Artificial Lawyer Predictions 2026 – Part 2

Welcome to Part 2 of the Artificial Lawyer 2026 Predictions, featuring a range of industry experts who all answer three key questions: what will definitely happen, what might happen, and what definitely won’t? Thank you to everyone who took part.

And just a quick word on predictions and the future. There are two main schools of thought. One is the ‘it’s out of our hands’ perspective: ‘Che sarà, sarà, whatever will be, will be, the future is not ours to see, che sarà, sarà.’

The other is what we can call the Sarah Connor view of life: ‘The future has not been written. There is no fate but what we make.’

AL hopes that adds an extra dimension to your thoughts as you’re reading through the following predictions. Plus, if you’d like to read the excellent Part 1 Predictions, they are here.

—

Sean Fitzpatrick, CEO, LexisNexis, North America, UK & Ireland

What will *definitely* happen in legal AI in 2026?

In 2026, agentic AI will become embedded in the core operating model of leading law firms, shifting legal AI from handling single actions to coordinated, multi-agent systems that handle complex workflows. Research, drafting, validation, and citation checking will no longer live in separate steps or tools; instead, personalized agentic systems will understand user preference and autonomously anticipate next actions, propose work product, and execute routine tasks once approved, all with traceable reasoning and citation provenance.

Voice and multimodal interfaces will make this interaction conversational, with legal professionals requesting outcomes rather than instructions. They will simply ask, ‘Find similar cases and draft a response,’ and receive defensible work product grounded in trusted sources. The firms that master orchestration and oversight of these agentic systems will gain exponential productivity and a decisive market edge.

What might happen this year?

In 2026, legal AI could reach a tipping point where it becomes the default productivity layer for many legal professionals, quietly embedded into everyday work rather than treated as a standalone tool. Lawyers are increasingly relying on AI for substantive tasks like drafting, summarization, and early-stage analysis within secure, trusted environments, reshaping both speed and quality of legal output. As this adoption deepens, firms and legal departments that successfully integrate AI into daily workflows will deliver faster, higher accuracy, and improved client outcomes.

At the same time, customizable automation is likely to gain momentum, with organizations expecting AI automation tools to adapt to their existing day-to-day processes rather than forcing teams to change how they work. Configurable, workflow-native automation could be a key differentiator for legal tech providers this year.

What definitely won’t happen in legal AI in 2026?

Despite the rapid pace of innovation, 2026 will not be the year that legal AI replaces lawyers or operates without human oversight. Fully autonomous, unsupervised legal decision-making, especially in high-stakes or regulated contexts, will remain both unrealistic and unacceptable. Trust, accountability, and professional responsibility still demand human judgment at critical points.

Also, legal work is too nuanced, jurisdiction-specific, and workflow-dependent for non-legally grounded AI tools to succeed. Solutions that aren’t grounded in authoritative legal content, transparent reasoning, and workflow integration will struggle to gain lasting adoption.

—

Kyle Poe, VP of Legal Innovation & Strategy, Legora

What will *definitely* happen in legal AI in 2026?

AI will create a new mode of collaboration between in-house teams and their law firms, fundamentally redefining that relationship. Forward-looking firms will move beyond episodic advice to deliver systems that embed their knowledge and experience into reusable, AI-powered workflows that clients rely on in their daily operations. This model allows firms to scale legal judgment rather than time. Firms that adopt it will remain structurally embedded in client decision-making as AI reshapes the market.

What might happen this year?

The biggest potential surprise might be how quickly advantage concentrates among lawyers who can translate legal judgment into systems, not just outputs. AI fluency will soon become a defining requirement of modern legal practice.

What definitely won’t happen in legal AI in 2026?

The death of the billable hour. AI will automate many low-value, repeatable tasks, but legal work defined by uncertainty, judgment, and accountability will remain difficult to price upfront. As routine hours may disappear, the fewer tasks that must still be done by humans will command higher hourly value, preserving the billable hour even as total client costs fall.

Shilpa Bhandarkar, Wild Ducks Advisory

What will definitely happen: We will see more hybrid deal teams that combine lawyers with other legal and technical professionals (eg legal engineers, data scientists) and tech solutions. For the first time (and finally!) non-legal skills will increasingly be valued at par with traditional legal expertise.

Law firms will invest more heavily in building their own bespoke technology solutions to complement the technology they buy from external vendors, including client-facing offerings.  

Recruitment and L&D approaches will be substantially re-designed as AI-enhanced legal delivery becomes more embedded.  

What might happen in 2026: We may see deal teams that formally include members of the client’s own team, or at least new working models that are genuinely more collaborative in nature. We may see elite law firms embrace agents so that AI is trusted on more stand-alone roles rather than simply supporting individual lawyers.

Consolidation in the legal tech vendor market. Finally,  the launch – or significant scaling – of an elite, global, native-AI law firm (defined as a law firm that is designed from the ground up around AI-first delivery models rather than retrofitted onto traditional structures).

