Overview: Seattle-based legal AI platform ($91M total funding: $25M Series A Aug 2024, $60M Series B Apr 2025 led by Sapphire Ventures with Thomson Reuters Ventures). Founded 2021 by former Microsoft engineers Jerry Zhou and Kyle Lam. 27,000+ cases processed, claims $1B+ in settlements supported. Purpose-built exclusively for plaintiff-side PI and mass tort.
Product Suite:
| Product |
Capabilities |
| Supio Inbound |
AI voice agent for 24/7 intake, multi-channel (voice/SMS/web/email), automated case qualification and scoring |
| Case Engine (pre-litigation) |
Instant Timelines (auto-generated medical chronologies), Case Signals (AI flags treatment gaps, missing bills, causation issues), Instant Demands (settlement-ready demand letters with ICD codes), Case Economics (automated billing ledger, lien tracking, settlement projections) |
| Case Bench (litigation) |
Drafting Agents (complaints, motions, interrogatories, expert disclosures), Lookalike Drafting (match firm's style from uploaded samples), Depositions (real-time transcription, contradiction detection, cross-referencing against full case file), Exhibit Builder (auto-generated packages with Bates stamping) |
| AI Assistant |
Chat + Deep Thinking modes over full case file, unlimited queries, prompt library, firm-wide knowledge base |
| Integrations |
Two-way sync with Filevine, Litify, SmartAdvocate, CasePeer, MyCase; Thomson Reuters CoCounsel partnership; Supio API for custom builds |
AI Architecture: "CaseAware AI" with 10 specialized expert models tuned for plaintiff law, handling 112+ case types. RAG with hybrid semantic/exact-match retrieval. Human expert verification on all outputs (claims 97% accuracy). LLM vendors not publicly disclosed.
Security: SOC 2 Type II (Vanta), HIPAA compliant, GDPR compliant, case-level data isolation, explicit no-training commitment.
Pricing: Custom only, sales-led. Positions as "flat pricing: no page limits, no demand-type tiers, no edit penalties" (competitive against EvenUp's per-demand model). Case studies claim $500-$1K savings per case.
Claimed Metrics: 28% higher settlements, 62% increase in caseload capacity, 97% accuracy rate.
What Supio Does Well (mirror or learn from):
- End-to-end lifecycle coverage: intake through trial in one platform. Creates genuine switching cost.
- Source-linked everything: every output cites specific document + page/line. Critical for attorney trust.
- Case Economics module: automated billing ledgers with CPT codes, lien reconciliation, settlement scenario modeling. A real workflow innovation.
- Two-way CMS integrations: not just sending data to Supio, but pushing structured data back. Technically harder and more valuable.
- Domain-specific AI models: distinguishing facts from opinions, understanding medical codes, recognizing treatment gaps. Genuine domain tuning vs. general-purpose.
- Human verification layer: medical/legal experts verify all AI outputs. Defensible accuracy claim.
- Rapid product velocity: from stealth (Aug 2024) to depositions, agentic drafting, voice intake, exhibit building in ~16 months.
Gaps that define our opportunity:
- Zero forensic evidence collection: Supio analyzes documents already in your possession. No web archiving, social media capture, screenshot preservation, or cryptographic hashing for chain of custody.
- Strictly PI/mass tort: family law, harassment, stalking, employment, and criminal defense are entirely unserved. Founders have acknowledged family law is on the horizon but are explicitly holding off.
- No client-facing portal: clients cannot see case status, review documents, or communicate through Supio. Platform is attorney/paralegal-only.
- No chain of custody infrastructure: security architecture protects data from breach, but nothing addresses proving evidentiary integrity for collected materials.
- Privilege architecture not articulated: heavy on HIPAA/SOC2 marketing, silent on how outputs interact with work product doctrine post-Heppner.
- No automated evidence collection from live sources: no ability to monitor social media, capture web pages, or ingest evidence from external platforms.
Strategic Assessment: Supio confirms the "vertical AI evidence tool" business model at scale ($91M funding, Thomson Reuters partnership). But it is architecturally a document analysis platform, not an evidence collection and preservation platform. Intactus occupies a different position in the evidence lifecycle: we start at capture, they start at analysis. The family law and harassment markets are available to own before Supio expands.
