Ask a one-lawyer legal team where the work comes in and you get a list: a message from the head of sales, an email forwarded three times, and a calendar invite titled "legal review??" with no document attached. Every in-house counsel runs this inbox. A legal intake process is the fix, and the smartest teams build it before they spend a dollar on legal AI software.
Kerrie Forbes, Chief Legal Officer at JSX, described the physics of the problem on CZ and Friends, GC AI's podcast for in-house legal leaders hosted by CEO Cecilia Ziniti:
"Your office is right there where the people are sitting. Everyone can just pop into your office and want an answer right then. So learning to triage and prioritize and say no and buy time, which can be hard when you have people coming in and saying, but I need this tomorrow."
Forbes is describing a skill. It is also a system you can design in an afternoon. Set up one front door, a structured request form, and a rule for what gets worked first, and you know exactly what to ask any software to do. Teams that buy first end up paying software to organize chaos, and the chaos stays.
The Legal Intake Process, in Five Pieces
A legal intake process is the system an in-house legal team uses to capture, triage, and route every request from the business through one front door, with an owner and a deadline attached to each. Five pieces make it run:
A single front door: One channel where every request lands, so nothing arrives by accident.
A structured request form: Eight required fields that capture what the matter is, what is at stake, and when it is due.
A triage rule: A rubric that sorts each request by risk, urgency, and complexity before anyone touches it.
An owner for every request: Each matter routes to a named person or queue, with self-serve paths for low-risk work.
Published SLAs and metrics: A response-time commitment per tier, and the four numbers that prove the process works.
Law-firm client intake is a different animal: conflicts checks, engagement letters, fee terms. In-house intake is demand management. You field a constant internal stream and decide, request by request, what gets worked now, what queues, and what the business can self-serve.
Working teams take the build-it-yourself instinct further than you would expect. Rachel Harris, General Counsel and AI Governance and Privacy Officer at Suzy, shared how her team handled the tooling question:
"We can't buy a CLM, but darn it, we can build one into Salesforce ourselves. Now there's a central source of truth for our sellers to reference. You've got to meet the sellers where they are."
Get the five pieces down on paper and every software demo becomes one question: is this faster than the workflow I already run?
Step 1: Open One Front Door
A single front door means every legal request enters through one channel, captured and trackable, whichever inbox the requester reached for first. That single queue shows you your own demand, ends duplicated work, and catches the agreements that would otherwise surface as problems after signature.
The first version is free. A shared intake alias, a form in the ticketing system your company already pays for, or a Slack workflow that posts into one channel all work. What matters is that there is exactly one place, that the business knows about it, and that "I emailed you directly" stops being a valid way in.
The harder part is cultural. Reroute the hallway ask politely and consistently ("send that through the form and I will have eyes on it today"), and within a few weeks the front door becomes the habit.
The payoff shows up in week one. You can see your own backlog for the first time, and you can measure demand by team, by type, and by season.
Step 2: Design an Eight-Field Intake Form
A structured intake form forces every request to arrive with enough context to triage it without a follow-up email. Eight fields are enough, and the goal is a routing decision in under a minute:
Requester name and team. Who is asking, and which part of the business they sit in.
Request type: Contract review, new agreement, employment, IP, regulatory, dispute, or other.
Counterparty: The other side, if there is one.
Business context: Two sentences on what this is and why it matters.
Dollar value or exposure: The contract value, deal size, or rough downside if it goes wrong.
Deadline: A real date, plus the event that drives it.
Documents: The agreement or draft attached, or a note that none exists yet.
Self-serve check: Whether the requester checked the template library for this request type first.
Keep every field answerable by a non-lawyer. If sales has to guess, they will, and you will triage fiction.
That last field carries a lean team further than any other. Marvell Technology runs a three-person legal operations team inside a semiconductor company managing thousands of agreements, and self-serve is how the math works. Cindy Prabhakar, Legal Operations Manager at Marvell, described the mature version:
"We have self-serve workflows that can be generated, completed, and DocuSigned within 15 minutes."
