Lawyers move into legal tech through several role families, and this page walks through them: legal engineer, product manager, customer success and implementation, knowledge management attorney, and legal operations.
Software companies need people who understand how law is actually practiced, and firms have to run the same work in-house — the tool choices, training and AI policies their ethics guidance calls for.
Here is what each role does, which skills transfer, and how practicing lawyers switch into one.
What “legal tech” covers
Legal tech is the software layer of legal work: the tools that firms and legal departments run on, and the companies that build and sell them.
The category is bigger than the current AI wave — practice management, document automation, e-filing and research platforms are established legal tech in their own right.
What is newer is the profession's own rulebook catching up: ABA Formal Opinion 512 (July 29, 2024) addresses lawyers' use of generative AI tools, tying that use to the duties of competence, confidentiality, client communication, supervision, meritorious claims, candor to the tribunal and reasonable fees.
The sector's main shelves, roughly in the order a practicing lawyer meets them:
- E-discovery and litigation support. Platforms that sort document sets. Opinion 512 names technology-assisted review as a well-known AI use in e-discovery — sorting documents as responsive or non-responsive and separating privileged ones.
- Research, drafting and contracts. Opinion 512 lists the tasks generative AI may help with: legal research, contract review, due diligence, document review, regulatory compliance and drafting.
- Practice and firm operations. Time, billing, matter and document management — the systems a firm runs on with or without AI.
- Legal operations tooling. Spend, vendor and workflow software for in-house legal departments. For legal operations, CLOC, the Corporate Legal Operations Consortium, publishes a “Core 12” framework of competencies.
This page is about the careers on the other side of that table — working for the tools' makers, or running the tool stack inside a firm.
If what you want is to use AI better in the practice job you already have, that is a different question, and our AI tools article owns it.
The roles: legal engineer, product, customer success, KM attorney
Titles vary from company to company; the families below are the ones this page walks through.
Think of them as a different use of your legal training rather than a step down from it.
Legal engineer.
Legal training plus hands-on technical work.
Legal engineers build and configure the tooling — document automation, workflow and intake builds, template systems and, at some companies, prototypes alongside the product team.
How technical the job is varies by employer, which is why two postings with the same title can describe different jobs.
Product manager.
Decides what gets built and why.
Lawyers do well here when they can translate how a practice actually works — the handoffs, the deadlines, the places matters go off the rails — into requirements a build team can act on.
Customer success, implementation and solutions.
The roles that stand between the product and the customer: onboarding, configuration, training sessions, escalations, and feeding what customers struggle with back to product.
Practice experience is the point — if you have sat on the customer side of a vendor demo, you know what a firm actually needs from one.
Knowledge management attorney and practice innovation.
The firm-side versions of this career.
KM attorneys curate the firm's precedents, model documents and research know-how; practice innovation roles redesign how the work gets done — and since Opinion 512 says managerial lawyers must set AI policies and supervisors must make sure lawyers and staff are trained, that work now includes the firm's tooling and AI setup.
These are lawyer roles inside law firms.
Legal operations.
The business-of-law function inside corporate legal departments — CLOC's “Core 12” competencies are one map of it.
It is where tool selection, vendor management and process work come together.
Skills needed (and whether you need to code)
The honest answer on coding: it depends which role you are aiming at, and no single skill list gets you into all of them.
Legal engineer and technical product roles reward hands-on technical comfort.
Customer success, KM and strategy-side product work run on practice judgment, communication and training ability.
The postings say which they want — read several before you decide what to learn next.
What transfers from practice, and what the record actually supports:
- Judgment about tool output. Opinion 512 warns that some generative AI tools are prone to “hallucinations” — ostensibly plausible output with no basis in fact — so uncritical reliance can mislead clients and courts. Vendors selling these tools and firms deploying them both need people who understand that failure mode from the inside.
- Fluency at the profession's own bar. Opinion 512 says lawyers need not become AI experts, but must reasonably understand the capabilities and limitations of the tools they use. That is also the level you can honestly demonstrate in an interview after deliberate hands-on use: functional fluency, not computer science.
- Ethics literacy as a work product. Opinion 512 ties generative AI use to competence, confidentiality, client communication, supervision, meritorious claims, candor to the tribunal and reasonable fees. A lawyer who can turn those duties into a tool policy, a diligence checklist or a training session is doing a job, not reciting one.
- Training and change management. Under Opinion 512, supervision includes training subordinate lawyers and nonlawyers on the ethical and practical use of relevant generative AI tools and their risks. Someone has to design and deliver that training.