Sources: supio.com, TechCrunch (Aug 2024, Apr 2025), Sapphire Ventures blog, BusinessWire, LawNext, PRNewswire
Overview: Atlanta-based, founded 2015 by Jeff Kerr. Bootstrapped, ~8 employees. Cloud-native litigation fact management / case chronology software. Won TechnoLawyer Top 25 (2016). Used by Washington Post for Afghanistan Papers investigation. The closest existing product to Intactus's timeline and evidence organization value proposition.
Product Suite:
| Feature |
Capabilities |
| Fact Chronologies |
Flagship: link facts to exact locations in source documents/audio/video. Dynamic timelines that evolve with new evidence. Filter by issue, person, date range. |
| Document Reviewer |
Patent-pending interface. Highlight text to create fact entries with auto-populated citations. Annotations, assignments, activity feed. |
| Deposition Transcripts |
Dedicated viewer with Bluebook-correct auto-citation, designation support, exhibit linkage, full-text search. |
| Audio/Video Review |
Auto-transcription (Google Speech), synchronized note-taking, time-stamped clip creation. |
| AI: Document Intelligence (Apr 2025) |
AI summarization of lengthy documents, entity extraction (people, orgs, dates, locations, events) with hyperlinked occurrences. |
| AI: Suggested Facts (Jun 2025) |
AI scans documents to propose factual statements with entity/issue linkages. Attorney accepts, edits, or dismisses. |
| AI Assistant |
Chat scoped to case documents only (RAG). Clickable citations to source passages. No general legal research. |
| Reports |
8 types: Facts by Issues, Entities & Documents, Facts Spreadsheet, Simple/Detailed Chronology, Entities with Facts, Case Archive Bundle. |
| Collaboration |
Role-based permissions, 2 free guest users per case (co-counsel, experts, clients). |
Pricing:
| Plan |
Annual (per user/month) |
Monthly |
Key Limits |
| Starter |
$30 |
$40 |
20 docs/case, no AI, no full-text search |
| Advanced AI |
$75 |
$100 |
Unlimited docs, all AI features, usage credits |
| Enterprise |
Custom |
Custom |
20+ users, SSO, dedicated support |
Plus usage-based overages: OCR ($1/100 pages), transcription ($0.06/min), storage ($10/GB/month).
AI Architecture: Google Vision OCR, Google Speech transcription, OpenAI models (per third-party analysis). Attorney-in-the-loop design: every AI output requires human review before entering the case record.
Security: TLS + AES-256 encryption, AWS VPC, 2FA, virus scanning, audit logging. HIPAA available as add-on. No SOC 2 or ISO certifications. No explicit data isolation or no-training commitments for AI processing.
What CaseFleet Does Well (and our strategic response):
- Fact-to-document linkage model: every fact links to an exact location in a source document. Response: Adopted. Intactus implements a Fact Schema with passage-level citations (character offsets, page/line for PDFs) baked into the ingestion pipeline from Phase 1b. Every extracted fact links to the exact source text span. This matches CaseFleet's core organizing principle while adding forensic provenance CaseFleet lacks.
- Attorney-in-the-loop AI: AI suggests, attorney approves. Response: Adapted with a two-tier model. Auto-computed metadata (classifications, summaries, entities) stores automatically, with no approval gate slowing ingestion. Facts and assertions require explicit attorney review with batch approve UX for efficiency. This gives us CaseFleet's professional responsibility defensibility without sacrificing automated intake speed.
- Deposition transcript management: Bluebook citation generation, designation support, and exhibit linkage. Response: Adopted for Phase 2. Dedicated transcript viewer with Bluebook citations, designations, exhibit linkage, and full-text search. Key differentiator: cross-referencing deposition testimony against the evidence vault (social media, emails, financial records) to surface contradictions, something CaseFleet cannot do because they lack the capture layer.
- Report variety: 8 report types covering the litigation lifecycle. Response: Matched and exceeded. 9 named report types across Phase 1b and Phase 2 (Case Timeline, Annotated Timeline, Evidence Index, Entity Map, Evidence Ledger, Vault Export, Custody Certificate, Entity Dossier, Case Brief). Several are structurally impossible for CaseFleet: Custody Certificate (chain-of-custody declaration), Vault Export (full manifest + hash verification), Case Brief (AI-synthesized narrative).
