How Dynamis Builds Its Own Legal Technology with Claude Without Hiring Developers

When Eric Rosen looked at Boston’s litigation market in 2023, he noticed that many of the city’s strongest independent litigation boutiques were joining larger national firms. Clients facing serious cases had fewer options for working directly with experienced trial lawyers outside Big Law.

Eric, a former federal prosecutor who led the Operation Varsity Blues investigation, and Constantine Economides founded Dynamis LLP to fill that gap. Their ambition was to build a national litigation boutique that could handle complex cases at more affordable rates, without adopting the staffing model of a large law firm.

Today, Dynamis operates across five offices, with lawyers licensed in approximately 25 states. The firm takes on white-collar criminal defense, government investigations, and complex civil litigation, often against much larger opponents.

Technology is central to how Dynamis competes. Rather than hiring developers or purchasing specialized software for every task, the firm gives every lawyer access to Claude. Eric and Constantine build many of the firm’s tools themselves, from litigation budgets and document review to financial reporting and federal sentencing research.

“For us, technology allows us to compete for bigger cases without having to scale up in manpower. It’s allowed us to do a lot more with far less.”

Eric Rosen, Founding Partner

Dynamis Builds Its Own Legal AI Tools Instead of Buying Them

Eric receives several pitches a week from legal AI vendors. Many offer tools that, he says, lawyers can now build themselves.

Two years ago, developing software required programming experience or outside technical support. Eric had taken programming classes in college, which gave him a foundation. Today, he describes what he wants to build, and Claude writes the Python code.

“Now it creates literally the Python programs and everything for you. We create all our own apps.”

Eric Rosen, Founding Partner

The firm has made Claude its standard AI platform. Eric prefers it for analytical work and its ability to connect with other tools, after experimenting with alternative models.

“I found Claude a lot better. It’s just intuitively a lot better, and the connectors are better.”

Eric Rosen, Founding Partner

This approach has changed what the firm builds internally. Instead of evaluating a new vendor each time lawyers encounter a repetitive task, Eric and Constantine can explore whether they can solve the problem themselves.

The result is a growing collection of tools built around the firm’s own work, data, and litigation experience.

How Dynamis Built a Federal Sentencing Database with Claude

In federal criminal defense, sentencing arguments often depend on what happened in comparable cases.

The U.S. Sentencing Commission publishes detailed sentencing data, but the underlying files are difficult for lawyers to use. They contain thousands of variables and require specialized software to analyze.

Eric used Claude to help turn that data into the Federal Sentencing Explorer, a free public tool covering 805,875 federal sentences from fiscal years 2014 through 2025. He built it in about three to four hours.

Lawyers can filter cases by offense type, offense level, criminal history, federal district, and cooperation status. The tool returns sentencing distributions, median outcomes, and a summary that can be used in preparing a sentencing memorandum.

Rather than relying exclusively on sentencing guidelines or searching individual decisions, defense lawyers can see what defendants with comparable circumstances actually received.

The tool is available to lawyers outside Dynamis, making the firm’s analysis accessible to the broader defense bar.

Claude Builds Litigation Budgets From 1,800 Past Invoices

Prospective litigation clients want to know what a case is likely to cost. But litigation budgets are difficult to estimate because the work depends on how a matter develops, what opposing counsel does, and how much discovery is required.

Dynamis wanted to base those estimates on its own experience rather than general assumptions.

The firm loaded 1,800 anonymized historical invoices into Claude and built a tool that uses past matters to generate budgets for prospective clients.

“Instead of just winging it, we fed our 1,800 invoices into the system and use AI to create budgets based on what’s actually been done in past cases.”

Eric Rosen, Founding Partner

The tool gives the firm a way to draw on its billing history when discussing costs with clients. It also makes the knowledge accumulated across previous matters more accessible when pricing new work.

The underlying data comes from the firm’s financial systems, including TimeSolv, which Dynamis uses for invoicing and accounting. The firm also uses PointOne to capture lawyers’ time automatically, helping ensure more of their work is reflected in billing records.

For Dynamis, these systems form part of a broader effort to use its operational data to make better decisions.

AI Document Review Without a Large E-Discovery Team

Large litigation matters often involve thousands or millions of documents. But not every production justifies the expense or setup of a full e-discovery platform.

For smaller productions, Dynamis built its own review tool.

The application analyzes documents, identifies relevant information, tags material, and organizes events into chronologies. For productions of approximately 1,000 documents or fewer, Eric says the analysis can take around 20 minutes.

“If you get something from the government, within 20 minutes you can analyze it and figure out if it’s useful or not.”

Eric Rosen, Founding Partner

That speed changes how the firm approaches incoming discovery. Instead of waiting for a lengthy review process, lawyers can quickly understand the contents of a production and decide where to focus their attention.

Dynamis has also developed tools for trading-pattern analysis in market-manipulation matters, work that previously required outside consultants.

The goal is not to eliminate lawyer review, but to give experienced litigators the information they need without relying on larger teams for the initial analysis.

Claude Extends Beyond Legal Research Into Firm Operations

Dynamis also uses Claude to build tools for running the business.

The firm’s internal applications include a searchable library of past briefs and an intranet that connects to TimeSolv through its API to track accounts receivable and financial information.

These tools give the partners access to information that would otherwise require manual reporting or separate software.

The same approach applies across the firm: identify work that takes too long, determine what information is needed, and build a tool around it.

Not every task is a good fit for AI. Eric remains cautious about using Claude to produce the first draft of a legal brief. He finds it more useful for developing arguments, reviewing ideas, and refining work that lawyers have already prepared.

That distinction matters at Dynamis. The firm uses AI to expand what its lawyers can accomplish, while keeping legal judgment and responsibility with the attorneys.

The Future of Litigation at Dynamis

Dynamis was founded on the idea that a litigation firm could operate nationally without growing into the traditional Big Law model.

Claude has given Eric and Constantine a way to put that idea into practice. The partners can build tools for their own cases, use historical firm data to inform decisions, and bring work in-house that previously required outside technical specialists.

Eric expects that role to expand. He believes AI-generated analyses that can be independently checked and tested may eventually become admissible evidence, changing how firms prepare trading studies, call chronologies, and other technical work for litigation.

For now, Dynamis is focused on finding the next task that technology can make easier.

The firm’s advantage is not a single AI application. It is the ability to build new ones as the work changes, without waiting for a vendor or adding an engineering team.

“[Claude] has allowed us to do a lot more with far less, and it’s really just changed the landscape.”

Eric Rosen, Founding Partner