What is unlikely to happen in 2026:  The emergence of AGI, nor advances in current models that enable fully automated, end-to-end delivery of global legal transactions without meaningful human oversight. Nor is it likely that private equity will invest in a truly global elite law firm. It is also unlikely that most of these predictions will play out as planned but they do help with business planning regardless!  

—

Omar Haroun, CEO, Eudia

What will definitely happen in legal AI in 2026?

In 2026, AI-native legal tech companies that are essentially selling LLM wrappers will come under intense pressure. CLOs—facing growing scrutiny from CFOs—will ask a simple question: what incremental value are we getting versus ChatGPT Enterprise, Microsoft Copilot, Google, Anthropic, Amazon Bedrock, or similar platforms we already pay for? That scrutiny will trigger real pricing compression and mark the end of ‘AI for AI’s sake.’ Buying decisions will hinge on provable ROI and measurable business outcomes—shorter cycle times, lower outside counsel spend, and materially increased internal capacity. Tools that aren’t embedded in day-to-day legal workflows, or can’t clearly outperform general-purpose AI, will struggle to survive.

What might happen in 2026?

2026 might be the year the legal AI market starts to meaningfully shake out. As ROI becomes non-negotiable, many point solutions built as thin LLM wrappers may be forced to cut prices, consolidate, or pivot toward services. At the same time, we may see a decisive shift away from the traditional SaaS model that requires customers to migrate data into proprietary systems. Legal teams are likely to prefer AI that integrates seamlessly with their existing enterprise stack—Microsoft 365, AWS and Azure environments, Salesforce, ServiceNow, data warehouses, and internal knowledge systems—bringing intelligence to the data rather than forcing the data to move. Vendors that can’t integrate cleanly into this ecosystem will increasingly be seen as friction, not innovation.

What will definitely not happen in legal AI in 2026?

What definitely won’t happen in 2026 is legal departments continuing to pay premium prices for generic AI layered on top of public models. Raw access to better LLMs will not be a durable differentiator. We also won’t see enterprise buyers embrace AI platforms that ask them to rip and replace existing systems—this old SaaS playbook is no longer viable. And despite persistent hype, 2026 will not be the year lawyers are replaced by autonomous AI. The real winners will be those who combine human judgment with AI that is integrated, accountable, and clearly tied to business outcomes.

—

Catriona Wolfenden, Product and Innovation Director, Weightmans

What will *definitely* happen in legal AI in 2026? 

Last year, I predicted that innovation would increasingly focus on people and processes and I think that this continues into 2026. Recent case law on AI hallucinations, both globally and in the UK, reinforces a simple truth: technology is only as effective as the oversight that identifies errors and intervenes. Just as the old saying in relation to data goes, ‘rubbish in, rubbish out’, if we neglect the human and process side of technology (especially with Generative AI) then we risk a future where the mantra becomes ‘no checks, no credibility’.

What might happen this year?

Whilst there has been a plethora of Gen AI job titles developed in the legal innovation sphere, this year may see a return to the oxymoronic ‘generalist expert’ innovator who has knowledge over many different types of technology and understands people and process too. With the answer to every problem not being a hammer, they have a variety of problem solving techniques at their disposal and are already fluid and well versed in navigating risk, clients, stakeholders, operating models and other business service teams to help accelerate from idea to solution.   

What definitely won’t happen in legal AI in 2026? 

Whilst there are academic papers exploring ‘Using Gen AI and blockchain for enhanced user experience in Metaverse-based therapy sessions’ I don’t think that in 2026 we will see this triumvirate collection of technology being applied to the majority of legal practice just yet. Technology, or various combinations of technology though will force us as a society to revaluate over the next couple of years what is the role of human lawyers and judges in a world where technology can increasingly augment or replace some tasks and decision making. As the Master of the Rolls opined at Legal Geek in 2025, it is up to us as a society to grapple with what decisions we want human judges, rather than machines to make and where that line should be drawn.   

—

Nick Fleisher, CEO, Sandstone AI

What will definitely happen in legal AI in 2026?

The industry will see an acceleration in legal tech consolidation and M&A activity, alongside further market fallout (akin to more ‘Robin 2.0’ scenarios). Law firms will increasingly shift toward multi-vendor adoption—simultaneously leveraging platforms like Harvey and Legora—while pairing these tools with proprietary, home-grown customizations.

2026 will be the ‘Year of the In-House Team.’ We will see the emergence of ‘Clio for in-house’ and ‘DMS for in-house’ solutions. General Counsels will utilize internal AI tools to execute work previously outsourced to law firms and ALSPs, leading to a marked reduction in billable hour spend for routine tasks. Data will move to the forefront as legal teams begin constructing ‘organizational brains’ that extend far beyond standard CLM and contract management. Finally, the way business stakeholders engage with legal tech will undergo a paradigm shift: traditional forms are out, replaced by conversational and agentic interfaces.

What might happen this year?