- Guest user collaboration: free external collaborators for co-counsel and experts. Response: Adopted for Phase 2. Guest role with read-only access to the attorney dashboard, scoped to specific cases. Distinct from the client portal. Two free guest slots per case.
- Audio/video review with auto-transcription: synchronized note-taking and auto-clip generation. Response: Split across phases. Phase 2: basic upload + Whisper transcription + searchable transcript as evidence item (covers 90% of family law use cases). Phase 3: rich media review with clip extraction, transcript-linked playback, and passage-level transcript citations.
- Accessible pricing: $75/user/month for full AI. Response: Competitive positioning. Intactus at $99/mo (Vault) and $149/mo (Vault + Analysis) is ~30-100% more than CaseFleet, but delivers forensic capture, chain of custody, client portal, privilege-safe AI, and no overage fees (CaseFleet charges $1/100 OCR pages, $0.06/min transcription, $10/GB/month storage). The total cost of ownership comparison favors Intactus for evidence-heavy cases.
Gaps that define our opportunity:
- Zero forensic evidence capture or preservation: CaseFleet is a post-collection tool only. No web archiving, no hash verification, no chain of custody at point of capture.
- No chain of custody tracking: no cryptographic hash, no custody declaration generation, no tamper-evident manifest. Their audit log is for team coordination, not forensic integrity.
- No automated evidence collection: all evidence entry is manual upload. No automated ingestion from email, web, or social media.
- No privilege-safe AI architecture: sends documents through Google and OpenAI services with no explicit no-training agreements, no SOC 2, no discussion of data isolation. Vulnerable post-Heppner.
- No client portal: guest users see the same interface as staff. No purpose-built client view with simplified evidence submission, case status, or secure messaging.
- No integrations ecosystem: no API, no Zapier, no Microsoft 365, no Clio/MyCase, no Relativity. Completely closed.
- No Bates stamping or document redaction: repeatedly requested by users, not yet built.
- Weak mobile experience: no native app, reported display issues on iPhone.
- Weak visual timeline: breaks down with incomplete data. Tabular formats are reliable but not visually compelling.
User Sentiment (41 reviews across platforms, 91% satisfaction): praise for chronology feature ("irreplaceable"), customer support, and ease of use. Criticism of pricing (especially new storage fees), Starter plan limits, no integrations, and no Bates stamping.
Strategic Assessment: CaseFleet is the tool Intactus is most likely to be compared to by attorneys evaluating timeline/evidence organization capabilities. But CaseFleet starts after evidence is collected; attorneys must capture and organize evidence elsewhere, then manually upload it. Intactus starts at capture and flows through to analysis and court output. The forensic integrity layer (hashing, chain of custody, tamper-evident manifest) is a category-defining differentiator CaseFleet cannot easily replicate. Their lack of SOC 2 and privilege-safe AI architecture are structural weaknesses in a post-Heppner market.
Sources: casefleet.com, Lawyerist review, ABA Journal, Capterra, SoftwareFinder, CaseFleet blog
Overview: Vancouver, BC-based. Founded 2024 by Alistair Vigier. Pre-seed (angel + Canadian government grants, no VC). ~4,000 users in 8 months with zero paid advertising. AI legal research and document automation, not a case management or evidence tool. Competes against LexisNexis and Westlaw, not Intactus.
Product Suite:
| Product |
Capabilities |
| Casey |
AI legal research over 100M+ court decisions (Canada + US). Natural language queries, jurisdictional filtering, document upload/review, contract risk flagging, legal memo/brief generation, multi-language (EN/FR/ES/ZH). |
| CaseForm |
AI court form auto-population from existing filings. Launched via AffiniPay/MyCase partnership (Jul 2025). California initially, 50-state expansion planned. |
| Bespoke Agent |
Custom AI trained on firm's own document history (e.g., 5,000 separation agreements → AI generates new ones matching firm's style). |
| Synthium DataHub |
Enterprise document governance for regulated industries (legal, insurance, healthcare, government). |
Pricing: CA$49.99/month (Standard), roughly 10x cheaper than Westlaw/LexisNexis. Bespoke Agent at ~CA$99/month. Enterprise pricing custom.
Security: SOC 2, ISO 27001, ISO/IEC 42001 (AI management), GDPR, CCPA. Zero data retention by default.
