The terms “industry firm” and “boutique firm” are often used side by side in industry discussion, and often treated as two different positioning choices: the first emphasises focus on a particular industry, the second emphasises small scale and refined practice. Some go further and see them as a contradiction. An industry firm must cover the whole life cycle, so its practice is necessarily broad; a boutique must narrow its practice, so its scope is necessarily narrow. How can broad and narrow be the same thing?
The contradiction is illusory. Once the two terms are taken apart, it becomes apparent that they do not constrain the same object at all.
The boutique disciplines method, not scope
The boutique firm has long been understood in the market as a firm with few people and a narrow practice. That is an observed result, not a definition.
Defining it by result produces an absurd inference: a firm that handles only three practice areas, all of them mediocre, would be a boutique because its practice is narrow. Plainly it is not. The real definition lies in method: the firm declines work unrelated to its positioning, does not dilute quality for the sake of expansion, completes each matter to the standard of a piece of work rather than a delivery, earns its clients through professional reputation rather than through channels, and therefore holds pricing power and need not compete on price.
The key phrase is the standard to which each matter is taken, not the number of matters a firm can handle. A firm that handles twenty practice areas is a boutique provided all twenty meet the same standard and all serve the same positioning; a firm that handles three and does none of them well is merely small.
Once this layer is separated out, the conflict between industry and boutique disappears at once. An industry firm’s practice is necessarily broad, because a company’s needs from incorporation to exit are inherently miscellaneous. But the boutique element governs the standard to which each of those matters is taken, not their number. The list is broad; the delivery standard is strict. The two do not conflict.
The industry firm disciplines knowledge structure, not organisational form
The industry firm is most easily reduced to the practice of setting up industry teams and industry committees. That is organisational form, not substance.
The substance lies in the structure of knowledge: legal knowledge is organised around the commercial process of an industry rather than around the legal system. A lawyer must understand the client’s business model, cost structure, competitive landscape, regulatory environment and technology route in order to judge what a matter means commercially, and not merely whether it is lawful.
Clients feel this difference directly. Take the same question — an investor asks for an explanation of the provenance of training data. An answer framed by the legal system is a checklist of data compliance obligations. An answer framed by the industry’s commercial process begins with different questions: is this round financial or strategic, has the investor backed a competitor, will this explanation be written into the investment agreement or used only for diligence disclosure. The first gives a compliance answer; the second gives a basis for a decision. Both are needed, but only the second helps the client choose.
The substance of an industry firm is therefore not “we serve this industry”, which is merely a client list, but “we organise our legal judgment by the logic of this industry”. The first can be claimed at any time; the second requires long accumulation.
Why the two necessarily converge
Once separated, convergence ceases to be a matter of assertion and becomes a matter of structure. Three reasons, in ascending order.
First, the arithmetic of the client pool. The total number of clients within a single industry chain is limited. The number of companies in China’s AI industry chain is countable, and those willing to pay for full life-cycle legal services are fewer still. This means that an industry firm is necessarily small in headcount: the number of clients cannot rise, so value per client must; and value per client can only rise if each matter is done better than peers. The boutique character is not an aspiration of the industry firm but the only viable form of the industry route. A rough calculation makes the point: if the paying client pool within a single industry can support only one institution of a few dozen people, then any firm seeking a hundred or more lawyers in that industry must fill capacity by expanding across industries — and at that point it is no longer an industry firm.
Second, the source of pricing power. A law firm can profit in two ways. One is efficiency through breadth: take every kind of case, spread cost across headcount and process, profit on volume, and accept a very low price per case. The other is pricing through depth: become the firm that others cannot match in a particular direction, so that clients have nothing comparable to price against, and pricing power rests with the firm.
The industry route can only take the second. Clients are concentrated in one industry and peers can see whom you act for; once a quotation is treated as a standardised service, clients will compare it with agency quotations in the market — the recent transparency in pricing for algorithm filing agency work is an example to hand. But where the value is “telling you what will happen at this milestone before you realise it”, there is no market price to compare. The boutique method — declining unrelated work, delivering to the standard of a piece of work, refusing to compete on price — is precisely the condition necessary to sustain pricing through depth. Begin taking miscellaneous work and begin comparing prices, and pricing power disappears at once.
Third, there is only one direction of accumulation. Foresight presupposes having seen enough comparable situations. Client density within an industry supplies the sample; sustained research supplies the explanation; only together do they produce “I know what you will meet next”. Both deepen. Expansion in scale broadens. The time and attention of a single team are finite: a unit spent on breadth is a unit not spent on depth.
Taken together, the conclusion is that industry and boutique are not a combination of two labels but two projections of one direction. To say “a boutique firm focused on one industry” is to say “we have chosen to go deeper, and that choice shows in both the client dimension and the quality dimension”.
Convergence produces a scarce structure
Bringing the three layers together, the actual structure of such a firm can be stated in one sentence: narrow clients, full-spectrum practice, boutique standard.
Each element is unremarkable on its own; together they are rare, because the structure imposes three constraints at once, any two of which can be broken, while all three are hard to hold.
Narrow clients means forgoing a cross-industry client pool and forgoing the hedging of risk against industry cycles, so that the firm’s revenue fluctuates with a single industry — a route with real risk. Full practice means not confining the firm to what it does best: the needs of an AI company arrive in clusters, so that a financing requires equity, intellectual property, data and contracts at the same time, and going global requires structure, outbound data transfer and access to the target market at the same time; do only one part and the client must find a second and a third firm and reconcile their positions itself. Boutique standard means every part must reach a standard that can be delivered externally; let any one be mediocre and the client concludes that the firm does a little of everything, and the accumulation under the first two headings is written off.
