For years, legal technology was largely a matter of operational efficiency. It helped firms type faster, store better, communicate more easily and move documents with less friction. But it rarely required the partners to rethink the firm itself. The core assumptions of legal practice remained stable – how work was scoped, staffed, priced, supervised, billed and delivered. Technology improved the machinery. It did not redesign the engine.
That phase is over.
In the space of a few weeks, the legal AI market has started to reorganise itself in public. Anthropic has launched Claude for the legal industry, with more than 20 MCP connectors into the software that legal teams already use along with 12 legal practice plugins. Microsoft has put a Legal Agent inside Word, aimed directly at contract review, tracked changes, citations and playbook-based redlining. Mike OSS has appeared as an open-source legal AI platform positioning itself as an alternative to Harvey and Legora. Harvey itself has open sourced a legal agent benchmark with more than 1,200 tasks across 24 practice areas.
For Irish solicitors, the immediate lesson is not to trial every new platform. That would be to misunderstand the moment. The real issue is that legal work is being reorganised around new capabilities like machine reading, automated drafting, structured data, embedded agents, workflow intelligence and increasingly formal governance obligations. The question for partners is no longer simply which software the firm should buy. It is how the firm intends to design, price, supervise and evidence the work it delivers in a market where technology is becoming part of the legal service itself.
The market is moving from tools to systems. Claude is no longer just a chatbot. It is being wired into Word, Outlook, Excel, PowerPoint, Docusign, iManage, NetDocuments, Box, Datasite, Thomson Reuters, Harvey and other legal infrastructure. Microsoft is no longer hovering at the edge of legal tech with a general Copilot. Now, its Legal Agent is designed to work inside Word, the place where contracts, advices, letters and pleadings actually live. Mike OSS is equally revealing because it shows how quickly a motivated lawyer/developer can now build a credible legal AI interface around frontier models.
The economics are also shifting. Thomson Reuters’ 2026 AI in Professional Services Report says AI adoption has moved beyond early experimentation into a strategic phase, yet only 18% of professionals say their organisations track AI return on investment. Wolters Kluwer’s 2026 Future Ready Lawyer report says 92% of legal professionals surveyed now use at least one AI tool, while the main barriers remain ethics, data privacy, inadequate training and resistance to change.
That is the gap Irish firms should care about. The gap is not between firms that use AI and firms that do not. The gap is between firms that have governed legal workflows and firms that have a growing collection of ungoverned experiments.
In Ireland, the adoption curve is still uneven. The Law Society of Ireland’s 2025 survey found that 69% of solicitors rarely or never used AI in their work, while 78% expected their use to increase over the next three to five years. This is a picture of a profession standing on the threshold. Cautious in practice, but aware that the centre of gravity is moving.
Regulation is moving too. The EU AI Act entered into force on 1 August 2024 and is being phased in. Prohibited AI practices and AI literacy obligations have applied since 2 February 2025. Governance rules and obligations for general-purpose AI models applied from 2 August 2025. The Act becomes broadly applicable on 2 August 2026, with high-risk rules now scheduled under the Commission’s simplification timeline for later dates depending on category.
Ireland is building its enforcement architecture around that timetable. The Department of Enterprise, Trade and Employment announced 15 national competent authorities for AI Act oversight, with a National AI Office to be established by 2 August 2026 as the central coordinating authority and single point of contact.
For solicitors, the EU AI Act is not just a client advisory topic. It is an operational topic. A firm using AI is likely to be a deployer of AI systems. The Law Society of Ireland’s generative AI guidance says solicitors using LLMs in daily work should assume they are deployers of a general-purpose AI system under the AI Act unless use is purely personal. It also urges firms to address permitted uses, accountability, safeguards, GDPR, confidentiality, privilege and professional obligations.
The professional risk is already visible. Reuters reported this week that a US lawyer apologised to a federal judge after phantom AI generated quotations appeared in a court filing, with the lawyer accepting responsibility for failing to vet the output. Even Sullivan & Cromwell, one of Wall Street’s most sophisticated firms, has had to apologise to a federal judge after a filing contained AI generated hallucinations including inaccurate citations and misstatements of law.
