Feb 17, 2026

5min read

How Harvey Is Defining Application-Layer AI

How Harvey Is Defining Application-Layer AI

Authors

Patrick McBride

Legal work has become one of the core proving grounds for embedding AI into the enterprise

Legal services are critical to modern economies. They underpin corporate activity from everyday contracts to once-in-a-generation transactions, help organisations scale across counterparties and borders, and enable firms to regulate risk. They comprise a trillion-dollar market.

Despite the industry’s crucial role, however, the actual delivery of legal work remains fragmented. Workflows are spread across disparate, disconnected systems, and institutional knowledge is difficult to uncover and time-consuming to share. As a consequence, talented, highly trained professionals still have to divert too much of their time each day from value-add work to manual, repetitive tasks.

For years, technology has promised to solve these challenges and unlock huge productivity gains but has ultimately fallen short of the hype. Experienced lawyers have witnessed successive waves of innovation, from contract lifecycle to knowledge management, but found that their daily life remains dominated by workflows built around 1990’s technology: email and Microsoft Word.

Recently, however, AI is allowing the legal market – long dismissed as a technology laggard – to set the standard for digital maturity across industries. This is not because AI will simply replace human judgment or automate away billable hours. Rather, AI is offering forward-thinking legal practices a catalyst to finally re-architect the way their work is organised, executed and scaled. 

Compared to the prior generation of SaaS tools, application-layer AI platforms cannot succeed on clever features alone. The underlying LLM capabilities are advancing at an incredible pace, and customers require vendors who can help them adapt as the technology evolves while also delivering value today. They need vendors who can back up exciting features with secure, scalable infrastructure and a partner ecosystem that helps users continue to leverage the legal tools they already find valuable. 

These are the challenges Harvey was designed to address.

An operating system built for how legal work actually happens

In just over three years since its founding in 2022 by Winston Weinberg and Gabe Pereyra, Harvey has scaled at an extraordinary rate, and today its team of 500+ employees serve more than 1,000 customers in more than 50 countries. By year-end 2025, Harvey surpassed $190 million in ARR, demonstrating that its marriage of world-class engineering talent with deep domain expertise in legal has earned the trust of many of the most demanding organisations in the world. 

From the outset, Harvey took a clear view: meaningful progress in legal AI would not come from siloed features, or from shoehorning established workflows into narrow point solutions. Instead, legal teams need a unified platform that can automatically assemble the full scope and context of a matter across multiple systems and knowledge sources, enabling research, drafting, document review, workflow creation and collaboration to occur within a single environment. They need flexibility to leverage the latest LLM breakthroughs without falling hostage to a single model provider whose future performance (or pricing) is highly uncertain. And they need all this to work in a way that is safe, reliable and repeatable. 

Legal teams do not need another standalone application or a plug-in that requires a DIY approach. They need an operating system to unite and then streamline fragmented work.

Why we have partnered with Harvey

At EQT, we look for companies that are reshaping how critical industries function, particularly where software can enhance both productivity and quality at scale. Harvey stands out not only because of what it has achieved to date, but because of the foundations it has laid to lead the next chapter of enterprise AI.

Our deep conviction in Harvey as an investor and customer stems from three key observations.

1. Harvey is building a foundational layer in application AI, not a collection of features

Most LegalTech vendors (and many early attempts at application AI in other industries) have approached AI by layering new capabilities on top of existing workflows, which are inherently fragmented. Harvey starts from a different premise: AI is only as good as the context it can reliably access, and in legal work that context already has a natural organising principle: the client matter.

Modern law firms already run on matters. Matters govern permissions and ethical walls, determine how work is billed, define who needs to collaborate, and serve as the reference point across disparate systems such as document management, e-discovery, research, timekeeping and billing. In practice, however, the information tied to a matter remains fragmented across dozens of disconnected tools.

As a result, lawyers are forced to reconstruct context manually - moving between systems, uploading documents, selecting sources and crafting prompts simply to give AI a chance of producing accurate work. Each tool competes to become the primary hub (with a bit of AI magic sprinkled on top), leaving users to adapt their workflow to the software rather than the other way around.

With Harvey, the tail no longer wags the dog. Firms and users are given the tools to redesign workflows around how they want to work today, rather than how they were forced to work by legacy solutions. 