Relevance to Intactus: Caseway is not a competitor; it occupies a different category (legal research vs. evidence management). It is worth monitoring because:
- Distribution strategy is instructive: embedding into MyCase/AffiniPay rather than building a full practice management stack. Intactus should consider similar integration plays with Clio/MyCase rather than competing with them.
- Bespoke Agent is a stickiness model: custom firm-specific AI trained on proprietary documents creates real switching costs. Intactus could consider practice-area-specific analysis lenses that learn from a firm's case history.
- Access-to-justice narrative: Caseway's mission framing resonates with courts, governments, and academic funders. Intactus's victim self-service angle carries similar narrative power.
- Pricing pressure: at CA$49.99/month for AI-powered legal tools, Caseway sets an expectation for what small-firm attorneys consider reasonable. Our $149/month Vault + Analysis tier needs to clearly demonstrate value beyond research.
Risks: CanLII copyright lawsuit (reportedly moving toward settlement), pre-seed stage fragility, Trustpilot complaints about billing/cancellation UX, US coverage still nascent.
Sources: caseway.ai, BetaKit, LawNext, IT Brief Canada, Vancouver Tech Journal, Advocate Daily
Overview: San Francisco-based legal AI platform. Founded 2022 by Winston Weinberg (ex-O'Melveny) and Gabriel Pereyra (ex-DeepMind/Meta AI). $1B+ total funding across multiple rounds. Series F (March 2026): $200M at $11B valuation, led by GIC and Sequoia, with a16z, Coatue, Conviction Partners, Elad Gil, Evantic, and Kleiner Perkins. 100,000+ lawyers across 1,300+ organizations in 60+ countries. ~$190M ARR (Jan 2026), up from $100M (Aug 2025) — 90% growth in 5 months. Majority of AmLaw 100 firms are customers.
Not a direct competitor — Harvey targets enterprise BigLaw ($1,200+/seat, 20-50 seat minimums). Intactus targets solo/small firms ($99-149/seat, no minimums). But Harvey is the dominant player shaping the legal AI market, and understanding its strategy, product architecture, and weaknesses is critical for Intactus's positioning.
Product Suite:
| Product |
Capabilities |
| Assistant |
General-purpose legal AI chat: research, drafting, analysis, contract review. Multi-model (routes across OpenAI, Anthropic, Google DeepMind depending on task). |
| Workflow Builder |
Drag-and-drop agentic workflow creation with 4 block types: user input, AI action, logic, output. 15,000+ custom workflows built by customers. |
| Vault |
Secure document analysis environment for deal rooms and due diligence. Large-volume document ingestion and risk surfacing. |
| Workflow Agents |
Autonomous multi-step processes for due diligence, contract review, compliance. Configurable conditional logic and classification. |
| Spectre (internal, April 2026) |
Internal autonomous agent that monitors company activity (incidents, bug reports, Slack, customer feedback) and makes engineering prioritization decisions without human prompts. Described as "a company world model." Being demoed to law firm clients as preview of autonomous legal agents. |
| LexisNexis Integration (2025 alliance) |
Lexis+ Protege service within Harvey: primary law content, Shepard's Citations, co-developed motion workflows (Motion to Dismiss, Motion for Summary Judgment). RELX (Lexis parent) is an investor. Expected to add ~$400-600/lawyer/year to effective cost. |
AI Architecture: Multi-model routing across OpenAI (GPT-5.2), Anthropic (Claude), and Google DeepMind (Gemini 3). One of 6 enterprise partners live in Claude's plugin ecosystem. "Trust Stack" hallucination detection: decomposes generated responses into individual factual claims, cross-references each against authoritative sources, applies legal reasoning patterns, flags inconsistencies. Claims 0.2% hallucination rate in internal evaluations.
Pricing:
- ~$1,200/lawyer/month with 12-month commitments and 20-50 seat minimums
- Annual contract floor: ~$288,000+ before add-ons
- Enterprise-only, no published tiers, no free trial, no self-serve
- Custom negotiation for every deal — some firms reportedly negotiate down significantly
- All-in cost expected to climb ~$400-600/lawyer/year once LexisNexis content is bundled
Go-to-Market Strategy (the playbook that built a $190M ARR business in 36 months):
- Ex-BigLaw salespeople: Hired lawyers from White & Case, Latham & Watkins, Skadden. Domain credibility that general tech salespeople can't match. The people selling Harvey spoke the buyer's language from day one.
