AI is already making parts of legal work cheaper and faster to produce. The question is no longer whether AI is coming. It’s what firms do about it.
Galley works with small and medium law firms to redesign how legal services are delivered, one service at a time.
We work with managing partners, COOs and legal operations leaders who want to understand how AI should change the way their firm works.
Rather than adding new tools to old processes, we start with one service. We look at what the client actually needs, how the service works today, where lawyer time goes, what it costs to deliver, where work gets stuck and where legal judgement genuinely matters.
Then we redesign it.
Using Galley, our operating system for legal work, we break the service into its component parts and decide what should be done by lawyers, AI or software based on what each does best. Controls, reviews and escalation points are built into the service from the start.
The result is a new operating model that can be tested against real matters, measured and improved.
The goal isn’t simply to lower the cost of delivery. It’s to make legal work faster and more predictable, while creating a better client experience without compromising trust or accuracy.
Galley Law is built on a reference architecture called Galley.
Galley is our blueprint for how an AI-native law firm operates. It maps the capabilities a firm needs and redesigns how legal work moves through them using lawyers, AI and best-in-class software.
The premise is that firms should own what makes their practice theirs: their knowledge, rules, processes, quality standards and client data. The technology used to deliver the work can change around it.
We believe that matters because the next wave of legal AI will bring an abundance of specialised models, agents and tools. Firms shouldn’t have to build their operating model around any one of them.
Galley is being designed to sit above that changing market. At a task level, it will allow work to be routed to the right lawyer, model, agent or software based on the quality, cost and controls the task requires.
The goal is to give firms a common governance and measurement layer across AI-enabled legal work. That means being able to see how work is being delivered, measure its quality and cost, change the tools underneath it, and continuously improve the operating model as the technology gets better.
Over time, that should allow firms to reduce low-value work, improve margins and concentrate their people where human judgement creates the most value.
Most firms are adding AI to operating models designed around human labour and the billable hour. Galley goes one level deeper: redesigning the operating model itself around what humans and machines are each best equipped to do, a genuine capability uplift in how legal work gets done, not just a speed or cost reduction.
Legal work is composed into tasks, decisions, evidence and controls, each routed to the right combination: lawyer, AI agent, deterministic system or workflow. This model-agnostic architecture lowers the unit cost of each matter without compromising the judgement, audit trail or accountability the law requires.
Intelligence is becoming abundant. What matters is where human judgment sits in relation to everything else. At the top: the practitioner’s experience, intuition and accountability. Below that: frontier models, lower-cost models, classifiers, deterministic systems and workflows. Galley keeps human judgment at the level where it creates the most value, and routes everything else accordingly.
We are reimagining not only where lawyers spend their time, but what clients experience throughout a matter: cost transparency from the outset, richer communication at every stage, and an audit trail clients can follow. The goal: a legal service that works better for everyone, with lawyers concentrated on the work that genuinely requires their expertise.
Rob Curtis spent two decades inside regulated industries. At Daylight he brought together product, technology, operations, compliance and customer experience to build a financial service from scratch. That pattern (decompose, build for the regulator, ship) is what Galley applies to legal services.
Galley is where the thesis becomes something we can test. We are assembling best-in-class solutions across the full capability stack (from client onboarding through agentic workflows to billing and financial management) to find out what it actually takes to build the best version of this firm. That means learning in the open, in real conditions.
Rob Curtis is a founder and transformation operator based in London, with a 20-year career building and redesigning businesses in regulated industries.
He tends to be drawn to moments when technology changes what is possible faster than established operating models can adapt. His approach is to go back to first principles: understand the customer, decompose how the work actually gets done, and rebuild the product and the organisation around the capabilities now available.
Rob previously co-founded and led Daylight, a US consumer fintech built to rethink banking for LGBTQ+ people. The team launched a regulated financial product in 90 days, raised approximately $25m, and took the company through acquisition in 2023. Earlier work includes major transformation programmes at Prudential, a restructuring and JV exit at Gaydar, and advisory work with technology founders and investors.
Now based in London, Rob is applying that experience to legal services. He is building Galley Law as a reference implementation of an AI-native law firm: working from the matter level up to understand how legal work can be decomposed across lawyers, AI agents, deterministic systems and workflow while preserving the judgement, controls, evidence and accountability the law requires.
He also writes Exchange Rate, a Substack on the AI decisions nobody wants to make.