How In‑house Legal Can Cut Intake‑to‑Decision to 24 Hours with AI Triage

Legal teams don’t lose weeks in negotiation; they lose them in intake. McKinsey estimates knowledge workers spend ~28% of their week on email and another chunk searching for information. When most legal requests still arrive via email or Slack—with missing context—delay is inevitable. The fastest in‑house teams are compressing intake‑to‑decision to 24 hours, not by adding another portal, but by orchestrating AI triage on top of their playbooks and positions.

The Hidden Drag: Intake Is Your Bottleneck

Intake is where work fragments: incomplete briefs, back‑and‑forth questions, manual classification, and handoffs that disappear into threads. Counsel context switches, duplicate requests pile up, and simple approvals wait behind complex matters.

The fix isn’t “more forms.” It’s an operating system that captures requests where they start, normalizes the data, applies your policy positions, and proposes the next decision—automatically. That’s the promise of AI triage when your institutional knowledge is structured and accessible.

Sandstone turns playbooks, fallback positions, and workflows into a living knowledge layer. Each intake is a chance to apply that knowledge—and improve it—so the next similar request resolves faster with less risk.

What AI Triage Looks Like in Practice

AI triage is a coordinated set of steps that happen in seconds:

- Capture: Ingest requests from email, Slack, procurement, or a simple link without forcing new behavior.

- Normalize: Extract parties, contract type, spend, data categories, and jurisdictions; map to your matter taxonomy.

- Classify: Determine risk tier and route based on business unit, counterparty profile, or data sensitivity.

- Decide: Compare request against your approved positions; propose “self‑serve,” “approve,” or “escalate with context.”

- Draft: Generate a response, summary, or first redline anchored to your playbooks and clause library.

- Record: Log the decision, rationale, and artifacts to your system of record for auditability.

Three high‑volume examples:

- Self‑serve NDAs: AI checks counterparty paper against your stance, applies standard edits or swaps in your template, and returns a signed‑ready NDA—no lawyer touch unless out of guardrails.

- Vendor reviews: Intake captures data flows and spend, classifies risk, triggers DPIA/DTIA questionnaires, and assembles a decision memo for security and privacy in one thread.

- Marketing claims: AI matches proposed copy against your substantiation rules, flags risky phrasing, and suggests compliant alternatives before legal ever sees it.

On Sandstone, AI agents operate within your layered knowledge—your definitions of “low risk,” your escalations, your fallback clauses—so outcomes are fast and defensible, not generic.

Design the Operating System, Not Another Form

Adoption lives or dies on friction. Meet requesters where they work, ask only once, and make the path obvious.

Principles that drive adoption:

- Natural integration: Intake from Slack or email; Sandstone collects what’s missing without a maze of fields.

- Layered data: Reuse company, vendor, and contract metadata; prefill whenever possible.

- Modular workflows: Snap in approvers (Finance, Security) based on context; no one waits on legal for non‑legal gates.

- Positions as code: Encode playbooks and thresholds (e.g., “<$50k SaaS with no PII = auto‑approve with standard DPA”). The agent enforces the line, not the calendar.

Legacy portals often centralize chaos; they don’t remove it. An AI‑first operating layer reduces touches, shortens queues, and leaves a clear audit trail.

Metrics That Matter: From Queue to Decision

What you measure drives behavior. Start with these:

- Intake‑to‑first response: Target <2 hours for all requests.

- Intake‑to‑decision for low risk: Target <24 hours (self‑serve NDAs, low‑spend vendor renewals).

- Auto‑resolution rate: % of matters resolved without attorney touch; 30–50% is common with strong playbooks.

- Rework rate: % of requests bounced back for missing info; drive this down with smarter prompts and prefills.

- Playbook coverage: % of volume covered by encoded positions; expand coverage weekly.

- Queue age distribution: How many matters >3 days old; a leading indicator of bottlenecks.

- Stakeholder CSAT: Short post‑resolution pulse in Slack or email; correlate sentiment with cycle time.

Instrument these metrics before and after deploying AI triage to show concrete ROI in weeks, not quarters.

Start Small: A 30‑Day Plan

You don’t need a transformation program to start. You need one high‑volume workflow and a feedback loop.

- Week 1 — Baseline and map: Identify your top request type by count (often NDAs or low‑spend vendor reviews). Document current steps, owners, and average cycle time.

- Week 2 — Encode positions: In Sandstone, translate your playbook into machine‑readable rules: thresholds, fallback clauses, “hard stops,” and auto‑approve criteria.

- Week 3 — Pilot AI triage: Enable intake from Slack/email, auto‑extract key fields, and route to self‑serve where in‑guardrail. Keep humans in the loop on edge cases.

- Week 4 — Expand and measure: Add one adjacent workflow (e.g., DPA addenda). Publish metrics and wins; adjust prompts and rules based on exceptions.

Actionable takeaway: Pick one workflow with high volume and low variance (self‑serve NDAs). Encode your approval stance and fallback positions, then enable AI triage for a 10‑day pilot. Success looks like a 40%+ auto‑resolution rate and <24‑hour decisions for in‑guardrail matters.

Why This Compounds: Knowledge as an Asset

Every resolved request enriches your legal foundation: new clauses, clarified thresholds, and sharper risk patterns. On Sandstone, that knowledge becomes instantly reusable—the next similar intake is faster, the exception list is smaller, and the audit trail is clearer.

This is how legal shifts from reactive queue to proactive force: layered data, crafted precision in workflows, and natural integration with how your business already operates. When intake turns into informed, automated triage, legal stops being a bottleneck and becomes the connective tissue of growth—moving in step with the business, with clarity and trust baked in.