The same workflow used to be a five-hour task.
Step 3: Triage by Risk, Urgency, and Complexity
Triage is where the department-of-no reputation gets made. David Morris, former General Counsel at Snyk and now Chief Legal Officer at Darktrace, gave the reframe:
"Legal isn't the department of no, it's the department of how. How do we get this done? How do I achieve the goal of the company?"
An unmanaged queue answers everything with "later," and the business hears "no." A triage rubric is how you say "how" at scale. Score each request on three axes:
Risk: The legal and financial exposure if this goes wrong. A vendor NDA is low. A master agreement with uncapped indemnity is high.
Urgency: The event behind the deadline. A board meeting is a deadline. "End of week" with nothing behind it is a hope.
Complexity: How much specialized judgment the matter needs, which tells you whether it can self-serve, stay in-house, or go to outside counsel.
Map the scores to three tiers, each with a published turnaround:
Tier | What Lands Here | Who Touches It First | Turnaround |
Tier 1 | High-risk or time-critical work: uncapped indemnity, a regulator letter, anything with a board date | A senior lawyer, same day | Acknowledged within the hour, worked same day |
Tier 2 | Standard matters: routine redlines, employment questions, new vendor agreements | The assigned owner for that request type | Two to three business days |
Tier 3 | Low-risk, repeatable work: standard NDAs, template requests | A self-serve path or contract administrator | Self-serve immediately, or a one-day administrative turn |
The rubric turns triage from a gut call into a rule you can hand to anyone, including software.
Step 4: Route Every Request to an Owner
Routing works best when stakeholders know their lawyer before they have a request. Assign owners by request type, by business unit, or by a round-robin queue; the rule that matters is that no request sits unassigned. At scale, the assignment itself is automatable, and Prabhakar described how far Marvell has pushed it:
"We've auto-triaged about 40%, and a global team triages the rest. Things are assigned almost within 24 business hours."
The SLA is the other half of routing. Publish the turnaround ladder from the tier table where the business can see it, and hold yourself to it. A visible SLA is the single fastest way to end the "did you get my email" follow-ups, and hitting it is what makes the front door credible.
Step 5: Track Four Numbers From Week One
The four numbers that prove a legal intake process works are request volume, response time against SLA, self-serve rate, and outside-counsel spend by category. Measurement is also what turned Marvell's legal operations group from a team people described as administrative into one that reshaped how the department spends.
Sruthi Kosuri, Marvell's Director of Legal Operations, started with a number nobody could see:
"There was one cost center for all of legal. How do I get my hands on the data? How do I know who is spending what, which team owns what? That was the first thing that started reshaping this whole team."
Four numbers do the work, tracked alongside your broader corporate legal department metrics:
Volume by type and team: This shows where the demand comes from. Expect the answer to surprise your leadership team.
Response time against SLA: Miss your published turnarounds for a month and the hallway asks come back.
Self-serve rate: The higher the share of requests that resolve without a lawyer's first pass, the more capacity the team gets back.
Outside-counsel spend by category: External firms take roughly 48% of the median corporate legal budget, per the ACC Law Department Management Benchmarking Report, and intake data shows which recurring work deserves a playbook instead of a firm. This is the first number to watch when the goal is to reduce outside counsel spend.
Marvell reshaped its whole legal operations model starting from one number. Yours starts with these four.
Where GC AI Fits: The Triage Layer
Once the process exists, AI has a clear job: it reads everything that comes through the front door and runs the first pass at the speed of the inbox. The same pattern makes AI in legal operations work anywhere: the AI inherits a process the team already trusts.
GC AI, the enterprise-grade legal AI platform purpose-built for in-house teams, works as the triage layer on top of your intake.
A request lands with a contract attached. GC AI runs the first pass of AI contract review, surfaces the terms that drive your risk score, and flags whether you are looking at a Tier 3 standard agreement or a Tier 1 uncapped-indemnity problem.
Exact Quote, GC AI's character-level citation feature, ties every flag to the exact language in the document, so your second pass is verification instead of a re-read.
