Enterprise software vendors built their pricing models around a simple assumption. More people in your firm needed access, so you bought more seats.

That model is starting to crack.

AI agents can now complete work that once required a user logging into a practice management system, legal research platform, document repository, intake tool, or CRM. An agent can classify an enquiry, update a matter record, prepare an intake brief, search a document set, and route work to the right person. It doesn’t need a seat in the traditional sense. It needs governed access to the right systems and a clear workflow.

That creates a problem for vendors that charge law firms per lawyer, per assistant, per administrator, or per occasional user. It also creates an opportunity for managing partners and firm owners.

If your firm is spending $1M to $25M in annual revenue, there is a fair chance you are paying for software capacity you don’t use. Some licences sit with lawyers who only log in a few times a month. Others are assigned to staff who use one function but are bundled into a much larger plan. Some platforms are retained because they cover a manual process that an AI agent can now handle.

This isn’t an argument to cancel every legal technology subscription. It is an argument to stop treating software renewals as an administrative task. Your practice management, legal research, intake, document automation, e-signature, and CRM contracts are now operating decisions.

For many firms, an honest review exposes $80K to $250K in annual leakage across unused seats, overlapping platforms, manual work, and lost matters. The bigger gain often comes from changing how work flows through the firm.

Why per-seat software pricing no longer fits legal work

Per-seat pricing made sense when software was a destination. A lawyer opened a platform, did research, recorded time, updated a matter, created a task, or reviewed documents. The vendor could reasonably count active users and charge accordingly.

AI agents change the unit of work.

An agent can operate across your systems based on an event. A prospective client fills in a form. A new email arrives in a shared inbox. A discovery batch is uploaded. A contract lands in a matter folder. The agent receives the event, performs a defined task, records the outcome, and routes the next action.

That is not seat-based work. It is usage-based work.

The shift is already visible in how firms use their technology. A junior team member may have access to a legal research tool but use it only for occasional first-pass queries. A practice manager might hold a full CRM licence mostly to assign leads. An intake coordinator could spend hours copying call details into a matter system. Those tasks are not best priced as permanent, named-user access.

They are workflows. You should measure them as workflows.

A useful question for every software renewal is this:

Are we paying for people to access this system, or are we paying for outcomes the system helps us produce?

If it is the second, you have room to renegotiate.

That doesn’t mean asking a vendor for a discount and accepting the same commercial model. It means bringing actual usage data to the conversation. How many matters were opened? How many intake records were created? How many documents were reviewed? How many research queries were run? How many users logged in at least weekly? How much activity was generated by an automated workflow?

Firms that can answer those questions can push for lower tiers, pooled licences, usage credits, API access, or a different bundle. Firms that cannot answer them usually renew what they had last year.

Start with the work, not the vendor list

A software audit often begins with invoices. That is useful, but it is not enough.

Start with the work your firm performs every week. Map the work from first contact through matter delivery and billing. Then identify who touches each system, what they do there, and whether that activity needs a human.

Consider client intake.

A potential client calls at 7:15 p.m. about an employment dispute. In many firms, the call goes to voicemail or an external answering service. The caller leaves partial information, if they leave anything at all. The next morning, someone listens to the message, calls back, asks the same questions, checks for conflicts, enters information into the intake platform, and then tries to find a suitable consultation slot.

That process may involve a phone system, an intake system, a CRM, a calendar tool, a practice management platform, and an email account. Multiple staff members may need access to several of those tools. Yet the actual job is straightforward. Capture the enquiry, identify the matter type, run an appropriate preliminary conflict check, collect contact details, record the urgency, and book the next step.

The same pattern appears in document review.

A litigation team receives a batch of discovery documents. A junior associate or paralegal opens documents, tags them, identifies potential relevance, summarises key points, and escalates significant material. The firm pays for document review licences and spends associate time managing the first pass. Associate time in the $200 to $400 per hour range is too expensive to use as a basic sorting engine, especially when review volumes spike.

The point is not that an agent replaces legal judgment. It doesn’t. The point is that much of the preparation, sorting, extraction, and routing work has never required a lawyer’s full attention.

This is where Omni Ops is useful. It is built around operational agents that take defined work from inbox to outcome, while retaining review points and accountability where the firm needs them.

The subscription audit law firms should run now

There are four practical steps here. None requires a six-month transformation program.

