Highlights

Watching a technology company emerge

I’ve been at 抖阴成年 long enough to remember when “technology company” felt more like an aspiration than a description. Today, it feels like a fact, one that we earn every day through the work we actually ship, not the language we use to describe ourselves.

I听co-lead the legal product portfolio here. I started my career as an M&A lawyer at a large New York City law firm, then became one of the founding employees of Practical Law in the U.S., and have been building legal technology ever since. I’ve watched this industry from a lot of angles, and what I see at 抖阴成年 right now is something genuinely different: a company that is doing the hard, unglamorous work of becoming what it says it is.听

That transformation doesn’t announce itself. It happens in product reviews and architecture decisions. It happens when a team pushes back on a roadmap because the underlying model isn’t ready to be trusted at scale. It happens when editorial experts and data scientists sit in the same room and argue, productively, about what “accurate” means and then come out the other side with a completely new approach to AI evaluations. That’s the 抖阴成年 transformation I experience, not a rebrand, but a rewiring.听

Why trust has always mattered

Those years in M&A practice taught me something that still shapes everything I build: the stakes in legal work are not abstract. A poorly drafted clause, a wrong interpretation, a confident answer that turns out to be wrong, these have real consequences for real people and real organizations. That instinct for precision and accountability followed me when I joined Practical Law in its early days in the US.听

Practical Law was built on a clear premise: legal professionals deserve tools that are backed by real expertise, maintained with rigor, and designed to be trusted. Technology was always in service of that commitment, not the other way around. We scaled with technology, but we never let scale become an excuse for cutting corners on quality. That philosophy is not old-fashioned, it’s exactly what the legal market needs now, more than ever.听

Which is why, when I took on the rebuild of CoCounsel, I wasn’t interested in incrementalism.听

The old product had strengths. But I’ve learned, that knowing when to start over takes more courage than knowing how to iterate. We made the decision to rebuild CoCounsel from the ground up, designed for power, designed for flexibility, and built to a standard I’d describe as fiduciary grade.听

That phrase matters to me. Fiduciary grade means the AI isn’t just useful, it’s trustworthy in the specific way that legal work demands. It’s transparent about what it knows and doesn’t know. It’s grounded in the authoritative sources that Westlaw and Practical Law represent, not just the open web, not just pattern-matching on training data, but the deep legal knowledge that our editorial teams have spent decades curating. It holds itself accountable the way a lawyer holds herself accountable.听

The market is responding to that framing because they recognize what it means. When you’re advising a client, filing a brief, or structuring a deal, you can’t afford AI that is merely impressive. You need AI that is reliable. Legal work product that looks good but has substantive issues is dangerous.听

Innovation is a team sport

None of this gets built alone. The rebuild of CoCounsel happened because of a team that was willing to think differently together. I brought a perspective shaped by legal practice and product leadership. My engineering colleagues do the work that makes complexity invisible, building across intricate systems so that what the user experiences is clean, fast, and reliable. Our data science team is where I look when I want to understand where this technology is actually going. They are not just validating what we’ve built, they are inventing it, pushing the boundaries of what’s possible, and asking the questions that keep us building for the future rather than the present. That orientation, always probing the frontier, is part of what makes the work credible. Editorial brought the domain authority that makes every answer more than a guess. Design kept us honest about whether real users could actually navigate what we were building.听

That is what I mean when I talk about the Changemakers narrative from where I sit. Change in a company this size doesn’t come from a single visionary at the top making pronouncements. It comes from people at every level who are willing to challenge the assumption that the way it’s always been done is the way it has to be done, and then do the patient, collaborative work of building something better.听

I’m a woman leading a significant portfolio in an industry that has become increasingly aggressive and male-dominated. I believe that the teams doing the most interesting work tend to be the ones that build differently, that bring in diverse perspectives not as an HR exercise but because the problems are genuinely hard and you need every angle you can get. That’s how I’ve tried to build, and it shows in what we’ve shipped.听

Building for a profession in transition

抖阴成年 is a company in genuine transition. The legal professionals we serve are also in transition, navigating an AI landscape that is moving faster than any prior technology shift in their industry, with higher stakes and less margin for error than most.听

What gives me confidence is that we are not selling them hype. We are building tools grounded in the same values that have defined this company’s relationship with the legal profession for generations: accuracy, trust, expertise. We are doing it with new technology and new ambition, but the foundation hasn’t changed.听

That’s the transformation I’m proud to be part of. Not the press release. The product.听

The right technology can change everything

The best don't just do their work. They change what is possible.

See how

The post appeared first on .

What a “frontier model” is, and how we built one for less

What the industry calls a 鈥渇rontier model鈥� is, in plain terms, a system at the leading edge of capability, the bar every other model is measured against. Until now, that bar has been set by a small number of heavily funded labs, each spending billions of dollars and years of infrastructure to reach it.

We took a different path. With Thomson, our first proprietary large language model, we built a system that competes with those frontier models, but did so with fewer than three dozen people, in three months from first experiments, with a final training run estimated at under $450,000 in GPU costs, specialized using 抖阴成年鈥� authoritative professional content and expert-created data, while preserving broad capabilities.