- Anchor logo strategy: A&O Shearman (Dec 2022) → Paul Weiss (Jan 2023) → signed prominent US firms → majority of AmLaw 100. Each logo made the next sale easier.
- Embedded staff at large clients: Harvey embeds engineers and ex-lawyer customer success staff on-site to customize workflows. Creates deep lock-in that no feature comparison can overcome.
- 10% of team is ex-lawyers in customer success: Drive change management, implementation, and adoption within firms. Ensure utilization thresholds needed for renewal are hit.
- Expand beyond BigLaw: Moved into corporate in-house counsel (PwC, Adecco Group) and professional services. This is the TAM expansion narrative.
- Revenue growth: $50M ARR (end 2024) → $100M (Aug 2025) → $190M (Jan 2026). 3.8x YoY. Weekly active users quadrupled in a single year. Customer base expanded from 40 to 1,300+ organizations.
What Harvey Does Well (mirror or learn from):
- Multi-model architecture as trust signal: Routing across OpenAI, Anthropic, and Google eliminates vendor lock-in concerns and lets Harvey pick the best model for each task. Enterprise buyers want to know they're not locked to one provider's roadmap. Intactus response: We use Anthropic exclusively, which is actually a privilege advantage (single ZDR agreement, one vendor to audit), but we should articulate why single-vendor is better for our use case, not a limitation.
- Trust Stack / hallucination detection: Decomposing AI output into individual factual claims and cross-referencing against authoritative sources. 0.2% hallucination rate is a credibility claim that moves enterprise deals. Intactus response: Our source-citation architecture (every fact links to exact text span) is the small-firm equivalent. We should quantify and market our accuracy rate.
- Workflow Builder as moat: 15,000+ customer-built workflows = 15,000 reasons not to switch. User-created workflows compound switching costs in a way a chat interface never will. Intactus response: Our instruction profiles (firm defaults, attorney preferences, case-type presets) serve the same function at a different scale. Each profile a firm builds makes leaving harder.
- Embedded customer success: 10% of Harvey's team is ex-lawyers doing on-site implementation. This is the enterprise playbook, but the principle applies at any scale: people who understand the domain drive adoption. Intactus response: Our design partner model with Nathan's firm is the bootstrapped equivalent. As we grow, attorney-background support staff should be a priority hire.
- Anchor logo GTM: Land 2-3 prestigious early customers, then leverage social proof. Harvey went A&O → Paul Weiss → AmLaw 100. Intactus response: Our version is local: land Nathan's firm, then leverage his bar association network. State bar CLE presentations with a reference customer are our anchor logo strategy.
- LexisNexis partnership: Rather than competing with the incumbents (Westlaw, Lexis), Harvey partnered with one. Smart positioning: Harvey provides the AI layer, Lexis provides the content layer. Intactus response: We should consider similar partnerships for primary law content rather than building our own legal research. Partner where we're weak, own where we're strong (evidence capture, chain of custody, AI analysis).
Harvey's Weaknesses (exploit these):
- The billing model paradox: Law firms bill by the hour. Harvey makes work faster. If Harvey replaces associate hours, it undermines the revenue model of its own customers. Firms are incentivized to not fully adopt it. Intactus opportunity: Our value prop is "handle more cases without hiring," not "do fewer hours per case." We grow the attorney's revenue, not shrink it. This alignment is a structural advantage over Harvey.
- The Claude gap: Multiple sources (lawyers, review sites, comparison articles) explicitly say "Claude gives 90% of Harvey's capability at 10% of the cost." When a $20-500/month general-purpose AI delivers 90% of a $1,200-2,500/month product's value, the moat is governance + workflow + brand trust, not AI capability itself. Intactus opportunity: We should never compete on "AI quality" alone. Our moat is forensic integrity, chain of custody, and privilege-by-design — things general-purpose AI cannot provide.
- Top-down adoption, bottom-up rejection: Management buys Harvey. Actual lawyers either prefer ChatGPT/Claude directly or don't use AI at all. One X user called it "universally panned by lawyers" despite firm-level adoption. A Singapore senior partner said "we use Harvey" while juniors on the same floor said "there's no AI, I'm proofreading documents." Intactus opportunity: Product-led growth. Build something lawyers choose to use because it makes their specific workflow better, not something management buys and mandates.