1. Build a clean software and access register

List every system that supports client acquisition, matter delivery, knowledge work, administration, finance, and reporting. Include the monthly or annual cost, contract renewal date, number of licences, named users, actual active users, integrations, and the process each tool supports.

Don’t let the list stop at major vendors. Smaller subscriptions create a lot of quiet overlap. Firms commonly carry separate tools for forms, appointment booking, e-signatures, transcription, document generation, task management, internal chat, and AI assistants.

For each platform, separate users into three groups:

• Daily users who need direct access
• Occasional users who may be able to share a pooled model or use an assisted workflow
• Users whose work can be triggered and completed by an agent

The third category is where the commercial conversation changes.

2. Measure activity over 90 days

A 90-day window is usually enough to cut through assumptions without being distorted by a single quiet week. Pull login reports where they exist. Review matter creation, document uploads, research activity, task completion, and usage by role.

Then compare this to the licences you pay for.

A law firm might find that 38 people have access to a practice management platform, but only 19 use it weekly. It may find that six people hold expensive research licences while two are responsible for most searches. It may find that the intake software is primarily used to store information after a human already collected it.

Don’t use this data only to identify inactive users. Look for workarounds. If lawyers are emailing assistants to update matters because the system is cumbersome, the licence count won’t reveal the real cost. The workflow map will.

3. Calculate the manual labour wrapped around each platform

This is the part vendors rarely highlight.

Ask how much time is spent moving information into and out of each system. Include checking shared inboxes, updating statuses, assigning tasks, summarising documents, creating first drafts, copying intake details, and chasing missing information.

Your lawyers may lose four to six hours a week to work that doesn’t reliably make it onto a billable invoice. Not all of that can be automated, and not all of it should be billed. But a meaningful portion is avoidable administration.

A 12-lawyer firm with even four unbilled administrative hours per attorney each week is giving away a material block of productive capacity. That capacity can be redirected into client work, faster response times, or a smaller support burden.

4. Renegotiate from a position of operating clarity

When renewal time comes, do not ask, “Can you reduce our price?”

Ask better questions.

Can our firm use pooled licences for occasional users? Can we move from named seats to usage credits? What does API or service-account access cost for controlled automated workflows? Can we separate the modules we use from the ones bundled into our plan? Can we keep a lower number of full licences and use an agent for repeatable administrative work?

You also need to check your contract language. Some vendors may restrict automation, bot access, data export, or API usage. Others may charge separately for integration access. Get clarity before you build a workflow around a platform.

The objective is not to force every vendor into a new pricing scheme. It is to make sure your firm doesn’t pay premium user-based pricing for work that is increasingly event-based and automated.

What an AI agent looks like in a law firm

An agent should not be a vague chatbot sitting on your website. It should have a specific job, a defined trigger, approved actions, a record of what it did, and clear escalation rules.

Here are three examples.

Intake Voice Agent handles the first response

The Intake Voice Agent answers every call, including after-hours, over lunch, and on weekends. It identifies the caller, captures the matter type, asks approved intake questions, performs a preliminary conflict-check workflow, and books an appropriate consultation directly into the firm’s calendar.

If a caller has an urgent family law issue, the agent can recognise the urgency rules you set and route the matter accordingly. If the matter falls outside your accepted practice areas, it can respond professionally and provide the next approved action. If a possible conflict is identified, it stops short of any representation statement and sends the issue for human review.

The agent records the interaction in your intake process, then gives the responsible person a concise brief. No voicemail transcription sitting unread. No repeated questioning. No high-intent prospect waiting until the next business day.

You can see where this fits within Omni Voice. The commercial effect is not just fewer missed calls. It can reduce the number of people who need broad access to every intake and calendar tool simply to perform basic call handling.

Matter Triage Agent creates a usable brief

The Matter Triage Agent monitors incoming forms and shared email addresses. It classifies the practice area, scores fit against your criteria, identifies missing details, and routes the enquiry to the right partner or team with a one-paragraph brief attached.

For example, an email about a commercial lease dispute arrives with attachments. The agent extracts the client details, identifies the stated issue, notes any deadlines, flags the relevant practice group, and creates a recommended next action. If the sender has not supplied essential information, it asks an approved follow-up question.

A partner doesn’t need to read every raw enquiry to decide what happens next. They see a standardised summary, relevant documents, and the basis for the routing decision.

