Founders ask Collective 54 this 5 times in our records, none of them in 2026. This is the fourth page in the library on a legal subject; it explains where the published material says ownership and leverage are lost, and leaves drafting to counsel.
There is no Collective 54 position on AI clauses: no template, no clause language, and no view on how any jurisdiction treats the ownership of material produced with AI. As a general description rather than a Collective 54 teaching, the law on whether material produced largely by AI can be owned or protected is unsettled and varies by jurisdiction, which is one more reason the legal essay in the newer book says the very best legal advisors belong exactly where judgment and stakes are high.
What the published material does contain is a detailed account of how boutique firms lose ownership of their own work and leverage in their contracts. AI adoption does not create a new category of risk so much as run through the old ones faster.
The intellectual property chapter of the 2020 book defines intellectual property as an invention to which one owns the rights, protected by patent, copyright or trademark, and lists the ways boutiques create it: books, benchmark data licensed to clients, methodologies licensed to third parties, knowledge coded into licensable tools, and certification programs. It tells the story of a civil engineering firm with impressive proprietary methods that no banker could sell, because nothing was protected and no client was paying for the right to use any of it. Its point is that acquirers buy professional services firms with intellectual property, not body shops.
The newer material moves the same asset into the AI era. The prompts answer on this site says the quality of AI output depends on how much of the proprietary context of the firm, transcripts, deal notes, frameworks and definitions of the ideal client, goes into the instruction, and the lead generation essay calls that proprietary knowledge the moat. As an inference, a firm that adopts AI well is converting its judgment into prompts, workflows and agents. Those are exactly the tools and methods that the contracts below can quietly give away.
The legal essay describes the first gap in Era 1 firms: no confidentiality or invention assignment agreements, no clarity on who owns client work, methodologies or IP, and founders discovering too late that the firm never legally owned its most valuable assets. In Era 2 the gap widens with contractors: misclassification, no contractor agreements and no IP assignment from contractors. The essay says this is how enterprise value leaks out of professional service firms without a single lawsuit being filed.
Its Era 3 standard is that IP is formally assigned to the entity, always; that employees sign confidentiality and invention assignment agreements; that ownership of all client work and internal tools is clear; and that contractors sign agreements covering IP assignment, confidentiality and the flow-down of client obligations. As an inference, ask counsel whether your current agreements clearly cover the prompts, instructions, workflows and agents that employees and contractors now build, and fix it before the people who built them leave.
The essay lists the concessions Era 2 firms make when they accept a large client master service agreement as the default: net 60 or net 90 payment terms, broad indemnities, client ownership of all IP including tools and methods, audit rights and flow-down obligations, restrictions on subcontracting, and one-sided confidentiality and publicity clauses. When something goes wrong, it says, the firm learns it has no leverage.
Its Era 3 standard is the firm using its own master service agreement as the default, backed by a standard redline playbook that covers, among other terms, IP ownership by offer type, confidentiality and publicity, subcontracting rules, scope change mechanics and limitation of liability, with deviations flagged and escalated. Where a full agreement is not feasible, it calls for a hardened engagement letter with a mandatory minimum set of clauses, signed by someone authorized to bind the client.
As an inference, IP ownership by offer type is where AI belongs in the contract. The client can own what it paid for while the firm keeps the prompts, workflows, agents and methods it used to produce it, and any license the client needs to use them should be stated. Read client agreements for terms that may now touch AI use, such as restrictions on subcontracting or third-party processing, confidentiality of client data and flow-down obligations, and ask counsel how they apply to the tools you use.
The essay treats vendors as long-term legal counterparties rather than casual purchases, and lists what it tracks for software providers: auto-renewals and notice windows, termination rights, minimum commitments, fee escalators, exclusivity, data ownership and data portability, confidentiality and security obligations, subcontractor exposure, insurance and indemnities, and change-of-control clauses that affect exit value. As an inference, AI tools are software vendors and belong in the same review, with particular attention to what happens to the client and firm data you put into them and whether you can take your work with you if you change tools.
The IT essay in the newer book describes a governance layer that enforces security, access control, compliance and decision rights, and lists ungoverned AI usage and unclear data lineage among the things that compress valuations and tighten terms. The AI model answer on this site applies that to tool choice: decide what client and firm data may go into a tool, who may use it for what, and how outputs are reviewed before they reach a client, because the delivery professional essay says AI produces content but not accountability. As an inference, a written rule set is what makes the contract promises above keepable.