That efficiency is the headline. What matters more is what鈥檚 behind it: how do we know any of it is true?

The answer is in the technical report we published alongside the launch of Thomson. It鈥檚 built to the standard we call Fiduciary-Grade AI: AI for professionals with duties of care, where 鈥渁lmost right鈥� is not good enough. The report is worth understanding at a summary level even if you never open the full PDF, because the methodology is what makes the claim credible.

Thomson: The technical report

Review our findings and details of the methods, data, and evaluations

Read the full report

Continual learning

The most interesting idea in the report isn鈥檛 the benchmark scores. It鈥檚 how the model was built. Thomson wasn鈥檛 trained from scratch, which would have required the billions in compute that frontier labs spend. But it also wasn鈥檛 simply fine-tuned on top of an existing model, which typically buys narrow domain gains at the cost of broader capabilities. Instead, we started with open-weight models and applied what the report calls Continual Learning: a training approach that materially reshapes the model across the entire training stack while deliberately preserving the capabilities it already had.

The result is what the report describes as a T-shaped performance profile: pronounced gains in the professional domains we targeted, alongside preserved and frequently improved performance in domains we didn鈥檛. That last part is the surprise. Most domain adaptation trades breadth for depth; Continual Learning, as we applied it, largely avoided that trade-off.

The report makes a broader case that this approach is repeatable: a blueprint for other institutions to build competitive models from open-weight starting points, at a fraction of the cost previously imagined.

Two kinds of evidence

The report doesn鈥檛 rest on a single number. It evaluates Thomson two different ways, and the distinction is the most useful thing to take from it: one tests the model in the lab, under identical conditions; the other tests it in the room, against the messy way professionals actually ask questions. Having both is what makes the claim worth taking seriously.

1. The first is a standard benchmark comparison: Thomson-1.0-Large measured against today鈥檚 leading models, under identical conditions, across legal, tax, journalism, safety, and general-purpose tasks. Thomson lands exactly one percentage point behind the top-scoring model on an overall, unweighted average (79.5 vs. 78.5), and it leads every model tested on two specific measures: instruction following and a political-neutrality evaluation. Those two measures matter more than their category names suggest. Specializing a model on dense professional content usually costs something elsewhere. Narrow training tends to buy domain performance by spending general reliability. Instruction following and neutrality holding up, rather than slipping, is evidence the specialization was done with more care than brute force. The report breaks all of this down further in a full comparison table for anyone who wants the underlying numbers.
2. The second kind of evidence is closer to how the model gets used day to day: a blind preference study comparing complete systems, not just models. Subject-matter experts rated thousands of real conversations without knowing which system had produced which answer. In legal conversations specifically, expert raters preferred Thomson鈥檚 answers to those of several named frontier systems a clear majority of the time. This is notable partly because this was a system comparison, not just a model comparison: Thomson had access to our legal tools and Reuters news, while the external models had broader web access. The report lays out the exact win rates against each named system, along with a separate, tighter set of results for general (non-legal) conversation, worth a look if you want the granular comparison. The per-system numbers is what a technical reader will want to scrutinize.

Where it loses, and why that鈥檚 the point

The most convincing part of the report is where it says, plainly, that Thomson isn鈥檛 ahead everywhere. For a skeptical technical audience, this matters more than any win rate: a launch document that volunteers a loss is one you can trust to report its wins honestly.

There鈥檚 one domain where Thomson-1.0-Large scores below its own starting model, attributed to mild forgetting during specialization and explicitly flagged as not a target area for this release. General-purpose reasoning, similarly, trails the strongest proprietary systems rather than leading them.

That鈥檚 worth sitting with. The version the data actually supports is narrower and more useful: Thomson is built and tuned for professional, domain-specific work, competitive with frontier systems on the tasks it was built for, and not represented as the best at everything. For the legal, tax, and compliance professionals this model is meant to serve, that鈥檚 the claim that should matter: not whether Thomson tops a general leaderboard, but whether it holds up on the specific kind of question they actually ask it.

The report鈥檚 constituent-benchmark tables are where that distinction gets precise, if you want to see exactly which tasks it covers and where the boundaries of the claim sit.

An audit, not an advertisement

It鈥檚 tempting to treat a technical report as supporting material for the announcement. It鈥檚 more accurate to treat it as the actual point.

The Fiduciary-Grade AI standard, the one this report exists to back up, only holds up if someone outside this building can check it, which is why the evaluation methodology carries as much weight here as any single score.

The two kinds of evidence only matter if they can be checked by someone other than us. The reason it holds up is that the evidence is open to outside scrutiny. The evaluations use publicly recognized benchmarks, re-implemented in a common harness so that any model is tested the same way. The report鈥檚 training record supports reproducibility and post-hoc analysis, so that for any released checkpoint the constituent datasets and training runs can be recovered exactly. And the smaller model is released as an open weight on Hugging Face for anyone to inspect directly.