Crucially, Harvey does not need to, nor want to, replace every system in the legal stack. Rather, it acts as a connective layer that unites data, workflows and institutional knowledge in a single, cohesive workspace. Harvey can unify context on a given matter across existing systems, approved user memory (including work patterns and drafting preferences), AI capabilities, and enterprise-grade controls for security, confidentiality and governance.

Rather than incrementally upgrading isolated features or simply applying LLMs to existing tasks, Harvey’s customers are given the opportunity to redesign workflows that have long been suboptimal - with the flexibility to keep the things that still make sense. In short, Harvey takes AI from overlay to infrastructure.

2. Harvey’s early decisions are now compounding into durable product advantage

Many of Harvey’s most important product decisions were made before they were strictly necessary. Investments in talent development, organisational scaling, security architecture, enterprise controls and extensibility were prioritised early, even when simple approaches and shallow demos would have sufficed.

Those investments are now beginning to compound.

Harvey’s ever-growing roster of blue-chip clients – from the world’s leading law firms (including more than 50% of the AmLaw 100) to global champion corporates (such as Walmart, PwC, Deutsche Telekom, Bayer, Merck, Comcast, Meta, Carrefour and EQT) – is not just a reflection of its success to date but also a clear source of ongoing advantage. As usage continues to ramp, Harvey benefits from ever-increasing domain expertise, sharper contextual understanding, and feedback loops informed by how leading legal organisations actually work. The breadth of the customer base across jurisdictions, practice areas and organisational models reinforces this effect, as does Harvey’s growing ecosystem of integrations and partnerships. 

While a great deal of attention has been paid to Harvey’s funding, it is talent, not merely capital, that will fuel this compounding product advantage. Harvey attracts world-class engineers, data scientists, researchers and former lawyers who are obsessed with solving challenging problems at the intersection of AI science and real-world complexity. And landing the world’s best talent requires a truly global presence - so Harvey is building its team not just in Silicon Valley but worldwide, including its most recent office openings in Bengaluru, Dublin and Paris.

We are already beginning to see the benefits of Harvey’s client network, partner ecosystem and talent pool reinforce one another, with recent releases including MatterOS, Shared Spaces and Memory setting the tone for the exciting road(map) ahead.

3. Harvey is helping change how legal teams create value, not just how quickly they work

Productivity gains are an important starting point, but they are not the ultimate goal. What will matter more is how legal teams allocate their time, manage risk and collaborate with key internal stakeholders and clients. 

Harvey’s ambition is not simply to help its customers complete current workflows more quickly or with less human intervention. Rather, it seeks to empower users to move beyond siloed, email-driven exchanges toward shared, governed workspaces where law firms and in-house teams can operate with greater transparency and coordination. Within these environments, AI supports not just efficiency but also consistency, institutional memory and informed decision-making. 

For clients, this translates into better outcomes, not just faster ones, alongside enhanced visibility and control. For law firms, it helps open the door to new service models and deeper client relationships. In fact, many of the most thoughtful law firms are already asking Harvey how they can leverage its capabilities to help them enter new practice areas, grow their market share or double down on existing areas of strength. Far from eroding firms’ distinct proprietary advantages, Harvey is giving teams the leverage to apply their strengths at greater speed and scale than ever before.

As Harvey redefines the scope and speed with which lawyers can operate, it is also investing in the next generation through its law school partnerships (which currently span the US, UK, Spain and Australia), helping to train law students for future AI-native roles.

From experimentation to expectation

The last 18 months have seen pilots and debates about adoption give way to a race to embed AI deep within the enterprise and to re-architect traditional processes into AI-native workflows.

While many AI experiments will inevitably fail and the precise penetration curve is difficult to predict, the direction of travel is clear. For enterprises, AI adoption has become an expectation.

Just as the rise of slick, easy-to-use consumer apps such as Amazon, Uber and AirBnB eventually drove businesses to demand similarly modern UX’s from their SaaS apps, the explosion in consumer adoption of ChatGPT, Gemini and Claude can only increase the pressure on enterprises to keep pace with the technological frontier. Human tolerance for friction continues to wane, and when it comes to ways of working, organisations are currently rethinking everything, everywhere, all at once.

With this backdrop, Harvey is building an operating system that legal teams can grow into – one designed for scale, trust and long-term relevance. In doing so, they have become an early leader in application-layer AI not just in legal but across industries. We are proud to partner with Harvey on this journey and look forward to continuing to support Winston, Gabe and the broader team as they continue to redefine the world of modern work.

More to read

20 articles