AI for Alternative Fee Arrangements Law Firms

The standard AFA pricing process at most firms looks like this: a partner recalls a vaguely similar matter from two years ago, adjusts for inflation and complexity, and names a number. That number has no connection to actual cost-to-serve data.

You cannot price effectively without understanding your true cost to serve, including non-billable operational work (Legaltech Hub, 2026). The variance hidden inside "average hours" is where profits disappear. A commercial lease review that averages 40 hours might range from 22 to 78 hours depending on counterparty complexity, negotiation rounds, and associate experience level. Price to the average, and half your matters are losers.

The fix is not more spreadsheets. It is auditing the last 24 months of closed matters per matter type, pulling average hours, median hours, and variance, then feeding that normalized dataset into a pricing model. When firms do this properly, the variance picture alone is enough to redesign their scoping approach.

This is also why structuring unstructured legal matter data with AI is a prerequisite, not an optional upgrade. If your matter data lives across disconnected emails, documents, and billing entries with no common taxonomy, you cannot run this audit at scale. The AI has nothing to learn from.

Once matter data is structured, AI can do something spreadsheets cannot: generate profitability probability scores at intake.

Here is the mechanism. A model trained on historical matters learns which signals correlate with cost overruns. Those signals include counterparty type, number of parties, jurisdiction, matter classification, staffing mix, and document volume at intake. When a new matter arrives, the model scores it against those learned patterns and flags whether it looks like a profitable flat-fee candidate or a time-bomb that needs hourly billing or explicit scope carve-outs.

The intervention window this creates is the real value. Predictive analytics can identify matters trending toward overruns early in the lifecycle. At this stage, there is still time to renegotiate scope, add a change-order mechanism, or adjust staffing. If identified only near completion, the firm is simply writing off time.

Ayora, which integrates with BigHand to generate risk-managed quotes, is one example of an enterprise-grade tool built for this kind of intake scoring. AltFee approaches the problem differently, providing over 470 template pricing guidelines alongside historical billing analysis so firms can standardize fee structures across practice groups rather than leaving each partner to price in isolation.

The right approach for your firm depends on whether your pricing problem is a data problem, a process problem, or both. It is usually both.

AI significantly streamlines document review timelines. That efficiency is real. Under hourly billing, it is also a revenue cut.

Firms that bill by the hour for AI-assisted work face a structural problem: the faster they work, the less they earn. Clients will notice, and they will demand rate reductions. Sixty-two percent of legal departments already believe AI will reduce the prevalence of the billable hour (Thomson Reuters, 2026). Firms still clinging to hourly billing on AI-assisted matters are negotiating from a weakening position.

Flat fees flip that equation. When you capture efficiency as margin rather than surrendering it as reduced hours billed, AI investment pays off. Firms adopting value-based pricing tend to experience stronger annual growth than those strictly billing by the hour. That performance gap compounds.

The pricing protection mechanism is a multiplier applied to your normalized historical cost. A common approach: calculate the true average delivery cost for a matter type, apply a 1.3x multiplier to account for variance and margin, and set that as the flat fee floor. This converts previously billable hours into predictable, higher-margin revenue (Legaltech Hub, 2026).

LeanLaw's Fixed Fee Mission Control feature tracks the effective hourly rate on flat-fee matters in real time, letting firms see exactly when internal delivery cost is eating into the client-facing price. That kind of monitoring is not optional. It is how you stay profitable after signing the engagement letter.

Pricing a flat fee is one challenge. Managing the matter to margin once it starts is a separate one, and this is where most firms have no tooling at all.

Casero is an AI-native intelligence layer that connects emails, documents, and case files into a living, case-level knowledge graph. That architecture is directly relevant to AFA management. When every document, email, and obligation in a matter is mapped and entity-extracted automatically, the supervising partner can see the actual state of the matter at any point rather than relying on associate status updates. That visibility is how you catch scope creep before it craters the margin.

Casero's entity extraction pulls out people, organisations, dates, events, and obligations from incoming documents and maps how they relate to each other within the matter. As new documents arrive, the knowledge graph updates automatically. This means a partner managing a fixed-fee commercial dispute does not have to chase down whether a new demand letter from opposing counsel triggers a contractual obligation the firm committed to handle inside the flat fee scope.

The similar cases feature is also relevant to AFA pricing. Casero surfaces past matters based on legislation, factual circumstances, and case classification, with multi-dimensional scoring showing why each case matched. That is exactly the historical comparison data you need to price a new AFA with confidence rather than instinct. Instead of a partner trying to remember a vaguely similar matter, Casero pulls the structured data from the closest actual precedents the firm has handled.

See the Law Firm Matter Lifecycle AI: Intake to Close piece for more on how AI can support the full matter journey, not just the pricing moment.

Seventy-one percent of attorneys now use AI in their work (industry reports, 2026). Only 4% of firms have a formal AI billing policy (Thomson Reuters, 2026). That gap is a liability.

When a lawyer uses AI to cut a 10-hour document review to 3 hours on a flat-fee matter, who captures that efficiency? It goes to margin, which is fine. But on an hourly matter, does the firm bill 10 hours or 3? Different partners will answer that differently, and clients will eventually compare notes.

AI usage is increasingly altering pricing conversations with clients. If those conversations happen without a policy framework, they happen inconsistently. That inconsistency creates malpractice exposure and client relationship damage.

The policy needs to address three things. First, how AI-assisted time is recorded and billed on hourly matters. Second, how efficiency gains are allocated on flat-fee matters. Third, what clients are told about AI use in their matters. These are not technology questions. They are governance questions that happen to be triggered by technology.

Building that governance layer is now a prerequisite for scaling AFA adoption. Firms that skip it will find that AI creates pricing transparency problems faster than it creates pricing advantages. See the Law Firm AI Governance Framework for a practical starting point.

Clients increasingly demand budget-to-actual visibility on their matters (Legaltech Hub, 2026). This is not a request firms can decline. It is a precondition for retaining sophisticated corporate clients under AFA structures.

The firms winning on value-based pricing are not winning because their flat fees are lowest. They win because they can demonstrate exactly where the money went. Reconciled, real-time data showing efficiency relative to budget builds client trust that rate discounts cannot buy.

Failing to provide that visibility pushes firms into margin-crushing rate discounts as the only lever they have left (Legaltech Hub, 2026). The discount conversation is a symptom of a transparency deficit.

Casero's audit trail, which records every action against every document with full source linkage, is part of the infrastructure that makes this transparency possible. When a client asks how a matter was managed, the supervising partner can show exactly which documents informed which decisions, which obligations were tracked, and when. Every AI-generated insight traces back to the original source passage.

For firms thinking about how to make the business case for this kind of investment, the Law Firm AI ROI: Making the Business Case article walks through the numbers in detail.