Sandstone triages and routes legal work coming through Slack, email, and Jira — then builds custom AI workflows for the in-house teams that Harvey was never built for.
ENTRY ANGLES
Build institutional legal knowledge base product for GC-level buyers — every Sandstone-processed decision becomes searchable precedent · Build outside counsel coordination layer: package in-house context as structured brief, receive external work back into same workflow · Create compliance infrastructure for regulated verticals (insurance, healthcare, financial services)
VERTICALS
CAPABILITIES
Legal domain expertise, Document AI and NLP, Workflow integration (Slack, email, Jira), Knowledge management
The in-house legal team at a technology company receives legal requests through every channel simultaneously. A salesperson pings legal on Slack asking whether a contract clause is acceptable. A product manager sends an email asking about a regulatory question. Engineering opens a Jira ticket requesting a data processing agreement. Each of those requests lands in a different queue, gets routed informally, and leaves no institutional record of how it was resolved. Sandstone exists to replace that with a structured intake, routing, and workflow system built specifically for what in-house teams actually do.
The platform connects to Slack, email, and Jira, pulls incoming requests into a unified view, and routes each by type and urgency. From there, teams build custom AI workflows for what follows: contract review against a company-specific playbook, NDA generation with pre-approved terms, regulatory question drafting. The AI handles routine drafts and flags edge cases for human review. The design distinction is deliberate — Sandstone isn't trying to replace the in-house attorney; it's trying to stop attorneys from spending most of their time triaging and templating rather than exercising judgment.
Nick Fleisher, who led McKinsey's legal technology practice, and Jarryd Strydom, a McKinsey alum who is himself an attorney and an engineer, co-founded Sandstone in New York. The company launched publicly in January 2026 with a $10 million seed led by Sequoia Capital, then raised a $30 million Series A led by Lightspeed Venture Partners six months later. Revenue grew 40x in the 90 days ending around the June close. Customers include Wayfair, Grindr, Mercury, Cox Media, and ElevenLabs.
The structural reason existing AI legal tools haven't solved this problem is that they were built for law firms. Harvey, Legora, and most other AI legal tools were designed for the billable-hour model: the work is organized around client matters, conducted by associates whose job is research and drafting, and delivered through formal written outputs. In-house teams operate on none of those assumptions. The in-house team is a cost center measured by deal velocity. It is small relative to the company it serves. Its work arrives through informal channels, not formal matter-opening processes.
The intake problem is the most important place to start because it precedes everything else. An automation built on top of a chaotic intake process will automate the chaos. Sandstone's choice to build the inbox and routing layer first — before the AI drafting, before the workflow automation — is the correct sequencing, even if it means the early product looks less impressive than a tool that can generate contracts from scratch.
The 40x revenue growth in 90 days is an unusually fast signal for legal software sales, which are typically measured in quarters. In-house legal teams are conservative buyers; a decision that adds a new tool to the legal workflow requires justification up to the GCO or CLO level. The fact that Sandstone is closing those sales quickly at companies like Wayfair and Mercury suggests the pain it's addressing is already broadly recognized at the decision-making level, not just felt by individual attorneys.
The most valuable long-run extension is institutional knowledge capture. Every legal decision Sandstone processes — a contract clause was acceptable, a vendor agreement was approved with these modifications, a regulatory question was answered this way — is a precedent. A legal team that has run its intake through Sandstone for two years has a library of every decision made, indexed by counterparty type, clause risk level, and resolution. That knowledge base makes each new attorney hire effective from day one, makes the team measurably more consistent across deals, and makes any incoming audit or litigation discovery significantly cheaper to respond to. The knowledge base also has acquisition value: it is institutional legal memory that doesn't walk out the door when a GC leaves.
The second opening bridges in-house and outside counsel. In-house teams don't do all their own legal work; they route complex matters to external firms. The coordination overhead — briefing outside counsel with the relevant company playbook and risk tolerances, tracking matter status, reconciling their output against internal position — is expensive and largely manual. A Sandstone integration that packages the institutional context and sends it as a structured brief to outside counsel, then receives their work back into the same workflow, would position Sandstone as the command layer for all legal work at a company, internal and external. The outside counsel market is too large and too fragmented to sell directly into; building the integration that makes outside counsel easier to direct through Sandstone is a wedge into that relationship without requiring Sandstone to become a law firm itself.