At least one major law firm will likely spin off an AI-first service offering. This entity would consist of a partner group compensated at traditional levels but powered by an ‘army’ of AI agents and engineers rather than a massive associate pool. This would represent the first true test of whether high-value, complex legal work can be successfully transitioned to AI-native firms.

What definitely won’t happen in legal AI in 2026?

Generic providers like OpenAI and Microsoft Copilot will not overtake vertical solutions in the legal AI market. Similar to how law firms use legal DMS instead of Sharepoint (despite previous Microsoft efforts), legal-specific workflows and requirements are simply too demanding for ‘one-size-fits-all’ tools. Until general reasoning models reach a level of sophistication where they can autonomously generate bug-free, bespoke user experiences and integrations, specialized legal solutions are here to stay.

Ian Nelson, Co-Founder, Hotshot 

What will definitely happen

Many more law firms, corporate legal departments, and law schools will move from debating if and how to implement structured AI training to actually doing it. Last year, we spoke with a number of leaders who told us that 2026 would be their “year of doing” when it comes to AI training and based on how the first few weeks of the year are going, that shift is already underway. This will also have broader implications for training more generally, as upskilling will continue to grow in strategic importance as AI takes on more of the routine work lawyers used to do.

What might happen

More firms may start looking beyond lawyers and include all professional staff in their AI training plans, and crucially, not just give everyone the same training. Different teams have different use cases, risks, and opportunities when it comes to AI. Professional development, business development, HR, finance, and knowledge teams will all interact with AI in ways that look very different from how lawyers use it. Firms that tailor training by role, rather than taking a one-size-fits-all approach, will be best positioned to capture real value from the technology.

What definitely won’t happen

Clients won’t go back to paying for associates to do routine work that AI can already do (and do well). The role of the lawyer is evolving, and that’s no longer a future prediction. Clients will continue pushing for efficiency and true value-add work, and firms won’t be able to reverse that trend. 

—

Logan Brown, CEO, Soxton

What will *definitely* happen in legal AI in 2026? 

Law firms will make moves to implement AI tooling, force adoption and play around with new types of models. We’ve seen an influx of investment into legaltech the last few years. I anticipate that increasing in 2026. We’re still in the early days of what is possible and what will be disrupted. Law firms are paying attention and starting to take action. We’ll see more startups, more large funding rounds, and new legal tech. We’ll also see law firms innovate. 

What might happen this year?

We might see priced fees become the norm for routine transactions. Goodwin is implementing flat fees across the firm. I anticipate that more startups will offer flat fee services. It might become more standard and a real market shift for transactions.

What definitely won’t happen in legal AI in 2026? 

The end of the billable hour will not be in 2026. AI tooling is great at sounding confident and giving concise answers. It’s still often wrong. Humans in the loop are still very necessary from both a tech and consumer perspective. We’ll see human lawyers continue this year. 

—

Noah Waisberg, CEO, Zuva

What will *definitely* happen in legal AI in 2026?

LOTS more launches of legal AI companies and products with AI embedded in them! 

What might happen this year?

A foundation model company (e.g., OpenAI or Anthropic) might focus on legal, either via acquisition or building themselves from scratch (i.e., launching a legal-tuned offering with features built for specific legal workflows, appropriate integrations, and compelling data governance).

What definitely won’t happen in legal AI in 2026?

The billable hour will not dramatically decline as a percentage of law firm revenue.

Dorna Moini, CEO, Gavel

What will definitely happen in legal AI in 2026?

Firms will start tracking time to understand where they can save it, not how to bill more of it. Billable hours won’t disappear in 2026, but their role will flip. Firms will start using time data less to justify invoices and more to expose inefficiency. This is a mindset shift that makes alternative pricing work very well. I can speak as a former Big Law attorney turned legal tech CEO that we track our time at Gavel, too. But we don’t do it to bill clients; we do it to find leverage. In 2026, certain legal teams will do the same: breaking work into categories like ‘deterministic,’ ‘judgment-heavy,’ and ‘client-specific,’ and using AI to eliminate the first category entirely.

What might happen this year?

We might see the first real wave of deal telemetry, meaning that AI won’t just draft, it will turn contracts into a dataset. As more work happens in Microsoft Word with AI in the loop, law firms and in-house teams can measure what actually drives time and friction. This will be data like which clauses trigger re-trades, which fallback positions shorten cycle time, and which counterparties consistently drag negotiations out, and even what leads to litigation. That data layer will lead from legal judgment to dictating outcomes, so playbooks evolve based on evidence, juniors learn faster, and pricing is less guessing because you can tie positions to time and risk.

What definitely won’t happen in legal AI in 2026?

We will not get ‘hands-off’ legal AI that reliably runs matters end-to-end without expert supervision, especially in contracts. Not because models won’t improve, but because the hard part isn’t generating language. Rather, it’s getting to the right language for the right deal context and proving why it’s right. Clients and firms are moving toward higher accountability: adoption is rising, but that only makes accountability more important for the critical issues lawyers face in negotiations and deal structuring. 