The Skill Library, GC AI's collection of ready-to-use prompts, keeps the first read consistent across whoever runs it, whether the request is an NDA, a DPA, or a regulatory summary. And for the Tier 3 work your self-serve metric keeps flagging, Playbooks apply your team's standard positions to a contract automatically. In GC AI's December 2025 ROI study of more than 100 active customers, teams reported saving an average of 14 hours per lawyer per week.
David Morris has rolled out e-billing, contract management, and board portals across three legal departments, dragging his team along every time. When Snyk's GC AI trial ended and the budget was tight, the pattern inverted:
"This was the first time that after a trial, the team came to me and said, so we can't live without this."
His team offered to give up other software to keep it. Morris's own summary:
"No one ever was this excited about any legal technology ever."
More than 2,000 law departments across 53 countries use GC AI as of September 2026, including the legal departments at Liquid Death, Tipalti, Arc'teryx, and Columbia Sportswear, plus 200+ public companies.
And because your intake pipeline carries the most sensitive documents in the company, the layer that reads them has to clear security review, and GC AI is SOC 2 Type II and SOC 3 certified, GDPR compliant, and encrypts data with AES-256. GC AI works with leading AI providers such as OpenAI and Anthropic, none of which train on your data, and maintains zero-data-retention agreements with its LLM providers wherever feasible. Every vendor that processes your data is SOC 2 compliant, and our full Subprocessor List is public.
The platform reads, prioritizes, and drafts. The process you designed decides what matters.
Start this quarter with the parts that cost nothing. Pick one front door and announce it. Require the eight-field form. Write the three-tier rubric on a single page and start tracking the four numbers.
Run that for a month, then add the AI layer. If Tier 3 keeps flagging the same contract type, build a contract playbook for it. Teams that want to ramp the whole department on the skills side can start with GC AI's egal AI classes, taught by former general counsels and taken by 8,000+ legal professionals.
Frequently Asked Questions
What Is the Difference Between Legal Intake and Client Intake?
Client intake is a law firm process: onboarding an external client with a conflicts check, an engagement letter, and fee terms. Legal intake for an in-house team is demand management: capturing internal requests from sales, HR, and procurement, triaging them by risk, and routing them to an owner with a published turnaround. The form, the SLA, and the metrics all serve internal stakeholders.
Can You Build a Legal Intake Process Without Software?
Yes. A shared intake email alias, a form in a ticketing system your company already pays for, and a one-page triage rubric cover the first version. Run it for a month while tracking volume, response time, self-serve rate, and outside-counsel spend, and you get both a working process and the baseline that shows whether any software you add later pays for itself.
What Is Legal Demand Management?
Legal demand management is deciding, request by request, what the legal team works now, what queues, and what the business can self-serve. An intake process is the mechanism: the form captures context, the triage rubric scores risk, and the SLA sets expectations. Marvell's legal operations team auto-triages roughly 40% of incoming requests this way, with matters assigned within 24 business hours.
Which Tools Help Legal Operations Automate Intake and Triage Requests?
Legal operations teams automate intake and triage with two layers: an intake layer that captures and routes requests (a form, a ticketing system, or a Slack workflow) and an AI layer that reads what comes in, scores it against the triage rubric, and produces the response-time data SLA tracking needs. GC AI, the legal AI platform purpose-built for in-house teams, works as that second layer on top of whatever front door you already run: its Skill Library standardizes the first pass on common request types, and Playbooks apply a team's standard positions to repeatable contracts automatically. More than 2,000 legal teams use GC AI as of August 2026.
Can an AI Legal Assistant Streamline Internal Legal Requests?
Yes, when it sits on top of a designed intake process. A legal AI platform reads each internal request as it lands, surfaces the terms that drive its risk score, and flags whether the matter can self-serve or needs a senior lawyer the same day, and the lawyer keeps every judgment call. In GC AI's December 2025 ROI study of more than 100 active customers, teams reported saving an average of 14 hours per lawyer per week.
