. Also avoid over-optimizing for keywords instead of matching conversational intent and evidence-based claims.

Q: How should small brands with limited resources start? A: Prioritize a small number of high-value SKUs. Cleanse ingredient lists and add Q&A to PDPs that answer the top customer questions. Ensure at least one reliable third-party reference (editorial or expert commentary) exists for each prioritized product. Use simulation tools selectively to validate where updates will matter most.

Q: Will generative search favor larger brands with more resources? A: Larger brands with established editorial coverage and broad retailer distribution may have an advantage because they produce many corroborated signals. However, smaller brands can compete by creating highly accurate, well-sourced content and pursuing targeted earned placements that demonstrate authority for specific product niches.

Q: How fast do results appear after implementing GEO changes? A: It depends on how often models retrain or update their knowledge and how quickly crawlers recrawl corrected content. Some effects can be seen in weeks if an influential source is updated; systemic changes across multiple publishers often take longer.

Q: What governance questions should executives ask before approving GEO budgets? A: Ask how the program will improve consumer discovery and conversion, what guardrails are in place for regulatory and legal risk, how content provenance will be tracked, and what KPIs will prove efficiency gains. Ensure cross-functional ownership among marketing, R&D, data and legal teams.

Q: Can generative search replace customer service content? A: Generative systems can field many common questions, but they should augment—not replace—official customer service and regulatory-verified guidance. Brands must ensure that any synthesized guidance aligns with approved claims, safety recommendations and usage instructions.

Q: Where should brands invest first: content, technical SEO, or earned media? A: Invest in the highest-leverage combination for your product mix. For most product brands, the priority order is: 1) data and content hygiene on PDPs (including Q&A and schema), 2) alignment of retailer pages, and 3) earned editorial coverage to build authoritative third-party references.

Q: How will GEO evolve over the next 12–24 months? A: Expect generative systems to become more sensitive to structured data and cross-source corroboration. Models will increasingly prioritize sources with durable editorial control and clear provenance. Brands that build robust, verifiable product entities and distribute them across authoritative channels will gain incremental discovery advantages.

Generative search is changing the mechanics of discovery and demanding a disciplined, product-centered response. Borghese’s Project PDP and RoC’s three-pronged program illustrate how legacy brands can turn product truth into discoverable narrative assets: clean data, long-form evidence, scalable Q&A, and distribution into durable, authoritative channels. The work requires cross-functional cooperation, careful governance, and an emphasis on measurable outcomes—an operational shift that rewrites the playbook for product marketing in an age where answers matter as much as links.