—

Nnamdi Emelifeonwu, CEO, Definely

What will definitely happen in legal AI in 2026?

The dynamics between law firms and their clients will keep shifting quickly as AI moves from ‘interesting innovation’ to an expected part of delivery where it makes sense. Buyers will be less impressed by demos and more focused on outcomes: faster turnaround, fewer errors, clearer matter visibility, and predictable cost. We’ll continue to see AI evolve from point solutions into workflow products that support end-to-end legal work such as drafting, negotiation support, transaction management, document lists, and matter coordination, with value measured in time saved and risk reduced, not novelty.

What might happen this year?

There’s a growing, pragmatic acceptance across the industry that AI isn’t 100% accurate. It can be incredibly powerful, but it often gets you to ‘mostly right’ rather than ‘fully reliable,’ especially on nuanced or heavily negotiated work. This year, we may see more solutions designed around controlled AI: systems that keep lawyers firmly in charge of the output, make provenance and changes legible, and focus on review and decision-support rather than autopilot. The ‘lawyer-in-the-loop’ model could mature from a slogan into a real operating approach where legal expertise doesn’t just check AI, but actively steers it to reach the standard clients actually need.

What definitely won’t happen in legal AI in 2026?

We won’t see a single, universal ‘legal AI’ that works equally well for everyone. Legal work is too contextual – shaped by deal structures, client preferences, jurisdictional nuance, and the messy reality of negotiation – for a one-size-fits-all model to replace legal expertise.

—

Rutvik Rau, Co-founder, August

What will *definitely* happen in legal AI in 2026?

In 2026, AI becomes part of the service, not the software stack. Law firms will offer AI-enabled advisory work as a core client deliverable. Legal AI will move decisively out of chat boxes and into the fabric of legal work—living inside email, documents, and matter systems, continuously aware of context, precedent, and intent.

What might happen this year?

We’ll see the rise of more boutique and small law firms. AI will bring in a new wave of entrepreneurial expansion for law firm, the 10 person law firm competing with the AmLaw. As clients are going to be increasingly aware of the benefits of AI, they’ll push for new pricing structures leading to innovative partners and their teams starting new firms to build critical advantages in client delivery. 

What definitely won’t happen in legal AI in 2026?

We won’t be out of demo mode. The models and tools will continuously be evolving with even more innovative solutions coming.

—

Melina Efstathiou, Legal Data Intelligence

What will DEFINITELY happen will involve, more demand from clients, more supply from existing and newly sprouted platforms (yes there will still be more platforms to emerge) and more advocacy towards ‘tidying up your house’ as I call it. Essentially, become more intelligent with your workflows, your use cases and your investment in tech as no doubt, these is currently a lot of haphazard investment in tools by coroporates which have overlapping features and use. Clarity will be sought and confusion will still try to sneak in the AI narrative with Legal Tech.

What MIGHT happen: A few will start following a cleaner, more responsible and more ethical route in the way they implement and incorporate AI in their day to day use (personal or professional). My most recent newsletter highlighted ESG and how it needs to be reframed but I’m not sure how many will embrace the fact that we are currently all ab-using AI for sometimes unnecessary tasks, not realising where true impact can be realised and is needed, disregarding all environment implications.

I’m also hoping to see a bigger, bolder movement from the Courts and Judges, to incentivise all to change their ways.

What will DEFINITELY NOT happen: Professionals are not replaced, adoption won’t be universal still, seamless workflows won’t be a reality but spasmodic adoption will still be the trend. BUT! Sceptics and critics will finally be silent as AI is our reality and not a myth or a hype and, their narrative will become irrelevant and might render them unreliable as professionals, as well.  

Ned Gannon, CEO, Coheso

What will *definitely* happen in legal AI in 2026?

An increased focus on change management, ease of adoption, and near-term ROI will take precedence over shiny demo-ready features as law firms and in-house departments incorporate tools within their workflows. Trust, candor, and collaboration will continue to be key to successful implementations and deepen relationships between vendors and their clients.

What might happen this year?

In line with other professions, we may see a dip in hiring for entry level roles as AI usurps more junior level work. AI adoption may also further accelerate the trend toward work staying in-house as companies are able to leverage technology to increase legal capacity closer to the business.

What definitely won’t happen in legal AI in 2026?

The pace of innovation and experimentation will not slow. In an uncertain and rapidly changing world, sound legal and compliance advice and processes will be more critical than ever as businesses and individuals navigate an evolving landscape.

—

Mathias Strasser, CEO, Scissero, (which just acquired the managed services arm of Robin AI)

What will definitely happen in legal AI in 2026?

Two things:

First, AI agents will be integrated more tightly into legal workflows, which will make lawyers noticeably more productive.

Second, it will be technically possible for law firms to ‘go local’ and run their own AI stacks without relying on Silicon Valley or hyperscalers. This represents a return to digital sovereignty: firms will be able to control their data the way they always have, without shipping it outside the firewall in order to benefit from AI.