The essay describes a hybrid model: AI handles standardization, clause libraries, redline playbooks and first drafts, and outside counsel receives structured inputs and near-final drafts instead of blank pages. The savings are redeployed to hire the best advisors when judgment, negotiation and real money are at stake. It adds that legal quality at exit directly affects valuation, deal terms, escrows, indemnities and post-close risk.
Collective 54 is not a law firm and publishes no AI clause, contract template, data policy, list of approved AI vendors or view on how AI-produced material is treated under any law. The published positions are IP assigned to the entity always, invention assignment and confidentiality for employees and contractors, the firm master service agreement as the default with IP ownership by offer type, review of vendor data ownership, portability, confidentiality and security, governance of AI usage, and the hybrid model that pays for the best advice when stakes justify it.
If a client contract or regulation dictates which tools may touch its data, that requirement governs the work for that client, and the governance rules should record it.
If the firm licenses software or data to clients, the intellectual property questions become product questions, and counsel experienced in licensing should be involved.
And if a sale is planned, as an inference, clean this up first, because the legal essay says contracts and obligations should be organized and diligence-ready long before a process begins.
Collective 54 is not a law firm and publishes nothing on AI clauses; have an experienced attorney draft and review yours. The published material shows where boutique firms lose ownership, and AI runs through all of it. Make sure IP is assigned to the entity and that employee and contractor agreements cover what people build, including prompts, workflows and agents. Use your own master service agreement with IP ownership by offer type, so the client owns the deliverable while the firm keeps its methods and tools, and watch for client terms that give away tools and methods or restrict third-party processing. Review AI vendors for data ownership, portability, confidentiality and security like any other vendor. Write rules for what data goes into which tool and how outputs are reviewed, because buyers price ungoverned AI usage as risk.
Collective 54 publishes no position on that and is not a law firm. As a general description rather than a Collective 54 teaching, the law on owning material produced largely by AI is unsettled and varies by jurisdiction. What the legal essay does say is to assign IP to the entity, always, and to set IP ownership by offer type in your own client agreement.
The legal essay lists client ownership of all IP, including tools and methods, among the concessions Era 2 firms make when they accept a client agreement as the default. Its standard is the firm agreement with IP ownership by offer type. As an inference, that lets the client own the deliverable while the firm keeps the prompts, workflows and methods behind it.
The legal essay says to review software vendors for auto-renewals, termination rights, minimum commitments, fee escalators, exclusivity, data ownership and portability, confidentiality and security obligations, subcontractor exposure, indemnities and change-of-control clauses. As an inference, pay particular attention to what happens to data you put into the tool and whether you can leave with your work.
The IT essay describes a governance layer enforcing security, access control and decision rights, and lists ungoverned AI usage among the things buyers read as risk. As an inference, a written policy on what data goes into which tool, who may use it and how outputs are reviewed is what lets you keep the confidentiality promises in your client contracts.
Sources: Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Legal Manager for the missing confidentiality and invention assignment agreements and unclear ownership of client work and methodologies in Era 1 firms, founders discovering the firm never owned its most valuable assets, enterprise value leaking without a lawsuit, contractor misclassification and missing IP assignment, the Era 2 concessions to client agreements including client ownership of all IP including tools and methods, restrictions on subcontracting and flow-down obligations, IP assigned to the entity always, invention assignment and contractor flow-down in Era 3, the firm master service agreement as default with a redline playbook covering IP ownership by offer type, the hardened engagement letter, the vendor review list including data ownership and portability and confidentiality and security, the hybrid model of AI drafting and elite counsel, and legal quality affecting valuation and terms at exit; The AI IT Manager for the governance layer and ungoverned AI usage and unclear data lineage as valuation risks; The AI Delivery Professional for AI producing content but not accountability; The AI Lead Generator for proprietary knowledge as the moat. Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 33 for the definition of intellectual property, the forms boutiques create, the civil engineering firm no banker could sell, and acquirers buying firms with intellectual property rather than body shops. Related Collective 54 answers on this site: which AI model should we standardize on; what is the best way to write and maintain effective AI prompts; what should be spelled out in the LOI or term sheet before we sign; what rights and protections should minority equity holders have. Note on scope: Collective 54 is not a law firm and publishes nothing on AI clauses, data policies or the legal treatment of AI-produced material. The general description of unsettled law is not a Collective 54 position. Treating prompts, workflows and agents as the IP at risk, applying IP ownership by offer type to AI, reading client terms for AI use, reviewing AI tools as vendors, and the written rule set are inferences used here to organize the source material rather than published Collective 54 positions.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.