Claude is the better drafter. Perplexity is the better researcher. Never cite Perplexity in a court filing. That's the framework, and it holds across every legal use case we've tested.

Claude (Anthropic) excels at long-form legal writing — briefs, memos, contract analysis — with careful, nuanced output that reads like a senior associate wrote it. Perplexity excels at finding and synthesizing current information with inline citations to real sources. The smart move isn't picking one. It's using both in sequence.

Claude's legal writing quality is the best in the general-purpose AI market. It handles nuance, qualifications, and multi-factor analysis better than any competitor. Ask Claude to draft a motion to compel and you'll get something that needs editing, not rewriting.

Key strengths for attorneys: - 200K token context window — upload entire case files and get coherent analysis - Careful reasoning — Claude hedges appropriately instead of stating wrong things confidently - Drafting quality — sentence structure, tone, and legal terminology are consistently strong - Artifacts — generates documents in a separate panel for easy copying and iteration

Limits: Claude has no internet access (unless using the search tool, which is limited). Its training data has a knowledge cutoff. It can't look up current case law, recent filings, or today's news. For research that requires current information, Claude alone isn't enough.

Perplexity is an AI-powered search engine that synthesizes information from across the web and provides inline citations to sources. For legal research, it's genuinely useful — but with critical caveats.

Key strengths for attorneys: - Real-time web search — finds current information, recent opinions, regulatory updates - Inline citations — every claim links to a source you can verify - Follow-up questions — refine your research iteratively - Academic and news focus — good at finding law review articles, news coverage, regulatory announcements

Limits — and they're serious: - Perplexity searches the open web, not Westlaw or Lexis. It finds news articles about cases, not the cases themselves - Citations point to web sources (news sites, blogs, law firm websites), not official legal databases - It can't access paywalled legal research platforms - Never cite a Perplexity source in a filing. The underlying sources are often secondary, summarized, or commentary — not primary legal authority

The Combined Workflow Every Attorney Should Use

Here's the workflow that maximizes both tools:

Phase 1 — Research (Perplexity): Use Perplexity to map the landscape. Search for recent developments, identify relevant statutes, find law review articles discussing your issue, and discover opposing arguments. Treat Perplexity as your research assistant who reads the internet for you.

Phase 2 — Verify (Westlaw/Lexis): Take every case, statute, and regulation Perplexity surfaces and verify it in your legal research platform. Pull the actual text. Confirm it's still good law. This step is non-negotiable.

Phase 3 — Draft (Claude): Upload your verified research to Claude. Give it the facts, the legal standards, and your argument structure. Let Claude draft the brief, memo, or letter. Its long context window means you can include extensive source material.

Phase 4 — Refine (Claude): Iterate with Claude on tone, structure, and argumentation. Claude's revision capabilities are strong — it takes feedback well and produces meaningfully different drafts.

Total cost: Perplexity Pro ($20/month) + Claude Pro ($20/month) = $40/month for a research-to-draft pipeline.

Pricing Comparison

Claude: - Free tier: Limited messages with Claude 3.5 Sonnet - Pro: $20/month — Claude Opus 4 access, higher limits, Projects feature - Team: $25/user/month — admin controls, longer context, team features - Enterprise: Custom pricing — SSO, advanced security, custom retention

Perplexity: - Free tier: Basic searches, limited Pro searches per day - Pro: $20/month — unlimited Pro searches, file uploads, dedicated AI models - Enterprise: $40/user/month — team features, admin controls, data privacy

For solo attorneys: Both Pro plans ($40/month total) deliver exceptional value. That's less than 15 minutes of billable time at most rates.

For firms: Claude Team + Perplexity Enterprise runs about $65/user/month. Still cheaper than a single Westlaw seat, though you absolutely still need Westlaw.

Why You Should Never Cite Perplexity in a Filing

This deserves its own section because attorneys keep making this mistake.

Perplexity aggregates web sources. When it says "In Smith v. Jones (2024), the court held..." it's probably pulling from a law firm blog post or news article that summarized the case. The underlying source might be: - A blog post that got the holding wrong - A news article that simplified the ruling - A secondary source that's since been corrected - A case that was later reversed or distinguished

Perplexity is a research starting point, not a citation source. Use it to find leads, then verify in Westlaw or Lexis. Judges have sanctioned attorneys for citing AI-generated case references. The fact that Perplexity provides links doesn't make those links citable authority.

Rule of thumb: If you wouldn't cite the underlying website in your brief, don't rely on Perplexity's summary of it.

The Bottom Line: Use Perplexity to find what to research, Claude to write the work product, and Westlaw to verify everything in between — $40/month for the pair, and never cite either one directly.

AI-Assisted Research. This piece was researched and written with AI assistance, reviewed and edited by Manu Ayala. For deeper takes and the perspective behind the research, follow me on LinkedIn or email me directly.