We already have frontier-level open-source models that can run comfortably on high-end on-prem hardware today, and forthcoming commodity hardware will make this even more accessible. In other words, the capability is no longer theoretical.

That said, possibility and reality are not the same thing. There is still strong momentum behind AI ‘wrapper’ companies in legal tech, and law firms tend to move with the herd. That momentum is likely to continue in the near term. The interesting question for 2026 is not whether local deployment is viable — it is — but whether any major law firms will take a genuinely contrarian position and adopt it early.

Over the longer term, local and firm-controlled deployments are a foregone conclusion. Whether any firms seize that opportunity in 2026 remains to be seen.

What might happen this year?

The AI bubble may begin to deflate or even burst. Several high-profile legal AI companies (who shall remain nameless) will act as bellwethers. Watch closely for announced IPOs that get delayed or are called off entirely (2000, here we go again), funding rounds on lower than expected valuations, and consolidation.

I also expect Microsoft to move more decisively into legal AI. Whether this will lead to a product launch this year or not is unclear – which is why it’s a ‘might’ for me. But when it does, it will put significant pressure on wrapper companies. Microsoft already sits on every law firm’s desktop. Once Copilot meaningfully supports legal research, document search, and drafting, many standalone tools will struggle to justify their cost.

Timing is uncertain — but the competitive dynamic is not.

What definitely won’t happen in legal AI in 2026?

Lawyers will not disappear in 2026. The long-predicted collapse of the legal profession will not materialize.

AI agents will continue to improve, but they will not be trusted to perform legal work without human oversight — nor should they be. The role of lawyers will evolve, particularly for repeatable and process-heavy work, but the profession itself is not going anywhere.

If I had to bet on which profession is more at risk in the near term — software developers or lawyers — I would bet heavily on developers. I do not expect a single lawyer to lose their job because of AI in 2026. What will continue to change is the nature of legal work, not the need for legal judgment or the need for oversight of the work product of legal AI agents.

—

Isabel Parker, Chief Innovation Officer, White & Case

What will *definitely* happen in legal AI in 2026? 

Client patience will finally run out – and clients won’t indulge any AI experiments that don’t deliver value.

What might happen this year?

More lawyers will realise that the hard part is domain knowledge, not tech – and vibe coding will quietly take over.

What definitely won’t happen in legal AI in 2026? 

An AI-native firm won’t top the league tables. Not yet.

—

Sean West, Co-Founder, Hence

What will *definitely* happen in legal AI in 2026? 

Industry consolidation. Big players will get bigger in order to offer fuller solutions to their clients under one roof by buying up adjacencies and competitors. 

What might happen this year?

Rising star companies crashing out. The competitive landscape is in total flux and things that looked like good businesses six months before will be on life support because of failure to compete with larger competitors; failure to raise money as early validation does not look like a durable moat; and, a realization that many workflows can be recreated using foundation models and assorted agentic coding tools in-house. 

What definitely won’t happen in legal AI in 2026? 

The death of the billable hour. It will now just come with an AI platform fee appended!

Peter Lee, Partner at Simmons & Simmons – leads the firm’s AI Governance practice

What will definitely happen in legal AI in 2026? 

In 2026, organisations will prioritise AI governance and allocate dedicated budgets to manage the risks associated with artificial intelligence use. Legal teams and Legal AI businesses will be important players in this work. Organisations will look to implement governance frameworks that include process management tools, audit trails, ethics boards, and clear reporting mechanisms to prevent compliance breaches, commercial risk, and reputational harm. They will provide training programmes, conduct stress testing, and use compliance metrics to ensure that the core principles of responsible AI (fairness, accountability, transparency, explainability, and privacy) apply throughout AI use cases. Expect to see real activity, hard work, and progress in this space, with legal operations and legal engineering professionals being heavily involved because they possess many of the relevant skills and relationships needed to deliver successful AI governance programmes.

What might happen in legal AI in 2026? 

Legal departments and law firms may start to move beyond isolated tools and pilots to embed agentic AI systems into their operational infrastructure. Analysts at Gartner predict that up to 40 percent of enterprise applications will feature task-specific AI agents, a significant increase from current levels of around 5%. By the end of 2026, we could see more organisations integrating self-executing AI agents into workflows to automate document review, legal research, and case management. Major legal technology providers say that they plan to launch multi-agent platforms, which will drive a shift towards more autonomous and orchestrated legal processes. These systems will likely improve efficiency and enable deeper analysis, but their success will certainly depend on advances in governance, data infrastructure, and user trust.

What definitely will not happen in legal AI in 2026? 

I have a feeling that most legal departments will not realise a full, enterprise-wide return on their AI investments, as many will remain in pilot mode, and struggle to achieve adoption and consistent productivity gains. I also do not think this is the year that agentic AI will replace human oversight; lawyers will continue to make final judgment calls and strategic decisions. Courts and regulators will continue to require clear lines of human accountability, regardless of AI’s increasing role in automation and legal work.