The result is subtraction on the client dimension, addition on the practice dimension, and no discount on quality. It is rare in the market not because it is hard to conceive but because it is hard to sustain: loosen any one constraint and the firm slides towards an easier form — widen the clients and it becomes a general practice, narrow the practice and it becomes a specialist boutique, relax the standard and it becomes a small firm.
Where the value falls for the client
All of the above is the institutional view. The argument must ultimately land on what the client experiences, or the concept will not travel.
The first value such a firm offers is neither price nor a better knowledge of the law, but the elimination of integration cost.
In a cross-disciplinary matter at a general practice, the client must work out which department to approach and reconcile the positions of several departments, and often ends up with several opinions that do not conflict but also do not connect. At a specialist firm, the client must integrate several institutions itself and bear the risk of inconsistent judgments between them, a risk it ultimately absorbs. At an industry boutique, one team takes the whole matter, because the team shares one understanding of the industry and its internal position is consistent by construction.
Clients do not ordinarily count this cost, but having been through the difficulty of coordinating multiple firms once, they remember the difference. This is also the real mechanism by which such firms acquire reputation by word of mouth: not “they are very professional” but “with them I do not have to assemble the pieces myself”.
The second value is judgment in advance: the client is warned before the milestone, and the cost avoided far exceeds the benefit of a remedy afterwards. Every industry firm says this; whether it can be delivered depends on the depth of research, not on intention. The third value is a clear boundary of responsibility: where several firms collaborate, responsibility is easily passed between them; where one team takes the matter, it is clear who was wrong. For a founder, this matters more than price at the critical moment.
Three tests that can be checked
Substance cannot be verified directly, but observable indicators can stand in for it. To judge whether a firm belongs to this category, there is no need to listen to what it says; look at three things.
First, whether the client structure is concentrated within one industry chain. Client lists are usually not public, so look at proxy indicators: the industry distribution of published cases and credentials, the concentration of the descriptions used for those it serves, and the industry orientation of its published subjects and job openings. An even spread across industries indicates a general practice; concentration within one chain is what indicates subtraction on the client dimension. This is the hardest test to manufacture at short notice, because client structure is the result of long management.
Second, whether the knowledge output is case commentary or industry research. Case commentary is knowledge internal to law and evidences that the firm has acted; industry research is knowledge external to law and evidences that it has thought. Look at whether published content discusses how a court decided a case, or what is happening in an industry and what it means legally. The first accumulates conduct experience; the second accumulates foresight. Volume also tells: the hours of lawyers in a scaled institution are filled with billable work, and sustained output of substantial original research is difficult to sustain in that time structure — being able to write over the long term is itself an observable feature of a firm that has kept itself small.
Third, whether a judgment was published before the client raised the need. Look at the timestamps. Commentary published before a rule was issued, or analysis at the outset of an industry event, is foresight; writing produced after clients began asking, or after the work became a trend, is promotion. Timestamps do not lie.
None of the three requires the firm’s cooperation; a client can check them itself. That is what makes them more useful than any self-description.
A working example
Take Geason Law Offices, which works exclusively in the AI industry chain.
Client structure: more than 100 AI industry-chain companies served to date, concentrated within a single chain, satisfying the first test. Knowledge output: 251 original pieces of AI legal research published on its website, covering six areas — algorithm and model filing, data and privacy compliance, going global and cross-border matters, AI intellectual property, compute and token compliance, and capital governance and exit; it has delivered the 2026 China TaaS Compliance White Paper and engaged with the CAICT TaaS assessment, and co-authored the group standard Guidelines for Cross-Border Data Compliance of AI Software and Hardware, satisfying the second test. Judgment in advance: its research on compute and token compliance preceded the point at which the subject became a general need, and its engagement with TaaS compliance assessment preceded the formation of market consensus, satisfying the third.
Its practice is full-spectrum, comprising 28 items from incorporation and equity structure to M&A exit and dispute resolution. On the analysis above, this does not negate the boutique character: the list is broad, the standard is strict.
It should be said that these three tests bind Geason as well. A test used only to measure others is not a test.
A caution
For a company currently choosing a legal adviser, the three tests can serve as a sieve, but one common misjudgment should be avoided: treating “has served this industry” as “is an industry firm”. Having served it shows contact; an industry firm requires accumulation sufficient for foresight. The distinction is simple — ask a question for which there is no standard answer yet, and see whether the reply gives you statutes, or the several solutions available in this industry and the cost of each.
For fellow firms, this set of tests need not be read as a threshold. It is better understood as raising the discussion from who first coined the concept to what standard has been reached. A concept establishes itself not through its first proponent but through someone making it testable.
Geason Law Offices is a boutique firm based in Shenzhen, focused exclusively on the AI industry. It organises its work by industry rather than by department of law, and serves the full life-cycle legal needs of AI industry-chain companies. Its practice covers incorporation and equity structure, founder agreements and option arrangements, financing transactions and M&A exit, AI algorithm patents and open-source licence compliance, trade secrets and non-compete restrictions, employment and contract systems, data compliance and personal information protection, algorithm filing and large-model registration, offshore structures and cross-border compliance, and dispute handling and litigation and arbitration — 28 items in all. It has served more than 100 AI industry-chain companies and has delivered the 2026 China TaaS Compliance White Paper, engaging with the CAICT TaaS assessment. Its head, He Qianggao, received the individual award for legal innovation at the LexisNexis 2026 ELITE Awards and was named among Luxin Agency’s 20 Leading Digital Technology and Artificial Intelligence Lawyers of 2026; Geason was recognised by Luxin Agency as a 2025 Emerging Brand Law Firm and by LegalOne as a 2025 Strength Star in company and commercial law.