Why SovereignAI matters

The report鈥檚 argument goes beyond building one strong model efficiently. Its larger stance is that organizations can own and control more of the AI stack than most assume: the model itself, proprietary data and tools, governance and values, infrastructure, and the economics of deployment. The report calls this SovereignAI, and frames it as a spectrum rather than a binary. Thomson doesn鈥檛 achieve full sovereignty on every axis, but it demonstrates meaningful progress across all of them, on a budget that makes the path viable for a much wider range of institutions than the current frontier-lab model implies.

If any of this is going to inform how you evaluate Thomson for your own work, read the report. The full benchmark tables, the preference-study breakdowns by system and by domain, the methodology behind each, and the open-weight model on Hugging Face are all there, worth reading firsthand rather than taking on faith.

Thomson

The purpose-built, proprietary LLM, engineered for high stakes professional work

Learn more

The post appeared first on .

President and CEO Steve Hasker and Alexander Kardos-Nyheim, Senior Director of TR Labs and co-founder of Safe Sign Technologies, joined Emily Colbert, Co-Head of Product for 抖阴成年 Legal, to defend that bet. What emerged wasn’t a product pitch. It was an argument about where AI competition actually gets won.

Why not just rely on the frontier labs?

Colbert put the obvious question to Kardos-Nyheim directly: why build your own model at all, instead of layering onto a frontier lab’s? His answer was direct. “It’s a good question, and one I’m often asked,” he said. His answer cuts against the industry’s dominant logic: capability was always going to become abundant, because every well-funded lab is racing toward it. That makes capability the wrong thing to bet a company on. Trust, precision, and a defensible chain from answer back to source are the harder problem 鈥� and general-purpose models, trained for breadth, were never built to solve it. Hasker framed the urgency behind that bet in blunt terms: “I’ve always thought about three to five year planning cycles. I think we’re in six month increments now.” That pace punishes any company still deciding whether to build.

What $40 million actually buys

The panel met the cost question directly, then reframed it. Instead of defending $40 million as competitive with the billions frontier labs spend, Kardos-Nyheim treated the figure as a signal of efficiency, not a ceiling 鈥� and pointed to where the money actually went: not compute, but the human layer underneath the training data. 抖阴成年 has described that layer elsewhere in detail 鈥� partner-level lawyers spent months building evaluation rubrics for the hardest research tasks the company could specify, and qualified attorneys logged thousands of hours choosing between model outputs against criteria no generic benchmark captures. The number makes a pointed argument to the rest of the industry: if compute really sets the ceiling on frontier performance, the price tag should look like the labs spending billions. If it doesn’t, a company holding the right proprietary content and the right in-house expertise can close most of that gap for a fraction of the price.

AI sovereignty: the product, not a feature

The sharpest insight from the conversation was about control. Hasker named exactly what customers fear when they raise “AI sovereignty”: that routing their data through a third-party frontier model, or a startup built as a thin wrapper around one, lets that data train a model their competitors, or even their own clients, can eventually query. “So it’s their IP,” he said. “It’s the source of their competitive advantage.” Owning Thomson outright solves that problem at the root. The model runs behind a customer’s own firewall, alongside the products built on decades of 抖阴成年 content: Westlaw, Practical Law, Checkpoint, and CoCounsel, instead of sending a firm’s most sensitive work through infrastructure the firm doesn’t control. Owning the model means 抖阴成年 controls the training data, the behavior, and where it runs, instead of depending on someone else for each.

How Thomson was built

The 2024 acquisition of Safe Sign Technologies, the company Kardos-Nyheim co-founded, planted the seed that grew into TR Labs. What stands out is how little of 抖阴成年’ own content the model has needed to get here: Thomson has trained on less than 10% of 抖阴成年’ proprietary content to date. The panel treated that not as a limitation but as headroom 鈥� the model “only gets better from here” as 抖阴成年 brings more of that archive to bear. Asked which evaluation result mattered most, Kardos-Nyheim didn’t hesitate: “If there’s one result I want to call out, it would be factuality鈥� not whether the model sounds authoritative, but whether its claims survive being checked against the sources it cites. 抖阴成年 has said elsewhere that Thomson clears that bar against general-purpose frontier models working over the open web.

Built to a Fiduciary-Grade standard

抖阴成年 coined a specific term for the bar Thomson has to clear: Fiduciary-Grade AI. The panel drew the distinction sharply; not about how smart the model is, but about who must survive scrutiny from its output. Strip back who 抖阴成年 serves, and you get fiduciary professions: lawyers, tax and audit professionals, people whose work needs to be right. A plausible answer that’s nine-tenths correct doesn’t count as close 鈥� it counts as a failure. That standard drives Thomson’s design goal directly. As the event’s closing message put it, 抖阴成年 built the model “so professionals can verify, cite, and defend their work鈥� AI that supports professional judgment rather than replacing it, where “accountability remains human.” It’s a sharper line than most AI vendors draw, and the panel clearly wanted it remembered:

Where the model already earns its keep

already runs Thomson inside Tabular Analysis, reviewing up to 10,000 documents against as many as a hundred questions at once, with every answer traceable back to its source document. That deployment sits inside a platform already operating at real scale: 1 million CoCounsel users across 107 countries and territories, and roughly 2,500 internal domain experts have built the work behind Westlaw, Practical Law, OneSource, Checkpoint, and CoCounsel. The panel described migrating more of the CoCounsel suite onto Thomson over time, extending the same sovereign, firewall-protected environment across a broader set of workflows.

