Contact

AI Search - when it works brilliantly, and when it struggles to perform

Is Search AI worth your investment? Examples of where it works well, and situations where it struggles.

AI search is changing how buyers discover, compare, and trust brands. Before you invest, use this quick, honest self-assessment to see whether a GEO + authority approach will pay off—for your category, resources, and goals.

Where this work lands best (high-fit)

  • Info-rich businesses with real proof assets
    e.g., B2B SaaS with docs/use-cases, clinics with treatment pages and outcomes, multi-location services, universities, professional firms. They already have testimonials, case studies, PDFs, webinars, data sheets—gold for entity building.

  • Brands with a defined positioning & ICP
    Clear “who we help / what’s uniquely true about us” lets us craft unambiguous entities and FAQs that answer engines can cite confidently.

  • Operational capacity to publish + maintain
    A small but consistent content cadence (subject-matter interviews → structured FAQs → references) and someone who can ship schema and page updates.

  • Tidy source-of-truth data
    Clean NAP, product/service taxonomies, bios, credentials, opening hours, locations—ideally in a CMS or sheet we can map to schema.org.

  • Regulated or credibility-sensitive categories
    Law, finance, healthcare, education, engineering. The need for verifiable claims, citations, and governance aligns perfectly with GEO’s emphasis on E-E-A-T.

  • Clear “change moments”
    Rebrand, site migration, product launch, market expansion, or compliance update—orgs are open to tidying data and processes right then.

Where it struggles (low-fit or likely to stall)

  • Thin or me-too businesses
    Dropship e-com with vendor descriptions, affiliates with no unique insight, local sites with boilerplate service pages—nothing new to cite.

  • No proof, no references
    No case studies, no expert quotes, no third-party mentions, no satisfied customers willing to review—answer engines can’t verify you.

  • Expectation of quick wins without ops
    Leadership wants “rank #1 in 30 days” but won’t free SMEs for interviews, won’t approve content, and won’t fix data inconsistencies.

  • Legal/bureaucratic deadlock
    Heavily centralised sign-off that takes months per page—momentum dies, coverage never reaches critical mass.

  • Fragmented local data
    Franchises/locations with inconsistent names, numbers, and hours; store finders out of sync with GMB—entity confusion persists.

  • Tool-chasing > truth-building
    Buying more AI tools instead of doing the unglamorous work: clarifying entities, standardising data, earning citations.

Classic failure modes (and what to do instead)

  • Over-indexing on markup alone → Great schema on weak content still underperforms.
    Fix: Pair schema with authoritative copy, references, and on-site proof.

  • Ignoring off-site signals → No reviews, no mentions, no citations.
    Fix: Lightweight PR, partner mentions, directory hygiene, talk-trigger review ops.

  • Unclear entity graph → Products, people, services not disambiguated.
    Fix: Maintain an entity map (Who/What/Where/IDs) and mirror it in schema + internal links.

  • One-and-done projects → Data drifts, pages decay.
    Fix: Quarterly data hygiene + FAQ refresh, review velocity targets, link reclamation.

  • KPI mismatch → Chasing traffic, not qualified demand.
    Fix: Track answer presence, branded uplift, lead quality, and “share of answers,” not just visits.

Great-fit sectors & scenarios (quick hits)

  • Multi-location healthcare/clinics: structured practitioners, treatments, FAQs, outcome frameworks, insurer affiliations.

  • B2B SaaS: docs, integrations, security/compliance pages, customer stories → superb for GEO snippets and comparisons.

  • Professional services (legal, accounting, architecture): credentials, case matter types, jurisdictional FAQs, citations to standards.

  • Education & training: course metadata, instructor bios, syllabi, accreditation.

  • Local services: NAP consistency, service area pages, job galleries, reviews.

Fast fit-score (8 levers, rate 1–5 each)

  • Strategy clarity (positioning, ICP)

  • Proof depth (case studies, reviews, third-party mentions)

  • Data readiness (clean, centralised facts)

  • Content ops (capacity to publish monthly)

  • Tech stack readiness (can implement schema, speed fixes)

  • Governance (legal/brand sign-off within 10 business days)

  • Budget realism (3–6 months to first durable gains)

  • Patience with compounding signals (citations/reviews accrue)
    ≥28 = strong fit; 20–27 = pilot first; <20 = groundwork before GEO.

Example mini-vignettes (anonymised)

  • Regional clinic chain (win): Standardised practitioner bios + treatment schema, insurer pages, review ops → 6 months later, consistent inclusion in AI answers for “[treatment] near me” and insurer-specific queries; bookings up via direct calls.

  • B2B SaaS (win): Integration pages + comparison FAQs + security attestations → cited in AI answers for “best [category] for [stack]”; demo quality improved (fewer tire-kickers).

  • Local trades (win): NAP cleanup + before/after galleries + FAQs + citations → AI summaries started naming the firm; lead volume steadier year-round.

  • Affiliate site (miss): Thin rewrites, no unique testing or data → zero citations in AI answers despite heavy schema.

  • Global enterprise (stall): 90-day legal cycles → content freshness died; pilot approach recommended.