One example of why that judgment matters: in Mata v. Avianca (S.D.N.Y., No. 22-cv-1461, June 22, 2023), the court sanctioned two lawyers and their firm for filing non-existent opinions with fake quotes and citations generated by ChatGPT, imposing a $5,000 penalty jointly and severally.
It is the concrete version of the failure mode that the judgment, policy and training work in this market exists to catch — which is why fluency in both the software and the duty is what these roles sell.
Firm innovation teams vs vendors
The legal tech employers this page covers come in two shapes, and the same training sells to both.
Inside firms.
Opinion 512 reads the existing Model Rules as calling for concrete AI governance work.
Under the opinion, managerial lawyers must set clear firm policies on permissible generative AI use, and supervisors must make sure lawyers and staff comply and are trained.
The opinion says client informed consent is required before information relating to the representation goes into self-learning generative AI tools of the kind it describes.
And it applies the same diligence to generative AI providers as to outsourcing vendors: reference checks, vendor credentials, security policies, confidentiality agreements, the vendor's own conflicts checks.
That work has to land somewhere — the firm-side homes for it are the knowledge-management and practice-innovation roles described above.
Either way it is lawyer-shaped work: policy drafting, tool selection, training, diligence.
Inside vendors.
The companies building these products put lawyers to work for the same reasons in reverse.
Their customers are firms whose ethics duties do not pause at the login screen: Opinion 512 points them to diligence on AI providers — reference checks, security policies, confidentiality agreements, the vendor's own conflicts checks — so a vendor selling to firms that follow it should expect that scrutiny.
The opinion also bears on how a tool gets used: under it, client informed consent is required before information relating to the representation goes into the self-learning tools it describes, and a lawyer billing hourly may bill only the time actually spent, even when a tool makes the work faster.
Product and content people who know those constraints from the inside can build and position a product around them.
Outsourcing guidance itself predates AI.
In Illinois, the Illinois State Bar Association's advisory Opinion 19-04 (October 2019) — a voluntary bar's opinion, persuasive rather than binding — notes that the ABA's 2012 Model Rule amendments added outsourcing guidance to the comments to Rules 1.1 and 5.3, and concludes that outsourcing is allowed where it contributes to competent representation with reasonable confidentiality and conflict measures — with client disclosure and informed consent ordinarily required, and always required when substantial responsibility is delegated to an unaffiliated lawyer.
It addresses outsourcing in general, not AI vendors specifically.
Opinions guide; your state's adopted rules bind
How to make the switch
The switch is a story you build in order: what you have used, what you have learned hands-on, how you tell it, and which role it points at.
The steps below run that order — and none of them requires quitting your job first.
Inventory the technology you already touch
List the platforms your practice runs through: e-discovery review, document automation, billing, research. Opinion 512 names technology-assisted review — sorting documents as responsive or non-responsive and separating privileged ones — as a well-known AI use in e-discovery, so review-platform experience is real experience.Get hands-on, on your own clock
Under Opinion 512, a lawyer may not bill a client for time spent learning a generative AI tool the lawyer will use regularly. Treat that as the scheduling rule: learning hours are your investment, not client time. Deliberate hands-on use is also what lets you claim the functional fluency interviewers can test.Reframe your experience as problems solved
Present your experience as the problems you can own rather than the tasks you have done: describe matters by the workflow you improved or the risk you caught, and be ready to explain how a tool should have made it better.Pick a lane from the postings, not the titles
A legal engineer posting can be a builder's job at one company and a configuration-and-enablement job at another. Read the actual responsibilities of the roles you find, talk to people doing them, and choose by the day-to-day work rather than the title.Get fluent in the ethics frame
Opinion 512 ties generative AI use to competence, confidentiality, client communication, supervision, meritorious claims, candor to the tribunal and reasonable fees. That is the frame firms work within when they deploy these tools, so fluency in it is directly usable in interviews and on the job — whichever side of the market you land on.
Legal tech is one direction among many.
If what you are really weighing is the shape of your career as a whole, start from the attorney careers hub and work outward from there.
Where to find attorney jobs
Two practical notes on the search itself.
First, these roles post under several titles — legal engineer, solutions consultant, knowledge management attorney, practice innovation — so search more than one.
Second, expect them in two different places: firm-side roles alongside other lawyer jobs, vendor-side product and customer roles on the software companies' own career pages.
On this board, attorney jobs at law firms are the live listing — the buttons on this page show what is open today.
For the career context around any application, the attorney careers hub and our AI tools article are the two nearest pages.
Career information, not legal advice. For the rules that govern your own use of AI tools in practice, confirm with your state bar.