Curious how Thomson holds up against the standard your work demands? Learn more about Thomson.

Why we built Thomson

You shouldn't have to settle for AI built for everyone. Now you don't.

Read the blog

The post appeared first on .

Highlights

Tax, audit and accounting professionals are operating at an inflection point. The profession is facing rising regulatory complexity, greater client expectations, larger volumes of data, compressed timelines and persistent talent pressures. At the same time, AI adoption is accelerating quickly, creating an opportunity to rethink not only how fast work gets done, but how work gets done.

New 抖阴成年 Future of Professionals data shows just how significant this moment is for tax and audit professionals. Nearly 89% say AI adoption will be transformational or have a high impact on their organizations over the next five years. Almost 70% say the same about the velocity of regulatory change, and 58% say the shortage of skilled labor will be transformational or high impact.

That combination matters. Alongside the focus of becoming more efficient and moving faster, the profession needs to become more strategic, more resilient, more attractive to talent and more proactive in the value it delivers to clients.

These are the challenges changemakers across the profession are addressing. are not adopting technology for its own sake; they’re using better tools to make better decisions, expand capacity, improve quality and create more room for professionals to focus on the work that requires judgment, expertise and client relationships.

AI is becoming part of everyday workflows

AI is already becoming part of that operating model. According to our data, 57% of tax and audit professionals use AI daily or multiple times a day, while 61% say their firm has provided access to Fiduciary-Grade AI tools and that they use them.

This marks an important shift. AI is no longer a future-state conversation centered on experimentation. It is becoming part of everyday workflows, from tax research and audit support to document review, data extraction, compliance checks, return preparation and client communications.

Speed alone is not enough

But in tax, audit and accounting, speed alone is not enough.

A tax position must be supported. An audit conclusion must be defensible. A compliance decision must be grounded in the right rules, facts and context. Professionals need technology that helps them produce work they can verify, explain and stand behind.

That is why to what comes next. The data shows that 95% of tax and audit professionals say it is essential or very important that AI outputs are grounded in authoritative, verified content. At the same time, 35% say they use AI tools not authorized by their organization at least occasionally.

That is a clear signal for leaders. AI adoption is moving quickly, but governance must keep pace. Firms need to reduce shadow AI risk without slowing innovation. They need to give professionals trusted tools that are accessible, useful and built for the high-stakes work they perform.

This is where Fiduciary-Grade AI matters. For professions , accountability, privacy, security and trust, AI must be grounded in authoritative content, shaped by domain expertise and designed with human verification at the center. The professional remains accountable for the outcome, so the technology must support that responsibility.

AI has important implications for talent

AI also has important implications for talent. The profession has long asked too much of too few people, especially during peak periods. Burnout remains a serious challenge, and many talented professionals begin to question whether the career they chose is sustainable.

AI gives us an opportunity to change that. By taking on more manual, repetitive and time-intensive work, AI can help professionals focus sooner on the work that develops judgment: interpreting, advising, challenging, communicating and leading.

The data suggests this is already becoming a talent issue. Nearly one in three tax and audit professionals say they would not accept a job offer from an organization that did not provide access to fiduciary-grade AI tools. And 54% believe AI will shorten the timeline for new accounting professionals to be trusted with significant professional judgment.

That should get every leader’s attention. AI is not only changing productivity. It is changing expectations for how careers are built.

AI is changing client service

It is also changing client service. Historically, many client relationships have been shaped by deadlines: collect documents, chase missing information, complete the work, deliver the final product and move on to the next obligation. That model leaves too little room for proactive advice.

. By reducing friction across document gathering, data extraction, communications, delivery and follow-up, firms can create more capacity to identify issues earlier, surface insights more continuously and engage clients throughout the year.

Clients are beginning to expect that shift. The data shows that 46% of tax and audit firms believe they would begin losing clients within two years if they do not demonstrate clear AI-enabled value. That means AI is no longer just an internal efficiency issue. It is becoming part of how clients evaluate quality, responsiveness and value.

The firms that see AI only as a way to save time will save time. The firms that see it as a platform for growth will change their operating model.

This moment calls for changemakers

That is why this moment calls for changemakers: leaders and professionals who understand the responsibility of the work, see the opportunity ahead and are willing to build a better model for the profession.

At 抖阴成年, that is who we build for. We build for professionals whose work carries consequence, and for firms and departments that need technology grounded in trusted content, domain expertise and professional-grade safeguards.

Because when the work matters, trust is non-negotiable. And for tax, audit and accounting, the future will be shaped by those willing to move from pressure to possibility.

The right technology can change everything

The best don't just do their work. They change what is possible.