—

Michael Grupp, CEO, Bryter

• Users will begin to measure ROI and ask for value beyond tests and individual enablement. Full transition from ‘ChatGPT for lawyers’ to ‘Task-specific automation’
• Something I hope for: we will see more real success stories with measurable benefits?
• Again, no law firm will go out of business because of AI.

—

David Silbert, Sr. Director of AI & Platform Strategy, DocuSign

What will *definitely* happen in legal AI in 2026? 

In 2026, the legal AI landscape will *definitely* shift from experimenting with standalone agents and solutions to embedding them directly into legal and contract workflow systems where governance and auditability are architectural requirements rather than afterthoughts. We will see a ‘reality check’ where organizations stop prioritizing maximum autonomy and instead demand systems that constrain AI through structured, logged processes, making explainability and traceability table stakes for adoption. This move will be driven by regulatory pressures, such as the EU AI Act, and a business demand for proven ROI. In practice, ‘show me your guardrails’ will increasingly mean ‘show me your workflow’.

What might happen in legal AI in 2026? 

We *might* see the current ‘gold rush’ of fragmented legal AI startups begin to consolidate significantly, as legal teams tire of running weekly RFPs for disjointed tools and instead seek fewer, more reliable platform partners that can support end-to-end workflows. This increased market maturity will bring convergence: specialized point solutions and AI agents folding into established ecosystems. Additionally, 2026 may finally be the year where AI moves beyond aspirational ROI. Instead, the ‘promise of tomorrow’ will convert into measurable value through the use of AI contract agents that successfully automate low-risk, high-volume tasks like intake triage, request prioritization, and simple renegotiations.

What *definitely* won’t happen in legal AI in 2026? 

What we definitely won’t see is the total removal of humans from high-stakes, complex legal work. For example, AI will not be independently sourcing, negotiating, and executing complex deals in the mergers and acquisitions space. Furthermore, AI will not ‘kill’ the billable hour and traditional law firms in 2026. Rather, AI will take over inefficient, low-value, and overpriced tasks, forcing firms to evolve their service models rather than disappear. Finally, organizations definitely won’t achieve success by deploying agents that ‘figure it out’ on their own. The idea of ‘autonomy without architecture’ will be rejected in favor of systems where risk is visible and manageable.

—

Tara Waters, TLW Consulting

What will *definitely* happen in legal AI in 2026? 

Market consolidation continues as buyer fatigue slows sales cycles even further. The industry is exhausted and still trying to justify initial legal AI price tags with proper adoption and ROI data. New providers will struggle to get air time, leaving them with few options. Vendors already in the door will seek entrenchment by providing additional value through acquisitions. Whether buyers will welcome this is unclear.

What might happen this year?

An AI-native law firm wins a significant client mandate setting off alarm bells at traditional firms. The shift of work away from law firms has been a looming threat. While many think this work will be retained by clients, the greater concern for firms is this work going to competitors, especially ones that they are not structurally capable of competing with. I’m cheering our new market players on.

What definitely won’t happen in legal AI in 2026? 

The major AI labs won’t launch legal-specific applications that compete with legal AI vendors. Notwithstanding Perplexity’s foray into patent search (really information retrieval rather than legal research), both Anthropic and OpenAI have already signposted that legal advice is a no-go area for them. They’re likely making better margins licensing their models to vendors and through their direct-to-consumer applications, and chatbot UI offers limited usefulness for legal work. If surviving the AI bubble is key for them, I don’t see how wading into the legal industry helps.

—

Shawn Curran, CEO, Jylo

What will *definitely* happen in legal AI in 2026?

We anticipate two major developments for legal AI in 2026:

First, law firms will increasingly prioritise deploying AI platforms within their own secure environments – such as their private Azure instances – driven by growing concerns around supply chain dependency and intellectual property. While technology integration was less of a concern when work was primarily human-driven, as legal AI automates greater portions of the workflow, we are seeing firms become much more attentive to who has jurisdiction of the technology layer that might well underpin the future of their practice. This will likely also lead to additional scrutiny on legal AI vendor incentives and alignment – and to seek out business models that do not ultimately need to compete with their own services.

Second, we foresee a significant rise in ‘citizen developer’ lawyers. From the outset, Jylo invested in a low-code/no-code approach, building a platform that supports lawyers in directly configuring their own workflows and user experiences. Although operationalising these custom workflows started out as too complex, we are now seeing a wave of practicing lawyers engaging in no-code automation and prompt engineering, using in-house data to create tailored solutions. Our vision has always been to abstract complexity through our Playbooks, and we’re well positioned and super excited to support this new movement – empowering lawyers to take macro-level control over AI-driven processes while enabling fine-grained customisation of the user interface and user experience. We believe this movement of lawyers as citizen developers will be a defining feature of legal AI in 2026.

What might happen this year?

