Middesk Replaces a Tangle of Forms, Slack, and Email with a Single AI-Native Legal Hub

The Challenge: Four Core Constraints

Before Sandstone, Middesk's legal function was running on tools that couldn't scale with the work:

1. Fragmented Intake: Requests arrived simultaneously through a Google Form nobody consistently used, ad-hoc Slack messages, and inbound emails. Nothing was centralized, triage was purely manual, and there was no single place to see what was open.
2. Unreliable AI Redlining: General-purpose models were being used for contract redlines, but outputs were inconsistent and Word documents were frequently hallucinated — making the tooling unusable for real negotiation work.
3. Repetitive Policy Questions: A significant share of requests were the same policy questions Sam had already answered many times (InfoSec, signing authority, and similar), with no automated way to surface those answers to the business.
4. No Codified Playbooks: With two product lines — each with its own direct and reseller MSA, plus annexes — negotiation guidance lived in Sam's head and in scattered past redlines, not in any reusable system.

The Solution: Why Sandstone?

Middesk needed a platform that could serve as a full legal operating layer on day one, without the overhead of a traditional CLM rollout. Sandstone was selected for five capabilities that mapped directly to the team's workflow:

• Unified Intake Across Slack and Gmail: Sandstone ingests requests natively from the #legal-asks Slack channel (via @Sandstone) and through a Gmail plugin, routing everything into a single Kanban board. This let Middesk deprecate its Google Form entirely and eliminate parallel intake channels.
• Auto-Detected Contract Review: Contracts submitted via Slack or email are automatically classified by type, matched to the appropriate playbook, and returned as a first-pass redline — all without leaving the platform. The team is starting with NDAs and MSAs and building out coverage across both product lines.
• Knowledge Bank for Policy Q&A: Internal policies are uploaded to Sandstone's Knowledge Bank, enabling the Slack bot to auto-respond to routine questions. Sam seeded it with a guidance memo distilled from the full history of the #legal-asks channel, turning years of past answers into a living reference.
• AI-Generated Playbooks: Rather than waiting on outside counsel to draft playbooks from scratch, Sandstone generates them from existing templates and a handful of past redlines. The Intelligence tab then tracks negotiation patterns over time — for example, where pushback on liability caps is trending — and suggests playbook refinements.
• Google Drive as the Document Backbone: Executed contracts and policies stay in Google Drive, where they already live, and sync into Sandstone — so search, precedent lookup, and the Knowledge Bank stay current without manual re-uploads.

The Results

Middesk's Sandstone rollout is in active pilot, with integrations live and playbooks in build-out. Early indicators include:

• Consolidated Intake: Google Form deprecated; Slack and Gmail now feed a single triage board, giving Sam one place to see every open matter.
• Reliable First-Pass Redlines: Contracts routed through Sandstone receive auto-classified, playbook-aligned redlines — removing the hallucination risk that made general-purpose AI tools unworkable for real contracts.
• Automated Policy Responses: Routine policy questions are increasingly deflected by the Knowledge Bank-powered Slack bot, freeing Sam to focus on higher-leverage contract work.
• Playbooks Under Construction: NDA and MSA playbooks are being generated directly from past executed documents, with the Intelligence tab set to refine them as usage accumulates across both product lines.

“The coolest part is that it gets better with more use.”