- Small/mid firm blindspot: 100% enterprise focus. No self-serve tier. No path for a 5-person firm. $288K+ annual minimum. Intactus opportunity: This is our entire market. The 650,000 solo/small firm attorneys that Harvey will never serve.
- No evidence capture or chain of custody: Harvey is a document analysis and drafting platform. It does not capture evidence, does not hash artifacts, does not produce chain-of-custody documentation, and cannot prove evidentiary integrity. Intactus opportunity: This is our category-defining differentiator. Harvey (and every AI legal tool) starts after evidence is in hand. We start at capture.
- Mobile is query-only, not capture (updated April 2026): Harvey now has a mobile app (Scan document, Record transcript, Upload, voice input — visible in their platform demo video). But their mobile is for querying the AI — not for capturing and preserving evidence with integrity metadata. Intactus opportunity: Our mobile evidence capture workflow (screenshot → hash → vault → chain of custody → searchable in seconds) serves a fundamentally different purpose. Harvey's mobile lets you ask a question; ours lets you preserve evidence.
- Pricing excludes the broader market: One X user summarized it: "The legal tech industry spent 3 years telling solo lawyers AI would level the playing field. Then they priced it so only BigLaw could afford it." Intactus opportunity: At $99-149/month, we're 8-16x cheaper than Harvey. We ARE the tool that levels the playing field.
- Valuation skepticism: $16.5B in combined legal AI valuations (Harvey + Legora) against a $2-4B traditional legal tech TAM. Harvey's pitch is "we're not legal tech, we're legal services — that's a $1T reframe." That's a bet, not a proven market. If the reframe doesn't hold, Harvey is dramatically overvalued. Intactus relevance: We're building a profitable small business, not chasing venture-scale valuation. Our economics work at 75 paying attorneys. We don't need the TAM reframe to be true.
User Sentiment (synthesized from X posts, review sites, and legal tech blogs):
| Signal |
Detail |
| Positive |
40% time reduction on document review. "Best in class" for transactional work (M&A, due diligence). Half of some firms use it daily. |
| Negative: restrictive UX |
"Using Harvey is actually a worse, more restrictive experience than just using ChatGPT directly" (practicing lawyer) |
| Negative: panned by users |
"Harvey is universally panned by lawyers, both in house and BigLaw. Business managers at firms sign up because 'if everyone is using it we need it too'" |
| Negative: uneven adoption |
Senior partner says "we use Harvey and ChatGPT"; junior associate on same floor says "there's no AI, I'm proofreading documents" |
| Negative: pricing |
Aggressive sales tactics, mandatory license counts, rigid 12-month contracts, no easy trial or flexible plan |
| Negative: small firm exclusion |
"CoCounsel: $900/month per user. Harvey: $2,500/month per user. They priced it so only BigLaw could afford it" |
| Competitor preference |
Multiple attorneys explicitly prefer Claude at $500/month over Harvey at $2,500/month, citing "nearly identical output" |
Strategic Assessment: Harvey confirms that legal AI is a massive, venture-scale market. Its rapid growth ($50M → $190M ARR in 12 months) proves attorneys will pay for AI tools. But Harvey's enterprise-only pricing, top-down sales model, and lack of evidence capture create a structural gap that Intactus is designed to fill. Harvey serves the top 0.1% of law firms by size. Intactus serves the other 99.9%.
The most important strategic insight: Harvey's weaknesses are not bugs, they're features of its business model. Harvey cannot serve small firms profitably at $1,200/seat with embedded staff and custom negotiations. Harvey cannot add forensic evidence capture without rebuilding its architecture. Harvey cannot solve the billing model paradox for its own customers. These are structural constraints, not product gaps that Harvey will close with a feature release. This makes Intactus's positioning durable.
Sources: CNBC (Mar 2026), TechCrunch (Feb/Mar 2026), harvey.ai blog, Bloomberg (Mar 2026), Artificial Lawyer (Apr 2026), Sacra, Contrary Research, Getlatka, Fast Company, Purple Law, claudeforlawyers.com, X/Twitter (24 posts, 4,800+ likes)