From our conversations within the industry, many within private practice law firms have yet to seriously embed their unique knowledge and expertise into legal AI tools – and are largely using out-of-the-box prompts or workflows which have low barriers to entry. Currently, the ‘triangle’ of AI deployment in legal looks like this: a large base of massive software engineering investment at the bottom, a narrowing layer of domain (legal) engineering in the middle (firm or vendor built), and a small high-level goal or outcome at the top. However, this structure is shifting. Soon, our view is that exec committees and boards of law firms will recognise that the triangle is inverting: the high-level goal will increasingly grow at the top of the triangle, supported by less domain-specific engineering, and an even smaller software base underneath – most of which may be owned or operated by external vendors, not the firm itself.

Firms will become increasingly concerned that their digital operations are precariously dependent on someone else’s infrastructure and models. As an analogy, Coca-Cola would never license its syrup recipe to a water company, and while private practice mostly avoid licensing their data to legal AI vendors, some vendors (not us!) still push for it as their starting position. Even so, like Coca-Cola preferring to keep its recipe internally and have optionality on water suppliers, law firms will likely start questioning the wisdom of deep dependencies or single-vendor lock-in. This realisation may well drive considerable churn in the market, and we expect may motivate a significant move back to on-premises solutions, as firms prioritise protecting their proprietary processes and data over convenience and speed of general legal AI adoption. At Jylo, we’re investing heavily in ensuring slick, seamless deployment of our technology into customer Azure environments to support this trend.

What definitely won’t happen in legal AI in 2026?

Computational legal AI will not advance as quickly or as substantially as AI for software development did in 2025. Unlike code, which benefits from comprehensive abstraction, documentation, and modular libraries or APIs, legal information is much less structured and lacks equivalent, battle-tested building blocks. The complexity and variability of law mean it cannot be easily ‘rolled up’ into reusable libraries with clear methods and rules. As a result, we do not anticipate legal AI achieving the rapid progress seen in software development until legal frameworks are restructured to be more compatible with computational models. Fine-tuning on legal data alone is insufficient as models are never up to date, and while enterprise search and retrieval-augmented generation (RAG) add some value, they are not transformative. The real breakthrough – and something we’re working hard on – will come when well-structured legal data – spanning legislation, regulation, case law, and internal know-how – is organised through context engineering and targeted for specific legal tasks.

—

J P Son, Co-founder, Verbit

What will *definitely* happen in legal AI in 2026? 

In the next 12 months, I think we’re going to start to see major changes in how junior associates are paid and trained, and that will also impact pricing overall. Clients are much less willing to pay for juniors to bill on their cases. The trend was already there, but now AI has made that so much more stark. This breaks the economics of how firms have always staffed matters. Junior development won’t disappear, but since clients are no longer going to fund it law firms are going to need to innovate to fill that gap.

What might happen this year?

It’s now a common view that the foundation model business doesn’t work at today’s prices. B2B customers are paying too little for what these models cost to build and the value they deliver. Some of that was subsidized by free use of third-party IP, but that’s ending. And even without the IP issue, the numbers don’t work. I expect a material price reset by the model providers, although it may not happen this year. When it comes, many legal tech business models on top of them will have to change or in some cases may not fundamentally work.

What definitely won’t happen in legal AI in 2026? 

There is already a strong push toward integration in legal AI, and that will continue. Firms and in-house teams want their tools to work together across (for example) contracts, research, litigation, compliance, and billing. But that doesn’t mean every firm will end up on a single platform. Legal work is too fragmented by practice area, risk, and jurisdiction for that to make sense. I believe the future is a connected stack of specialized tools and/or firm-built systems, not one product that tries to be everything.

—

Paulina Grnarova, Co-founder, DeepJudge

What will definitely happen in legal AI in 2026

2025 was the year of prompt engineering. 2026 will be the year of context engineering. For legal AI to really work and add value, it first and foremost needs to be grounded in real knowledge and data, ideally a firm’s own. 

We’ll also see legal AI moving from experimentation to operational dependence in day-to-day legal work. Law firms and in-house legal teams will no longer ask whether to use AI, but how they can best leverage it to remain competitive.

What might happen in 2026

We may see early consolidation, with legal teams moving away from lots of disconnected tools and toward fewer, more deeply integrated legal AI platforms.

What will definitely not happen in legal AI in 2026

The billable hour will not go away, even as AI becomes more capable. What will change is how lawyers use their time. Legal AI will increasingly assist lawyers in their most crucial, high-value work by empowering them with depth, speed and confidence.

—

Alex Zilberman, CEO and Co-founder, Chamelio

What will *definitely* happen in legal AI in 2026? 

In 2026, legal AI shifts decisively from ‘chat and draft’ to ‘run the work.’ The winners will be the platforms that turn legal tasks into repeatable workflows across intake, triage, routing, approvals, negotiation, obligations, and reporting, with tight permissions, audit trails, and handoffs into the tools teams already live in. The real value will come from orchestration and outcomes, not clever prompts, and buyers will judge AI by cycle time, control, and measurable impact.

What might happen this year?