See how

The post appeared first on .

Highlights

Throughout my career, one belief has remained constant: technology should make professional work better, not just faster. That belief shapes how I think about the future of legal, tax, risk, and compliance teams and the role technology should play in helping them succeed. So when people ask me what it means as 抖阴成年 accelerates its transformation through technology, my answer comes from what I see in my own business every day.

Corporate legal, tax, risk, and compliance teams keep organizations moving safely through complexity. They don’t get to be “almost right.” A missed clause, misapplied regulation, an unverified answer 鈥� all carry real consequences for the businesses they serve. When I look at who’s leading through that pressure well, I see the same pattern we celebrate in our Chosen by changemakers stories: people who decide to work differently before they’re forced to.听 We’re not just digitizing old workflows. We’re rebuilding the tools these teams rely on to meet a standard worthy of the responsibility they carry.

The Future of Professionals Report 2026 found that 96% of professionals say AI must safeguard confidential data, 94% say its outputs must be grounded in authoritative content, and 90% say it must produce reasoning that can be explained and defended. Yet the gap between expectation and delivery is stark: 78% of corporate clients say receiving AI-enabled quality improvements from their providers is very important or essential, but only 6% say most or all of their providers actually deliver.

That’s the thinking behind what we call Fiduciary-Grade AI: AI built for professionals who operate under real duties of care, not just general productivity gains. It means:

• Authority you can verify: outputs grounded in trusted, domain-specific content, not the open internet.

• Transparent reasoning: the professional can examine and defend the answer.

• Accountability that stays human: responsibility never shifts to the tool; it exists to support judgment, not replace it.

I care about this distinction because I’ve sat across the table from corporate customers who don’t need another AI demo. They need something they can trust with work that matters.

‘s legal team is a good example of what this looks like in practice. Their privacy impact assessments connect directly to the private health data of real patients, so speed can never come at the cost of rigor. Using , they’ve reduced the review process by roughly 60%, cutting what once took five days down to two. Proof that moving faster and moving carefully aren’t in conflict when the technology is built for the responsibility the work carries. It’s an example of how trustworthy AI can help professionals move faster while maintaining the standards their work demands.

That鈥檚 the changemakers story听that matters to me: the general counsel redesigning how their department engages with the business, the tax leader moving from compliance to strategic planning, the compliance team building always-on risk intelligence instead of quarterly reviews. It鈥檚 not a technology story. It鈥檚 a leadership story, and it鈥檚 the one I want my team to write alongside our customers.

I don’t think the corporate function that emerges from this period will look like the one that entered it. Legal, tax, risk, and compliance leaders are redefining how their teams create value, influence strategy, and manage complexity. I want us to be part of shaping that change, not just reacting to it. Our role isn’t simply to provide technology. It’s to help those professionals meet a higher standard, with tools they can trust when the stakes are high.

That’s what Fiduciary-Grade AI听is about. And that’s why I believe the changemaker mindset isn’t a trend. It’s the bar.

The right technology can change everything

The best don't just do their work. They change what is possible.

See how

The post appeared first on .

Highlights

There’s a moment every legal professional knows. You’re deep in a document review 鈥� hundreds of contracts, thousands of clauses, a deadline closing in 鈥� and you ask an AI tool a question that shouldn’t be complicated. The answer comes back confident, well-written, and subtly wrong. Not wrong enough to catch on first read. Just wrong enough to matter.

That’s why we built Thomson. Not just because general models get the details wrong sometimes, but because owning the model outright changes what we can do about it.

抖阴成年 doesn’t just deploy AI. We build it.

Why general-purpose AI isn鈥檛 enough

The past few years have brought extraordinary AI advances. General-purpose large language models (LLM) can write code, summarize articles, and hold nuanced conversations. But legal and compliance professionals work in a different environment. Imprecision has consequences. A misread clause or a missed precedent isn’t an inconvenience; it’s a risk you can’t take.

So what separates a tool that’s merely useful from one you can stand behind?

Not all AI is built for the same stakes. At one end are general-purpose tools 鈥� broadly useful, but shallow on any one domain. A step further are professional-grade tools, built for a specific field, in environments where an occasional error is tolerable. And then there’s a third tier: Fiduciary-Grade AI, built for work where a small error doesn’t just cost time, it can mean a lost case or a client’s trust.

That’s the tier legal and compliance work lives in. It’s the tier Thomson was built for.

The decision to build our own

We’ve always been an expertise company. Grounded by 175 years of intelligence, we’ve combined proprietary data with domain knowledge to give professionals the intelligence they need for their most important work. Thomson is the next chapter of that mission.

Owning more of the AI stack gives us greater control over performance, economics, deployment, and sovereignty 鈥� how the model behaves, what it costs us to run, where it runs, and how our clients’ data is protected. Just as importantly, it lets us continuously improve Thomson using the content and expertise only 抖阴成年 has, rather than depending on what a third party chooses to build next.

