Once a client works inside your workspace, switching stops being about the relationship.
Start a conversation with the AI Adoption Concierge, already scoped to client-facing AI. Pick a starting point, or describe your situation directly.
The most strategically interesting development in this area is not a tool a firm uses internally — it is a category of product that lets a firm expose AI workflows and document sets to its own clients, under the firm's brand, with the firm's lawyers validating output. Two of the leading legal AI platforms shipped versions of this within days of each other in late 2025, and at least one design allows a client to use it without holding a licence of their own. The strategic observation made at the time is the one to internalise: clients may start choosing firms partly because of the workspace they get. That is a genuinely new form of switching cost for law firms, and unlike relationship-based retention it does not leave when a partner does. It also carries risks that mostly have no bar guidance yet.
From lowest commitment to highest, and most firms should stop well short of the end.
The lowest-cost version. A diagram, a register, a searchable summary — an artefact the client returns to rather than a document they file.
Clients submit questions; the firm's lawyers validate the cited answers before they go back. The human gate is the product.
Collaboration on document sets and reviews with matter-level permissions, so the working surface is shared rather than emailed.
The firm's workflows exposed under the firm's brand, in some designs without the client needing a licence.
Answering procedural questions from firm-approved material only. Guardrail design is the entire exercise.
The far end. A handful of very large firms have done it, including with revenue-sharing arrangements. Not a small-firm strategy.
How to approach this without creating exposure.
Mostly where the relationship lives, which is a different question from how good the work is.
Relationship-based retention walks out with the partner. A client whose working process lives inside your workspace is held by something else entirely. That is why the first-mover argument here is stronger than for anything else in this department — and why a mid-market firm should at minimum understand what its competitors are showing clients.
It depends entirely on what it does, and the profession has not settled the boundary. A tool that retrieves documents, answers procedural questions from firm-approved material, or surfaces cited passages for a lawyer to validate is a long way from practising law. A tool that answers a client's substantive question without a lawyer in the loop is much closer to the line, and the line itself is jurisdiction-specific. The consistent design answer is the human gate: the firm's lawyers validate anything substantive before it reaches the client, which is also what makes it worth paying for. Any actual determination is a question for your own counsel.
The portal products are priced and sold to firms considerably larger than most, and a small firm should not start there. What is realistic at any size is the underlying idea: give the client a structured artefact they return to rather than a document they file. A plan diagram, a searchable obligation register, a matter dashboard, a maintained summary. That achieves a meaningful part of the retention effect without a platform purchase, and it is the sensible first step regardless of eventual ambition.
Three that matter. Cross-matter leakage, where a client sees material from another matter — the failure that ends the relationship and possibly the engagement. Prompt injection, since anything a client can upload is content the firm's AI will read and a client's counterparty may have authored. And scope drift, where a tool built to answer procedural questions gradually gets asked substantive ones because it is convenient. The first two are addressed by permissions enforced at retrieval and by red-teaming before launch; the third is a governance problem and needs a defined boundary written down.
Some will, and it is a reasonable request rather than an insult — corporate legal teams have been explicit that a substantial share expect to depend less on outside counsel as they build internal capability. Firms have responded across a spectrum: some license tooling to clients, at least one large firm on a revenue-share with its vendor; some use the workspace as a retention device without unbundling the advice; and some decline. The strategic point is that this conversation is now happening whether or not a firm has a position, and having one is better than improvising it in a pitch.
Describe your client base and what you deliver today. The Institute will help you find the realistic version.