In house legal teams will increasingly treat outside counsel’s AI posture as a procurement requirement: how they use AI, what data touches what systems, what gets logged, who reviews outputs, and what gets disclosed. That pressure will show up in billing guidelines, engagement letters, and panel reviews, and it will also hit pricing: clients will resist paying for first drafts that can be generated cheaply, while being more willing to pay for judgment, strategy, and high-stakes work that is defensible and accountable. Firms that can demonstrate speed plus rigor will win share, and firms that can’t will get squeezed on rates and preferred status.

What definitely won’t happen in legal AI in 2026? 

Generic foundation models will keep getting smarter, but on their own they still will not deliver reliably useful legal outcomes in the way buyers actually need. Without your playbooks, clause standards, matter context, approvals logic, jurisdictional constraints, and your internal source-of-truth data, they produce answers that are too inconsistent, too hard to audit, and too risky to operationalize. In 2026, ‘model quality’ becomes table stakes, and the differentiation moves to domain grounding, integrations, guardrails, and workflow execution that turns intelligence into action inside the organization.

—

Raymond Blyd, LegalComplex

What will definitely happen in legal AI in 2026

• Focus on AI infrastructure: Legal AI vendors will prioritize AI orchestration, custom workflows, evaluation, and governance. These won’t be flashy new features, but they are essential for completing legal tasks with accurate and consistent outputs. 
• More investment in experimental legal AI: A significant amount of VC capital (dry powder) is sitting on the sidelines, searching for defensible legal AI startups. The phenomenon of ‘Kingmaking’ will continue, resulting in more capital for market leaders. Nevertheless, experimental ventures will also get a chance to try new approaches to solving legal problems. 
• More legal SaaS companies will quietly fold: Even though many will rush to pivot and deploy AI, the SaaS business model requires certain operational margins. Legal AI economics demand a nearly constant fundraising cadence that many software-as-a-service companies may not be able to sustain.

What might happen this year

• AI progress slows: Despite increased capital expenditure on AI, we’ll witness fewer AI breakthroughs. 
• Legal AI adoption grows exponentially: More legal teams will start using AI even as breakthroughs slow. Corporate legal teams will shepherd safe AI adoption across enterprises. Law firms will also start building their own legal AI stacks to maintain control over legal service delivery.
• FOMO fades: We will less frequently hear new revenue growth announcements as growth cools. As momentum moderates, legal AI customers will feel less pressure to adopt purely out of fear of missing out and will take more time to shop around.

What definitely won’t happen in legal AI in 2026

• A clear leader in legal AI: No vendor or solution will achieve an unassailable, lasting competitive moat.
• Artificial general intelligence (AGI): It’s like the agents we were promised in 2025 that never materialized. Even if some claim to have achieved AGI, legal AGI is still a long way off. 
• Harvey and Legora merge: Consolidating the legal AI space into their product portfolio would be attractive for public investors in OpenAI. Thus, OpenAI absorbs both Harvey and Legora in an all-stock deal before the IPO of the century. Don’t worry; it’s highly unlikely that OpenAI will go public in 2026.

—

Eduardo Rastelli, IHG

What will definitely happen?

Legal AI will be expected as a core part of the legal technology infrastructure rather than a ‘nice to have,’ embedded into everyday workflows and tools. While legal work will continue to require a human in the loop, a ‘new lawyer role’ focused on a more ‘cockpit’ and ‘problem-solver’ function (supervising, validating, and applying judgment) rather than acting as the primary doer, particularly for routine and lower-value tasks, will evolve. At the same time, AI governance and usage policies will be strengthened and become business as usual, reflecting regulatory and ethical pressures. Clients, in turn, will increasingly expect the efficiency gains delivered by AI, even if they remain unwilling to accept fully automated legal services – this includes businesses, in-house departments and law firms.

What might happen?

New billing models are likely to continue emerging, building on early examples already in the market and accelerating a shift toward outcome-based or value-based pricing for AI-enabled legal work. In-house legal teams are also likely to push harder than law firms to implement and scale AI, viewing it as a key lever to unlock capacity and efficiency without the traditional linear increase in headcount and costs. 

What won’t happen?

AI will not replace lawyers, though it will continue to transform how they work, shifting roles toward supervision, problem-solving, and validation rather than traditional task execution. Fully autonomous legal decision-making will remain out of reach in 2026, likely only becoming plausible around 2027 depending on organizational risk appetite. Universal regulatory alignment across jurisdictions will not occur, and AI systems will continue to produce errors, meaning the want for ‘perfect accuracy’ remains out of reach.

—

And that ends Part 2 of the Artificial Lawyer Predictions 2026. I hope you enjoyed them. Here is the link to Part 1, in case you’d like to peruse those as well.

Thanks to everyone who took part and sent in their thoughts and insights. Now let’s see what actually happens!

P.S. the main pic shows a group of legal innovators standing around a Sphinx, a symbol of the search for mortal wisdom in a hard to fathom world, and with an Egyptian pyramid – which once again represents the law firm business model – in the background.

Richard Tromans, Founder, Artificial Lawyer