That decision started in 2024, when we acquired Safe Sign Technologies, an AI research company founded by leading AI and legal minds. The result is a model that’s built on open-source foundations we can evolve as the AI landscape shifts, while the legal intelligence we put into Thomson only deepens over time. We’re not locked into any single foundational system 鈥� and we’re not standing still.

What we built

Thomson is our proprietary LLM, purpose-built by 抖阴成年 for high-stakes professional work. It was built on our proprietary content, including Westlaw and Practical Law, two of the most authoritative legal data sources in the world, and reflects input from our vast network of human legal experts.

Think of it as a “T-shaped” model. The horizontal bar is broad general capability 鈥� language understanding, reasoning, and the fundamentals any strong AI needs. The vertical column is what makes Thomson different: authoritative legal content and deep subject-matter expertise, refined by expert guidance, producing a level of legal comprehension optimized specifically for professional work. We built it to be both broad and deep.

Thomson and CoCounsel aren’t the same thing, and that’s by design. CoCounsel is the product 鈥� the agentic workspace, skills, and workflows legal professionals use every day. Thomson is one of the LLMs CoCounsel calls on.

What it’s already doing

Thomson’s first implementation just launched, powering CoCounsel Legal’s Tabular Analysis feature. You can extract structured answers from up to 10,000 documents across up to 100 questions you define 鈥� at a scale and consistency that used to be impossible without significant time and cost.

Powering the engine behind Tabular Analysis, Thomson gives you deeper comprehension of complex legal documents, more consistent extraction across question types, lower variance across large document sets, and verifiable, citation-grounded outputs your attorneys can review and rely on. These aren’t incremental improvements 鈥� the numbers show it. As of Q1 2026, CoCounsel has crossed one million users. The share of our business that’s Gen AI-enabled has doubled in a little over a year. Monthly CoCounsel SKUs in legal have quadrupled year-over-year. That’s not a pilot program. That’s a shift in how the work gets done.

Independent testers agree. A Washington University law professor pitted Thomson against ChatGPT and Claude using real questions from his own Corporate Tax class and preferred Thomson’s answers overall, pointing to its linked citations to treatises as especially useful. A separate review from Queen’s Conflict Analytics Lab and Cornell Legal AI Lab found Thomson’s citation quality held up against leading frontier models 鈥� even on Canadian employment-law questions the model wasn’t specifically tuned for.

We hold Thomson to the same standards you hold your own work to. It’s developed and deployed within enterprise-grade governance frameworks grounded in the principles of Fiduciary-Grade AI: transparency, traceability, verifiability, structured evaluation, and expert validation. Thomson is built for repeatable performance you can trust.

Built for the people who make change real

Speed is the easy part. Any AI tool can give you an answer fast. What you need is one you can stand behind 鈥� in front of a judge, a client, a board. That’s what we mean by Fiduciary-Grade AI: intelligence built to the standard your profession already holds itself to.

The legal professionals who rely on us every day aren’t passive consumers of information.

You’re the people who protect businesses from harm, hold institutions accountable, and shape the frameworks that govern how society works. You’re changemakers 鈥� and you deserve tools built with the same seriousness you bring to your work. don’t just need answers. You need answers you can defend.

You’re in good company: a million professionals like you have already chosen to trust 抖阴成年 with the work that can’t afford to be wrong.

We built Thomson for you. Not because AI is a trend worth chasing, but because the next phase of professional AI won’t be defined by general capability 鈥� it’ll be defined by domain mastery, measurable outcomes, and trust. That means higher-quality outputs and greater confidence when you’re using AI in the workflows that matter most.

You shouldn’t have to settle for AI built for everyone. Now you don鈥檛.

Thomson

The purpose-built, proprietary LLM, engineered for high stakes professional work

Learn more

The post appeared first on .

But the thing that听hasn’t听changed is why we exist.听

We are still helping professionals navigate complexity, make informed decisions, and drive better outcomes for their clients, businesses, and communities.听

That’s听why the听Chosen by Changemakers听campaign resonates so deeply with me.听

When most people hear the word “changemaker,” they think of inventors, entrepreneurs, or industry disruptors. Those people certainly matter. But I believe changemakers are all around us.听

They’re the lawyer听helping a family through a life-changing decision.听

They’re听the tax professional听guiding a small business through uncertainty.听

They’re听the journalist pursuing facts and accountability.听

And increasingly,听they’re听the technologists, marketers, communicators, product leaders, and customer teams working together to build the tools that make that work possible.听

From where I sit as Chief Marketing and Communications Officer, transformation isn’t just about technology. It’s about people.听

I’ve spent my career helping organizations navigate change, and one thing I’ve learned is that successful transformation never happens because of a new platform, a new strategy, or even a new technology. It happens because people are willing to challenge assumptions, learn new skills, take risks, and reimagine what’s possible.听

I see that every day at 抖阴成年.听

I see it in our engineers and product teams who are building AI solutions designed specifically for professional work.听

I see it in our marketers and communicators who are helping tell a different story about who we are becoming.听

I see it in our leaders who have had the courage to invest boldly in the future while staying true to the trust our customers place in us.听

