Skip to main content
LITT AGENT FACE · SUPERVISED AI TEAMMATE

Insource your legal front line.

Agent Face is LITT's supervised AI teammate. It picks up every routine legal request — NDAs, vendor agreements, policy questions — across Slack, Teams, WhatsApp, email and your portal, runs it end-to-end on your playbooks, and routes only the genuine judgment calls to a lawyer.

Sign up free
DPDP-aligned·India data residency·SOC 2 · ISO 27001·Decision Trace on every action
Built for
Agent Face · Live front line
Intake
Triage
Outcome
Intake · channels
SlackTeamsWhatsAppGmailPortal
Incoming · Slack · #sales-ops
NDA for the Acme pilot
In queue
WhatsApp · Sales
Can we share customer data with a vendor?
Gmail · Procurement
Vendor MSA — TechCorp redline
+ 24 more today~40 / day
Agent Face · triage
Detected intent
NDA review
Context pulled
Playbook v4Approved sources#sales-ops
Matched task
Triage & routeReview & redlineDraft documentObjection handlingResearch & answerCustom promptBespoke workflow
·Identified intent
·Pulled playbook context
·Matching task · scoring
97%
Confidence
Gate at 85%
Within gate → auto
Outcome · under your gate
awaiting intake…
Resolved · today
NDA — Northwind pilotResolved
DPA — cloud vendorResolved
FEMA — ODI structuringEscalated
1,286
Handled this month
37
Escalated to counsel
4 min
Median resolve
62%
Outside-counsel saved
The problem

Your legal team isn't slow. It's overwhelmed.

Thousands of low-complexity, high-volume requests — NDAs, vendor agreements, policy questions, DPDP and GST queries — bury the high-value work and slow the whole business down. Tools that merely assist don't remove the bottleneck. Tools that merely assist don't remove the bottleneck.

Today · without Agent Face
Triage & routewaiting
Review & redlinewaiting
Draft documentswaiting
Negotiate & objectionswaiting
14in the backlog ↑ & rising
Business teams ping lawyers on Slack and email
Counsel manually draft, review, redline and file
Routine work builds a backlog and a queue
You pay outside counsel for volume work
With Agent Face
Auto-resolved · Decision Trace
Triage & route
returned in minutes
1,284cleared · backlog 0
Requests arrive through the channels teams already use
An agent picks them up automatically — 2am, quarter-end, no backlog
It executes on your playbooks, end-to-end
Lawyers see only the flagged items — five minutes, not sixty
The difference

An agent, not an assistant.

Assistants help one person do one step faster — and stop at every exception. Agent Face owns the whole request, and only stops for genuine judgment calls.

A chat assistant
Ask
Draft
Exception

Helps a lawyer draft or search a little faster. Assists individual steps, hands back the work, and stalls the moment something is non-standard. You still drive every request.

LITT Agent Face
Intake
Execute
Review
Sign
Done

Owns the request end-to-end — intake, drafting, review, negotiation, signing and filing — on your playbooks. It escalates only the genuine judgment calls, with a Decision Trace attached.

How it works

A request arrives. The agent goes to work. A lawyer reviews what matters.

01

Intake & routing

Requests land embedded in Slack, Teams, WhatsApp and your portal. Agent Face reads the intent and pulls the relevant context.

Classify request type
Read intent
Pull context & playbook
02

Execution

It runs the right task end-to-end against your playbooks, templates and fallbacks — drafting, reviewing, redlining, negotiating.

·Draft / review
·Apply fallbacks
·Flag exceptions
03

Supervision & send

Output lands in a review queue with flagged items and reasoning. Approve, adjust or escalate — then it signs and files.

·Review queue
·Approve / escalate
·Sign · file · learn
What it handles

Routine legal work, on autopilot.

Agents are built from configurable task types — trained on your playbooks, not generic data.

Task · Legal
Triage & route

Classify every request, extract key details, and route to the right workflow or lawyer — filtering misdirected items.

resolved · Decision Trace attached
Supervision

Autonomy, under your gate.

Supervision is the whole point. Turn it on and off — by agent, by task, by threshold. Start fully supervised, then dial it back for routine requests. Every decision, edit and negotiation is logged in a Decision Trace.

The supervision gate · drag the threshold
3 auto-resolved5 to review
gate 85%
auto-send →↓ review queue
Items below the gate route to Your own lawyers. Higher confidence over time → dial the gate down.
Supervision by agentSupervision by taskApproval thresholds by riskEscalation thresholdsFull Decision Trace audit
Lives where your teams work

Embedded in the tools you already use.

Nothing to adopt. Agent Face meets requests on the channels your business already lives in, and connects to the systems you already run.

Channels · requests in
Agent
Face
Routing
Acts on · systems
A request lands on Slack → Agent Face triages & executes → writes back to CLM.
5 min
to review, not 60 to draft
1,200+
requests handled / month
62%
less outside-counsel spend
2–4 wks
to launch your first agent
Built for Indian regulated environments

Enterprise security. DPDP by design.

Your request
Files, queries, PII
Encrypted India tenant
AES-256 · region-locked · isolated
Discarded after session
No retention · no model training
Data residency in India

Dedicated regional tenant. Your matters stay inside their home jurisdiction.

Zero model-training retention

Your files and queries never train base models, and aren't stored beyond the session.

AES-256 · TLS 1.3 · SSO + RBAC

Encrypted in transit and at rest, with least-privilege access and single sign-on.

SOC 2 · ISO 27001 · DPDP

Independently audited, with a full Decision Trace audit on every action.

Questions, answered.

Does Agent Face replace our team?

No. It absorbs the repeatable, high-volume work so your team focuses on real risk and judgment. Your reviewers always retain final authority — nothing leaves without sign-off when supervision is on.

How accurate is it, and how do you control errors?

Every agent is trained on your playbooks and approved sources — not generic data. Low-confidence or non-standard matters escalate automatically, and every action carries a Decision Trace you can audit.

Where does our data live?

In a dedicated India tenant, encrypted in transit and at rest. Your files and queries are never used to train base models and are not retained beyond the session.

How long until we're live?

Most teams build and launch their first agent within days to a few weeks — starting fully supervised, then dialling supervision back per agent and task as confidence grows.

Put your legal front line on autopilot.

Insource the routine, escalate the rest to a lawyer. See Agent Face run your legal requests end-to-end — under your gate.

Sign up free