That warning has now reached the Irish courts. In Guerin v O’Doherty [2026] IECA 48, Ms Justice Caroline Costello, President of the Court of Appeal, addressed AI generated submissions that cited authorities which simply did not exist. She laid down practical principles for the use of AI in legal research stating that it may assist the work but it does not relieve any party -represented or not – of the duty to verify the law before putting it before a court. One of the greatest risks now is invisible AI use inside the firm: work being shaped by tools the partners has not approved, cannot supervise, cannot audit and may not even know exist.
The answer is not panic. The answer is design.
Richard Susskind has long warned that the deeper legal AI shift will not be simple task automation. In a Law Society of Ireland Gazette report, he argued that the short-term impact of AI is automation, efficiency and productivity, but the longer-term opportunity is new forms of legal service and “preventative lawyering”. In 2025, he put it even more bluntly: the mistake is thinking AI is only about lawyers, when it is really about the future of law and justice.
That is the point Irish solicitors cannot afford to miss. AI is not merely a faster way to produce old work. It changes what clients will expect, what juniors will learn, what partners will supervise, what compliance will require and what fixed-fee legal services can look like.
The practical response is a five part reset.
First, map the work before buying the tool. Client intake, AML, conflict checks, document review, advices, correspondence, billing, WIP, client updates, file closing and knowledge reuse should be mapped as actual workflows. Most firms do not have an AI problem. They have a visibility problem.
Second, classify AI use by risk. Summarising a meeting note is not the same as drafting advice on limitation, producing a pleading, triaging a vulnerable client, reviewing discovery or analysing employment data. The firm needs clear categories – permitted, restricted, prohibited and partner-approved use.
Third, build AI literacy into the firm. Article 4 of the AI Act requires providers and deployers to take measures, to the best of their extent, to ensure sufficient AI literacy among staff and others dealing with AI systems on their behalf. This should not be a one-hour webinar. It should be role-specific training for partners, associates, trainees, support staff, compliance officers and accounts teams.
Fourth, move from experimentation to evidence. A firm should be able to quantify that this workflow used to take X minutes, now it takes Y, this output is checked by Z, this category of data may not be entered into consumer tools, this tool is approved for these use cases, this is how we preserve confidentiality, privilege, audit trails and human review.
Fifth, talk to clients before clients talk to you. Corporate clients increasingly expect their outside counsel to use AI, yet Thomson Reuters notes that fewer than 20% mandate it through guidelines or RFPs. That creates an opening for Irish firms to lead with confidence. “Here is where we use AI, here is where we do not, here is how it improves speed and cost, and here is how your matter remains supervised by solicitors.”
The strongest future firms will be those that can turn new technology into better-managed legal work withclearer workflows, safer supervision, faster delivery, stronger compliance and a more confident client proposition.
AI governance now belongs inside the operating model of the firm. It should be visible at matter opening, risk review, billing, supervision, complaints handling, data protection, AML administration, training, precedent management and client communication. AI is no longer a side issue for the technology file. It is becoming part of how legal work is run.
For small and mid-sized Irish firms, there is a real opportunity here. Large firms may have bigger budgets, but they also have more complexity. Smaller firms can move faster if they start in the right place and that is with a focused review of the work.
Praxis helps Irish solicitors’ firms make sense of this moment without the hype. We start with the work, not the software. The Praxis Legal Tech Audit reviews client intake, matter workflow, AML and GDPR administration, billing and WIP, client communication, technology use, automation opportunities, EU AI Act readiness and practical AI governance.
The output is not a theory deck. It is a 90 day roadmap – what to fix, what to automate, what to stop doing, what to govern, what to train and where technology can create measurable value.
The firms that act now will not simply be “AI ready” but will be easier to run, safer to supervise, clearer for clients and better positioned for the legal market that is already arriving.