And I see it听in听our customers.听

In many ways, they are the ultimate changemakers.听

The professionals we serve are operating in fields where accuracy, trust, and accountability matter deeply. They don’t have the luxury of getting it wrong. That’s why our approach to AI has always been grounded in helping professionals work with greater confidence, not replacing the expertise they bring.听

Our听Future of Professionals research听continues to show that AI is transforming the way work gets done.听But the real opportunity听isn’t听simply about efficiency.听It’s听about enabling people to focus on higher-value work, solve harder problems, and create greater impact.听

That’s听the transformation听I’m听most excited about.听

As communicators and marketers, our role听is to help people understand not just what听we’re听building, but why it matters.听

Because behind every product launch, every innovation, and every breakthrough are people trying to make a difference.听

Those are听the changemakers.听

And at 抖阴成年,听we’re听proud to听build for听them.听

Every day.听

The right technology can change everything

The best don't just do their work. They change what is possible.

See how

The post appeared first on .

Highlights

A year ago, agentic AI was mostly a talking point 鈥� something professionals had heard of but few had actually put to work. That’s no longer true. According to the most recent AI in Professional Services Report, today, 15% of professionals say their organization already uses agentic AI tools, and another 53% say they’re actively planning or considering it, 听Three-quarters of professionals expect it to be central to how they work by 2030.

In other words, agentic AI has quietly moved out of the experimentation phase and into the growth phase. The question for most businesses isn’t whether to adopt it anymore. It’s whether they’re prepared to use it well.

Why the conversation has changed

Generative AI’s rise was fast and visible. Professionals started using tools like ChatGPT to draft, summarize, and research within months of the technology going public, and usage has kept climbing 鈥� 74% of professionals now say they use AI several times a week. Agentic AI is following a similar trajectory, just a step behind. Where GenAI reacts to a prompt and produces something, agentic AI takes an objective and runs with it: researching, drafting, checking its own work, and adjusting course across multiple steps with far less hand-holding.

That shift from “creates content” to “completes a process” is what makes agentic AI valuable 鈥� and also what makes getting it right harder than getting GenAI right. A generative AI tool that gives an average answer is a minor inconvenience. An autonomous system making decisions across a multistep workflow needs to be trustworthy at every step, not just the last one.

The gap that actually matters right now

Here’s what the data says businesses are getting wrong. It isn’t that they’re moving too slowly to adopt agentic AI 鈥� it’s that adoption and value have quietly come apart.

Clients and stakeholders have already raised their expectations: 78% of corporate clients say it’s very important or essential that the firms they work with deliver AI-enabled quality improvements. Only 6% say they’re actually seeing that from most of their providers. That gap hasn’t gone unnoticed 鈥� nearly a third of corporate clients say they’re already reconsidering relationships with firms or providers that are falling behind.

Internally, the picture looks similar. More than a third of professionals (34%) admit to using AI tools their organization hasn’t officially sanctioned 鈥� usually because the tools they’ve been given aren’t good enough or the strategy behind them isn’t clear, a telltale sign that adoption is outpacing governance. And even among professionals actively using AI at work, 41% say they still don’t have access to tools built specifically for professional work and grounded in verified content. Put those together and a pattern emerges: the challenge for most businesses isn’t getting people to use AI. It’s giving people AI that’s actually worth using.

None of this means agentic AI isn’t worth adopting. It means adoption alone was never the finish line. The organizations pulling ahead right now aren’t necessarily the ones that moved first 鈥� they’re the ones that closed the distance between having the technology and getting real value from it.

E-book

An essential guide for bringing AI agents to your organization

Access report

What “good” agentic AI actually requires

Not every tool marketed as agentic AI is built to handle high-stakes, multistep work responsibly. For professionals in law, tax, audit, compliance, and similar fields 鈥� where an error carries real consequences 鈥� “almost right” isn’t an acceptable bar.

That’s the thinking behind 抖阴成年’ Fiduciary-Grade AI standard: AI that’s grounded in authoritative, domain-specific content; protected by rigorous privacy and security safeguards; built and continuously refined by credentialed subject-matter experts; and designed to produce reasoning that’s transparent enough to explain and defend under real scrutiny. Just as importantly, taking on more of the work doesn’t mean AI takes on the accountability. That responsibility stays exactly where it’s always been: with the professional.

It’s a useful checklist for evaluating any agentic AI solution, not just 抖阴成年’ own. If a tool can’t tell you where its answer came from, how it was validated, or what happens when it hits something outside its lane, it isn’t ready for work where the stakes are real.

Where to start

If your organization is somewhere between “we’ve heard of agentic AI” and “we have a strategy that’s actually working,” you’re in good company 鈥� most businesses are exactly there right now. The organizations that get ahead from here won’t be the ones that panic about falling behind. They’ll be the ones that take a clear-eyed look at what agentic AI actually is, what it’s good at, and what to demand from any solution before they commit to it.

