DeepDive Research
An AI research assistant built specifically for analysts, journalists, and consultants that runs a multi-step investigation — broad discovery, entity disambiguation, source triangulation — and produces a memo with every claim linked to a verifiable source the user can audit in one click.
management consultants, investigative journalists, and equity analysts producing research memos under deadline
- Multi-source discovery that scans reports, filings, podcasts, and primary sources — not just top-10 web pages — and ranks them by evidence quality, not SEO
- Entity disambiguation layer that flags when "John Smith (CEO)" and "John Smith (Senator)" are being conflated and asks the user to confirm before synthesizing
- Claim-level citations so every assertion in the memo hyperlinks to the exact paragraph supporting it, with a confidence flag when only one weak source backs it
- Counter-argument generator that deliberately searches for the strongest case against the user's working thesis so memos don't fall for one-sided narratives
Perplexity's WANDR benchmark (116 stars) formalizes the gap: even frontier models are weak at wide-and-deep research — entity disambiguation, evidence-backed synthesis. A product that scores well on WANDR-style criteria becomes the trusted research partner for professionals who can't afford an AI hallucination in a client memo.
Perplexity's WANDR benchmark confirms real, named workflows (competitive mapping, due diligence, market analysis, due diligence, talent sourcing) for exactly this audience, and established vendors (AlphaSense, Hebbia) already sell heavily to consultants/analysts.Perplexity's Open-Source WANDR Benchmark Exposes How Badly Research Agents Fail ↗WANDR Benchmark: Evaluating Research Agents That Must Search Wide and Deep ↗
Market is crowded: Perplexity Deep Research ($20/mo, 300/day), ChatGPT Deep Research, Google Deep Research, plus AlphaSense, Hebbia, Elicit and journalism-specific tools (AURA, HyperWrite AI Journalist) all already produce cited long-form reports for this audience.Google Deep Research vs Perplexity vs ChatGPT (2026) ↗AlphaSense vs Hebbia ↗AURA: AI Research Tool for Journalists ↗
Clear willingness to pay at multiple tiers: $20/mo Perplexity Pro with 300 DR queries/day for pros, while AlphaSense and Hebbia charge enterprise pricing to the same consultant/analyst buyer who bills $500+/hr.Perplexity vs ChatGPT Deep Research: Ultimate 2026 Guide ↗AlphaSense vs Hebbia vs Bloomberg Terminal for front-office research ↗
Cited, auditable research memos are a permanent professional deliverable; liability and trust requirements for client memos and published journalism ensure the verification layer only becomes more valuable as AI-generated content proliferates.WANDR Benchmark Open Source — July 2026 | explainx.ai Blog ↗
Doable with current LLMs but non-trivial: WANDR shows even Perplexity's own Search-as-Code system hits only ~0 on the benchmark, confirming that entity disambiguation plus source triangulation at memo-grade quality remains genuinely hard engineering.WANDR Benchmark Open Source — July 2026 | explainx.ai Blog ↗Wide Research - Perplexity ↗