ILTACON 2026: Context is king – but transformation is still the hard part

For the past three years the legal technology industry has moved at relative breakneck speed, from initial curiosity around generative AI to pilot programmes, firm-wide rollouts and now, agentic experimentation.

Across conversations with senior leaders at iManage, NetDocuments, LexisNexis, Intapp, Harvey, Aderant and others, a common set of themes repeatedly surfaced: context engineering, data quality, governance, agent orchestration and a growing focus on ROI. There’s been huge progress since last year, but transformation remains the hard part.

The race for context

If there was a single word that defined ILTACON 2026, it was probably “context”.

Virtually every major AI announcement in the run up to and during the conference centred on providing legal AI systems with the information, history and organisational knowledge needed to become a useful assistant.

NetDocuments in May launched a Legal Context Graph and speaking to Legal IT Insider at the conference, CEO Josh Baxter said: “Context matters, otherwise it’s like instructing an associate, then that associate goes away and you get someone who knows nothing. Agents will speed up work, so you really need the context.” With the cost of moving from subsidised use of tokens to a consumption-based model on everyone’s mind, NetDocuments’ director of product marketing Michael Hill added: “The answer isn’t less AI, but AI that’s more effective in getting to that right answer faster.”

iManage’s forthcoming new platform, which goes into GA in October, is built around the iManage Context Fabric, a governed intelligence layer.

“The big focus for customers right now is ‘how do I prepare my data for the age of AI,’ said Neil Araujo, CEO and co-founder of iManage.

The new platform will enrich documents with information (think jurisdiction, type of law, parties etc and then link the information about the document with the information in you PMS, for example) to make it easier to find and interrogate.

Harvey’s latest product direction points in the same direction. Rather than lawyers repeatedly starting from scratch, Harvey is increasingly organising work around matters, memory, preferences and institutional knowledge.

Tracking the developments of the past four years in his company update at ILTACON, Harvey’s co-founder and CEO Winston Weinberg said that 2026 is the year of transformation but the second half of 2026, is about how AI becomes matter-centric and personalised, with chief product officer Anique Drumwright saying: “Harvey II is centred around the matter, context, best practices and preferences.” Michelle Mahoney, chief innovation officer at Mallesons, said the firm is working on transforming practice with Harvey, commenting: “We rebranded recently and voice and tone are really important so we’re digging deep on memory, both the firm and individual memory, and we really want to add that context.”

If ever there was an imperative for firms to do the hard yards when it comes to data hygiene it is now. Wilson Sonsini’s CIO Jeffrey Lolley, speaking on stage with Salesforce’s global account director Kiel Parker about the work they are doing together in building a sophisticated CRM system, said: “The technology is the easy part. The hard part is the data. Getting people to understand the value and getting clean data has been the biggest challenge. You hear about people who went out, bought the first thing available and sat it on top of data that wasn’t clean. The data needs to be the first part.”

Data governance challenges will get worse

Governance is notably no longer being viewed or positioned as a brake on innovation, but the enabler of broader adoption. However those who live and breathe governance warn that the challenges will become significantly more complex as agents become more capable and widely adopted.

Speaking to Legal IT Insider, Intapp’s chief product officer and co-founder Thad Jampol said: “Governance challenges will get worse before they get better. As we move away from ‘question and answer’ and have autonomous agents doing work that increases compliance risks by an order of magnitude.”

Intapp during ILTACON announced a partnership with Legora that will connect Legora with Intapp Walls and Intapp Time, for automated AI time capture. Jampol said that the partnership rounds out AI for the practice and business of law. With firms saving and storing information beyond their document management system, Jampol observes that “everything done in Legora is protected by ethical walls.”

Other notable announcements included iManage’s new platform’s ability to disable AI at the client or matter level to accommodate outside counsel guidelines and client requirements.

And one to watch – Wilson Sonsini described building an AI architecture designed to make data access decisions automatically rather than leaving lawyers to determine whether information can be used in a given context.

“We didn’t want to create a new data environment with data that we thought was permissible,” said Lolley. “We wanted to build an engine that could make that decision for our lawyers and give them access appropriately, taking that decision-making out of the hands of lawyers.”

Agentic AI progress

Talk of agentic AI was everywhere in Nashville, and it’s important to cut through the noise, if you can.

Aderant, which is hunting for a new CEO, at ILTACON announced early access to the Aderant Agent Center, which is built on the technology brought in through its acquisition of Hercules. Aderant has seven agents across appeals, collection, talent and time and SVP for product management and AI, Lisa Erickson said: “We will have seven back-office focused agents by the end of this conference, and there’s another eight to ten we’re working on.” Rather than general purpose AI tools, agents start with the firms own business context.