That’s exactly what our听Agentic AI 101: What Your Business Needs to Know听e-book was built to do. It walks through the real differences between generative and agentic AI, how professional-grade agentic AI works inside a business, the concrete benefits worth expecting, and 鈥� most usefully 鈥� the specific questions to ask any vendor before you buy. If you’re trying to move from “we should probably look into this” to a plan you can actually execute, start there.

The post appeared first on .

Highlights

Knowing that AI is reshaping professional work is easy. Knowing what to do about it this week, in your firm or your function, is the hard part. And the answer looks different depending on where you sit. Law firm, corporate, and a tax leaders are all wrestling with the same underlying shift, but the pressure points, and the fixes, aren’t identical.

The pattern repeats across all three, though. Adoption has outrun governance. Client and stakeholder expectations have outrun delivery. The organizations pulling ahead aren’t the ones talking about AI, they’re the ones putting it to work. As Raghu Ramanathan, President of Legal Professionals at 抖阴成年, puts it: “The firms creating a true competitive advantage today are not the ones talking about AI. They’re the ones operationalizing it.”

Law firms: clients stopped waiting a while ago

Law firm leaders are caught between two forces moving at once. Clients want proof, not promises: 77% say AI-enabled quality improvements from their firms matter enormously, but only 5% say they’re actually seeing them. Seven in ten in-house legal professionals expect firms to change how they charge as AI use grows. Only 28% of firms have changed their pricing at all.

Talent is watching the same gap. Nearly a quarter of law firm professionals would turn down a job offer that didn’t include real access to professional-grade AI, and when a lawyer’s own view of AI clashes with their firm’s actual direction, they become nearly three times more likely to leave within the year.

There’s also a liability many firms haven’t priced in yet: a third of law firm professionals are using AI tools their firm never approved, often on matters involving privileged, sensitive client information. The firm carries that exposure whether it chose to or not.

What should law firm leaders do? Two moves matter most.

1. Train lawyers to lead the AI conversation with clients themselves, in plain language, grounded in real matters, honest about where the firm is still figuring things out.
2. Open the pricing and value conversation early rather than waiting for clients to force it.

The industry has already moved past “should we use AI?” The real question now is “can we afford not to?” The action report for law firm leaders walks through how firms are already answering it.

Law firms under pressure

Practical actions law firms can take to remain competitive

Read the report

Corporate leaders: this is a leadership problem, not a licensing one

For corporate leaders, Liz Zimick, President of Corporates at 抖阴成年, names the challenge without softening it: “Taken together, these findings point to a change management and leadership challenge that no software license alone will solve.”

Well over half of corporate professionals feel real stakeholder pressure to move faster on AI, and 15% are already seeing financial consequences from moving too slowly, with another 29% expecting them within the year. 59% lack access to tools built to the accuracy standard their fiduciary work demands, so more than a third fill that gap with AI their organization can’t see or govern.

What should corporate leaders do?

1. Stop waiting for direction to arrive from the C-suite. Audit where your people are already using AI, formally or otherwise, and where real demand goes unmet.
2. Plan for every stage of the journey at once, since most departments start by scaling capacity and only later discover they need to rebuild the underlying workflow to reach something better.
3. Lead the AI governance conversation across the enterprise rather than simply joining it.

General counsel, in particular, are well positioned to define what responsible AI use means for the whole organization, not just their own function. The action report for corporate leaders breaks down where to start.

How corporate functions are responding

Strategic paths leaders can take to prepare for the future

Read the report

Tax and audit firms: talent already sees where this is going

Elizabeth Beastrom, President of Tax, Audit & Accounting Professionals at 抖阴成年, frames the stakes around a single insight: “Accounting and audit firm employees increasingly see professional-grade AI tools as a non-negotiable part of their jobs.”

Eight in ten tax and audit professionals now use AI several times a week, yet more than a third are doing it with tools their firm never sanctioned. Clients are watching just as closely: 89% say AI-enabled quality improvements matter enormously, and nearly a third of professionals believe their firm risks losing clients within the year if it can’t keep pace.

What should tax and audit leaders do? Four priorities scale from a two-person practice to a national firm.

1. Govern the tools your people already use instead of pretending they don’t; approving one professional-grade AI tool built on verified content changes your risk profile far more than a policy with no tooling behind it.
2. Know your direction even while the path evolves and decide early whether freed-up time should fund growth, work-life balance, or higher-value advisory work, because chasing all three at once dilutes each.
3. Ask your people honestly what they want AI to change about their work, then build your deployment, and your recruiting pitch, around the real answer.
4. Decide now what work your early-career professionals still need to do by hand, before AI quietly removes the chance to teach it.

The action report for tax & audit firm leaders goes deeper on each of these four moves.

Explore key findings

Four calculated moves for tax & audit firms

Read the report

Choose your path. Don’t drift into one.

Across law, corporate, and tax alike, the organizations getting real value from AI share one thing in common: they named a direction instead of waiting for one to emerge. The biggest gaps were never really technical, they were organizational: the right tools not yet in place, people never trained to work the new way, a strategy that never made it past the leadership deck. Closing that AI value gap takes leadership, not procurement.

The post appeared first on .