8 Harvey Legal Research Agent Alternatives Our Panel Actually Ships

Shared AI research workspaces for teams to annotate and build together

“The direct competitor here is 'Notion AI plus a shared doc,' and Perplexity beats it on one specific axis: the research artifact and the annotation layer are the same object. You're not copy-pasting AI output into a doc and losing provenance. Where this breaks is at scale — the moment a team has 50 Research Pages and no folder structure or cross-page linking, it becomes a graveyard of orphaned reports. Perplexity has 12 months before Microsoft Copilot Pages ships something functionally identical inside Teams, so the clock is running.”— The Skeptic

AI search for regulated teams — with SSO, audit logs, and data residency

“Perplexity Enterprise is checkboxes done correctly: SAML SSO, EU data residency, audit logs — these aren't differentiators, they're table stakes for any Fortune 500 procurement conversation, and Perplexity finally has them. The real question is whether enterprise IT buyers trust a 2-year-old AI search company with their data over Microsoft Copilot, which ships the same compliance stack with an existing vendor relationship and a known legal team. My prediction: Perplexity wins in the departments that have already bypassed IT to use Pro, and loses everywhere IT controls the procurement process. What would flip this? A marquee referenceable customer in a regulated vertical, announced publicly, with a case study.”— The Skeptic

Run Python & R code inside your search sessions, sandboxed and persistent

“The primitive here is a REPL with persistent session state embedded in a retrieval interface — that's actually a non-trivial thing to ship correctly, and sandboxed container isolation per session is the right call, not a toy iframe. The DX bet is that you never leave the search context to crunch numbers, which works until you need pip installs beyond the pre-loaded environment or you want to pull in your own data files without pasting CSVs into a chat box. The moment of truth is asking it to analyze a dataset you found in the same session — if that works end-to-end without copy-paste, that's genuinely useful. It's not replacing a Jupyter notebook for serious work, but it doesn't need to: it earns its keep for quick validation tasks where spinning up a local environment is the thing that was stopping you.”— The Builder

RAG model with citation-level grounding for regulated enterprise search

“The primitive is clear: a RAG model that returns answers with document-level citations baked into the response structure, not bolted on post-hoc. The DX bet is on the connectors — pre-built integrations to Salesforce, SharePoint, and Confluence mean the 'connect your data' step doesn't require you to write a chunking pipeline at 2am. The moment of truth is whether those connectors handle real enterprise data shapes (nested Confluence spaces, Salesforce custom objects) without breaking — the docs suggest yes but I haven't stress-tested edge schemas. What earns the ship is that citation grounding is a first-class output type, not a hallucinated footer: the API returns source references as structured fields, which means downstream auditing is an engineering problem you can actually solve.”— The Builder

Extended thinking for grad-level math, science, and coding

“The primitive here is straightforward: a reasoning model that allocates more inference compute to hard problems before returning a result. The DX bet OpenAI made is to hide all of that behind the same ChatGPT interface you already use — no new API surface to learn, no config, just select o3 Pro from the model picker. The moment of truth is dropping a genuinely hard coding problem or a graduate-level proof and watching whether the extended thinking trace actually catches errors that o3 misses — in my experience, it does on non-trivial linear algebra and dynamic programming. The honest caveat: if you're accessing this via API you're paying per-token and the latency is real; this is not a drop-in for production pipelines. Ship for the specific use case of hard reasoning problems where correctness matters more than speed.”— The Builder

Grounded AI research assistant with internal knowledge and audit trails

“The direct competitors here are Glean, Microsoft Copilot with SharePoint grounding, and — honestly — a well-configured Notion AI with a few connectors. Perplexity's actual differentiator is its search-grounded citation chain, which is real and meaningfully reduces hallucination risk compared to raw GPT-4 deployments. Where this breaks: any enterprise with a complex permission model — the moment you need row-level security across data connectors, the 'grounded' story gets complicated fast. Prediction: Microsoft eats 60% of this market within 18 months by bundling Copilot deeper into M365, but Perplexity survives as the default for companies that haven't standardized on the Microsoft stack yet.”— The Skeptic

Web search + your docs, synthesized into cited briefs inside Notion

“The job-to-be-done here is sharp: a knowledge worker needs to produce a research brief without leaving the document they're already writing in. Notion's bet is that context-switching to a browser and back is the actual friction, and Research Mode eliminates exactly that. What earns the ship is that it doesn't require the user to set anything up — the AI add-on subscribers just get it, which means time-to-value is measured in seconds, not configuration screens. The gap to watch is whether the document synthesis is meaningful or decorative — if internal pages surface as citations but don't actually change the output, users will notice within a week and stop triggering it.”— The PM

Web browsing and cited sources baked into your Notion workspace

“The job-to-be-done is unambiguous: synthesize external information into a Notion doc without leaving the tab. That's a real friction point for anyone using Notion as a second brain or team wiki — the copy-paste-cite loop from browser to doc is genuinely painful and Research Mode kills it. Onboarding is effectively zero because it surfaces inside a workflow the user already has; there's no new app to learn, no new mental model, just a new slash command or AI prompt. The gap is completeness around source control — users can't currently filter by date range or exclude domains, which means research tasks with recency requirements still need a dedicated tool running in parallel.”— The PM