Also focusing on applying agentic AI to the business of law is Intapp, which (alongside Litera) talks about ‘firm AI’, with Celeste able to take care of the intake process (including emails), taking away some of the administrative burden. Jampol says Celeste, which is available in the UK and US, has the biggest customer pipeline in the company’s history.

Litera unveiled Firm AI Search, which connects the data across Foundation, matter systems, and expertise database into a single interface searchable by its agent Lito. Lito now spans the lifecycle of a matter from intake to close, CEO Avaneesh Marwaha said. Harvey discussed its agentic capabilities built around longer-running workflows and matter-centric work management – each matter has a vault, agent centre, library and task list. Opus 2 launched its agentic AI framework Ingenuity to provide AI agents for legal case management and dispute resolution. And LexisNexis and Thomson Reuters both speak of an agentic ‘harness’ that enable legal professionals to leverage multiple tools in one workflow.

Run MCP

MCP was inevitably high on the agenda at ILTACON, with first prize going to the Autologyx booth for its ‘Run MCP’ theme – with Run DMC-style caps and backdrop. Autologyx announced a new set of MCP capabilities in June and if you haven’t read our article on that, do that here.

iManage’s new integration with Thomson Reuters leverages MCP so you can surface iManage content in CoCounsel, but it will also improve the quality of the integration with HighQ. Litera enables customers to connect their data and workflows with Lito using MCP. NetDocuments has an open, AI-tool-agnostic approach to MCP that allows customers to connect to the tools they choose. With Google announcing during ILTACON the launch of Gemini for Legal, the landscape is ever more complex, and law firms need to be able to create their own ecosystem.

The verification tax isn’t going away

One of the more candid discussions at ILTACON centred around the efficiency gains that firms are achieving, or not, leveraging AI.

Despite substantial AI adoption, many organisations are still wrestling with a ‘verification tax’ that takes away from time savings, meaning that firms are doubling up on effort and time and not recouping costs.

Actionstep’s chief product officer Triona Saunders noted that firms have not yet fully realised efficiency gains because significant human review remains necessary. Actionstep this acquired Trace for passive time tracking, which is now generally available. “Time and again we hear that firms need visibility into what people are doing,” she says. “You need an audit trail and the ability to track what agents are doing. Because we have the permissioning piece in our platform, every user has a set of permissions, so by connecting AI we can control that.”

Tokenomics arrives

A topic that barely featured last year emerged repeatedly this year: tokenomics, for want of a better word. The days of subsidised AI experimentation looks to be ending. Speaking to Legal IT Insider, Brad Kolacinski, a partner at Control Risks, said: “Attorneys treat AI as free but tokenomics are coming. A lot of these frontier models will move to a token model so they can make money, which will change the dynamic.”

Weinberg suggested that frontier model prices will keep going up thanks to constraints on infrastructure – you can’t buy more computing power before 2029 because demand for AI exceeds our ability to build and power new data centres. Harvey is focusing heavily on benchmarking performance and “building the muscle to prepare for these changes.”

Hill points out that this conversation around the cost of tokens doesn’t have to be negative, observing: “It’s great that we’re talking about tokenomics now. The conversation is how do we optimise and make sure we get value for every dollar we spend. That’s really positive. People are getting serious.”

The billable hour and productivity

Of particular interest to those of us who have worked in the sector for decades were the conversations around the billable hour, pricing and changes to law firm structure.

These conversations are often laughed at due to the fact that change has seemed to be on the horizon forever. But speaking to Legal IT Insider from her new role in Hike2, principal of legal advisory services Catherine Krow, former founder of Digitory Legal said: “For the first time we are seeing a move off the billable hour. Empowering law firms and departments to have value conversations in different ways is very exciting.”

 

Productivity and ROI also take centre stage. At legal tech consultancy Opensity, chief intergration officer Andrew Dober says:  “People have solved what technology they want to use, but how does it work together and drive ROI that helps them evolve as an organisation? Historically people focused on use of technology in practice. Now there is more engagement around one platform and how data comes together.”

Change management and transformation

Tied in with this and perhaps the final important takeaway from ILTACON 2026 is that despite all of the talk of progress, change is really hard. We’ll be bringing you a series of IT director interviews shortly, where we will focus heavily on  change management, and it remains a struggle.

At Epiq, managing director Kieron Champion told us: “We’re seeing a lot of tools being deployed and licensed but I’m not convinced that users are getting value. You’ve got to make it sticky, intuitive and elegant. Firms are buying licences and throwing them over the fence. They won’t get the full value of AI. I think only the most mature firms around this AI journey are looking to transform the way work is done. Rather than seeing it as a productivity tool that transforms the way they work, they haven’t changed their processes or identified what work can be done by AI and where humans need to be in the loop or lead.” 

Major progress was evident everywhere this year at ILTACON, but firms have to change the way they work to fully take advantage of it.