That protects response time while giving the firm a more consistent view of its pipeline. It also means your CRM or practice management system becomes the system of record, not the place where staff manually reconstruct an enquiry after the fact.

Document Review Agent manages the first pass

The Document Review Agent performs first-pass review across contracts, discovery batches, and matter files. It can flag clauses, summarise positions, identify dates and entities, group similar documents, and produce an associate-grade memo for review.

For a commercial contract review, the agent can identify limitation of liability provisions, indemnities, termination clauses, governing law, renewal dates, and non-standard obligations. For discovery, it can help surface documents related to defined issues, create summaries, and identify potentially significant communications.

The supervising lawyer remains responsible for legal advice, judgment, privilege decisions, and final work product. The agent accelerates preparation. It gives the associate a structured starting point instead of a blank screen and thousands of pages.

This is also why a software audit must include document review spend and labour cost together. A firm may pay for review capacity while also paying skilled people to do repetitive first-pass work. The right answer might be fewer licences, a different platform tier, or better integration between the tools you retain.

Protect confidentiality and professional responsibility

Law firms cannot treat AI adoption as a generic productivity project. Client confidentiality, privilege, data residency, supervision, record retention, and conflict obligations all matter.

That is why the workflow should be designed before it is deployed.

Define what sources the agent can access. Set rules for what data it can write back to systems. Require human approval for actions that create legal, financial, or client-facing commitments. Maintain activity records. Test workflows against real edge cases, including incomplete instructions, conflicts, sensitive matters, and documents that contain contradictory information.

You should also be clear about where data is processed and how vendors handle it. Your firm’s existing engagement terms, regulatory obligations, insurer expectations, and client requirements may all affect the design.

The goal is controlled delegation. You are not handing professional judgment to a machine. You are removing the repetitive handling that prevents professionals from applying judgment where it matters.

Use intake as your first commercial test

If you want a practical starting point, use our AI Client Intake Checklist for Law Firms to map every intake question, handoff, conflict step, response target, and calendar action. You can also access the direct downloadable checklist for use in your next partner or operations meeting.

Intake is a strong first test because the workflow is visible and measurable. You can track call answer rates, speed to response, consultation bookings, qualified matter conversion, and staff hours removed from copying information between systems.

For many firms, after-hours intake is the most obvious weakness. If 30% to 40% of after-hours enquiries fail to convert, the issue isn’t usually lead generation. It is speed, follow-up, and consistency. An agent can address those points without asking a partner to be permanently available.

Once that process is stable, move to email triage or first-pass document review. Build one governed workflow at a time.

This is where an outside review can help. See Omni for law firms if you want to identify the workflows, systems, and commercial terms that are creating avoidable leakage.

What an Omni Audit delivers in 60 minutes

An Omni Audit is not a software demonstration and it is not a generic AI strategy deck.

In a 60-minute working session, we look at your operating reality. That includes where enquiries are lost, where lawyers are spending unbilled time, where staff are rekeying information, and where software licences no longer match the work being done.

You leave with three useful outputs:

1. A prioritised workflow map showing where an agent can remove manual handling
2. A software and subscription review lens for seats, usage, overlaps, and renegotiation opportunities
3. A practical implementation path that identifies what to build first, what should remain human-led, and what controls are required

There is no need to prepare a deck. Bring your current stack, a rough view of licence counts, and two or three processes that frustrate your team.

If your firm is heading into a renewal cycle, this work should happen before the renewal quote arrives. If contracts have already renewed, it can still shape the next 90 days and give you the data needed for the next negotiation.

Book a 60-min Omni Audit and we will identify where your firm is paying for seats when it should be paying for outcomes.

The firms that act early will have better leverage

AI agents do create a difficult pricing problem for software vendors. Their old model depends on growing the number of human users. Your firm does not need to inherit that problem.

Your responsibility is to make sure the technology budget reflects the way work now gets done.

Keep full licences where lawyers and staff genuinely need rich access. Reduce or pool access where usage is occasional. Replace repetitive data movement with governed workflows. Insist on pricing that reflects matters processed, documents reviewed, or actions completed when that is a better fit than named seats.

The immediate benefit is lower waste. The more important benefit is a firm that responds faster, gives lawyers more usable time, and has a clearer picture of what its technology is actually doing.

For a closer look at the operating model, visit the AI audit for law firms. If you are ready to put numbers against the opportunity, Book my Omni Audit.