The engagement closes, the final invoice goes out, and your revenue from that client returns to zero until somebody sells the next engagement.
A new matter, a new audit, a new project, a new search: each one takes a sale, a scope, a staffing decision, and a ramp-up period before the firm earns a dollar. When the engagement ends, the counter resets.
This is the episodic model. It has been the default in professional services for decades, and it worked well enough until AI compressed the delivery time on each episode and the revenue per engagement fell with it.
Firms across law, accounting, and consulting are shifting to continuous delivery, and their revenue no longer resets to zero.
The problem with episodes
The episodic model carries costs that most firms accept as facts of life but should not.
Every new engagement requires selling the work. Even with existing clients, there’s a scoping conversation, a proposal, a sign-off. That overhead consumes partner time, delays revenue, and creates gaps where you earn nothing from that client.
Each new episode requires rebuilding context. The team reviews prior work, re-learns the client’s situation, asks questions the client already answered last time. Someone pays for that ramp-up, the client or the firm. It’s waste either way.
And episodic revenue is lumpy: tax season, litigation waves, M&A cycles. Your revenue spikes and dips based on when clients happen to need you. Planning hiring and cash flow around that volatility is expensive and imprecise.
AI makes all of this worse. When delivery time compresses, each episode generates less revenue. You need more episodes to hit the same number. The selling overhead and ramp-up costs stay fixed. The math gets harder every quarter.
An engagement earns once. Selling the next one, scoping it, and getting your team back up to speed on the client all cost the firm again, before a dollar comes in.
Continuous, by vertical
Instead of waiting for the client to call with a problem, you’re embedded. You monitor, analyze, and surface findings on an ongoing basis. The client pays monthly. The relationship doesn’t end when the project does.
This is already happening in accounting and staffing, and the same mechanics carry over cleanly to consulting and law.
In accounting, firms running the AICPA’s Dynamic Audit Solution feed client transactions into an AI engine and update the analysis continuously rather than once a year at audit time. They turn the annual engagement into an always-on advisory subscription. The 2024 CPA.com/AICPA Client Advisory Services Benchmark Survey found that firms earning significant revenue from CFO-level advisory work post more than 30% higher monthly recurring revenue than firms without it.
Mastercard runs TEAM, its Talent Excellence Always-on Model, which McKinsey profiled in April 2026: candidates, including people not actively job hunting, get engaged on an ongoing basis instead of the firm starting from zero once a requisition opens. A staffing firm that copies the model embeds itself in how the client manages talent, not only in how they fill seats.
Consulting firms can use the same structure for advisory work: a base fee for steady access to the firm’s judgment, a usage-based component for work beyond it. The firm earns predictable recurring revenue instead of chasing the next project sale, and the client gets ongoing access to expertise instead of buying it in expensive one-off engagements.
The same shift fits regulatory surveillance in law firms: a monthly retainer buys standing coverage, AI runs the scanning across new filings and rule changes, and attorneys step in for the judgment calls when something surfaces. The client pays for visibility they didn’t have before, and the firm earns twelve months of retainer instead of one engagement fee.
Why the economics are better
This is more than a different billing arrangement.
Monthly recurring revenue is forecastable. You can plan hiring, investment, and capacity around a number you can see, not a pipeline you hope converts.
A client paying a monthly retainer for 36 months generates more total revenue than a client who engages you for two or three projects over the same period. That’s where the AICPA advantage comes from: more months of revenue per client, and a lower re-acquisition cost for every dollar earned.
The ramp-up cost disappears because your team already knows the client’s situation. They keep the context instead of rebuilding it, AI runs the monitoring at near-zero marginal cost, and your people spend their hours on judgment and client conversation rather than re-learning what they already knew.
And the relationships are stickier. A firm embedded in how a client operates is hard to displace. The competitor has to replace your expertise, your ongoing access to the client’s data, and the monitoring infrastructure you’ve built around their business. Episodic relationships have almost no switching cost. Continuous ones have a lot.
The bottleneck is not technology
The tools for continuous delivery already exist: real-time data ingestion, anomaly detection, automated monitoring, AI-powered analysis. Most firms already own them.
What’s missing is the commercial decision. Moving from episodic to continuous means rethinking the offering itself, what the service is, how it reaches the client, what the client pays, and when. That’s a business model question, and the pricing model has to change alongside the delivery model.
It also changes the client conversation. An episodic engagement sells a defined scope of work. A continuous relationship sells ongoing value, so the proposal is different, the pricing is different, and the client is agreeing to a relationship rather than approving a project.
Most firms haven’t had that conversation. Plenty have renamed the old work advisory and left the model exactly where it was.
Where the counter stops resetting
Most firms still deliver work in episodes. AI keeps shrinking the revenue inside each one, and the cost of selling the next one has not moved. A firm on a monthly retainer earns in the months when nothing goes wrong, and reaches the client first when something does. Its revenue starts each January at last year’s number instead of zero.