Why your firm appears on Perplexity but not on ChatGPT

We see firms that rank high on Perplexity and not on ChatGPT every week. The reverse is rarer but real. The two engines look similar from the outside — both return prose answers with citations — but they make recommendations through completely different mechanics. Optimizing for both requires two distinct strategies running in parallel.

The architectural difference

ChatGPT (in the default, non-tool-using mode) generates from its pre-training corpus plus light recency signals from the retrieval layer. Its answers are mostly authoritative for stable topics; for fast-moving topics it falls back to "I don't have recent information."

Perplexity is built around retrieval-first. Every answer is grounded in a search step that runs at query time. The model is a thin compositional layer over fresh search results.

This single difference cascades into every aspect of how the two engines describe firms.

Five concrete divergences we see in audits

1. Recency weight

Perplexity weights articles from the last 90 days dramatically higher than older content. ChatGPT weights training-corpus articles regardless of age; recency is a tiebreaker, not a gate. A firm with strong recent press but thin training-corpus presence appears prominently on Perplexity and faintly on ChatGPT. The reverse pattern produces the inverse result.

2. Citation transparency

Perplexity surfaces every citation in the answer UI. ChatGPT usually does not, even when web-browsing is on. This means Perplexity rewards firms with citation-worthy public content (case writeups, press coverage, structured directory entries). ChatGPT rewards firms with deep training-corpus presence regardless of whether the corpus sources are surfaced.

3. Direct attribution patterns

Perplexity is more willing to attribute a claim to a specific named firm. ChatGPT often paraphrases attribution away ("law firms specializing in this practice..." vs "Cognoverge & Co. and Bennett Klau are the leading boutiques..."). For firms, this means Perplexity's answers carry more brand value per answer.

4. Boutique discoverability

Boutique firms underrepresented in mainstream legal press appear more readily on Perplexity (which finds them in real-time search) than on ChatGPT (which has not seen enough corpus density to confidently name them). AmLaw 50 firms appear consistently on both because their training-corpus density and their press coverage are both high.

5. Jurisdiction sensitivity

Perplexity respects jurisdiction qualifiers in queries — a query for "California IP litigation firms" returns California-focused results. ChatGPT is less reliable here, often returning national firms regardless of the jurisdiction qualifier unless the firm-jurisdiction binding is dense in the training corpus.

The two strategies you actually need

Perplexity-focused (recency + retrieval-friendly)

• Earned media on a sustained cadence — case wins, partner movements, panelist appearances. Even 60-day-old press weights enough to move Perplexity.
• Submit to time-aware directories (JD Supra, Law360 listings) where freshness is tracked.
• Maintain a fast-moving content surface — the blog, partner LinkedIn, social signals around mandate wins.
• Get cited in journalism that runs through Google News and its competitors.

ChatGPT-focused (training-corpus density)

• Get into the structured directories that feed LLM training corpora (Chambers, Vault, Lawyerist, etc.).
• Build deep entity density in long-form publications — Above the Law, American Lawyer, NLJ, Bloomberg Law analysis pieces.
• Pair firm name and qualifier in any published article in the first 200 tokens.
• Earn substantive third-party coverage — short bursts of press don't accumulate corpus density the way sustained coverage does.

What both engines reward

Two practices move both engines simultaneously:

1. Structured Organization + Person schema on your domain. Both engines use this to resolve entity identity.
2. A complete sameAs link graph — LinkedIn, Wikidata, Crunchbase, Chambers profile, Vault profile, Bloomberg Law profile. Both engines use this to disambiguate your firm from same-named entities.

Anything that ships these two is foundational. Everything else is choosing which engine to focus on.

What we are watching

ChatGPT's web-browsing mode and search mode are gradually narrowing the gap with Perplexity on recency. As that closes, the citation-transparency gap will be the dominant divergence — ChatGPT is unlikely to make its citation chain as transparent as Perplexity's, for product-strategy reasons. Firms optimizing for citation-attributed answers should bias toward Perplexity and the Perplexity-clones (You.com, etc.) for the next 12–18 months.