Wordsmith, an emerging Scottish legal tech startup that has managed to secure backing from two well-known venture capital firms, offers an AI platform aimed at in-house legal teams and law firms that can be set up to assist other employees within the company, allowing anyone in the company to ask for help with legal tasks such as reviewing contracts or answering specific questions about documents.
Founded in October last year, the Edinburgh-based company is the brainchild of former TravelPerk senior executives Ross McNairn (CEO) and Robbie Falkenthal (COO), and CTO Volodymyr Giginiak, who held various engineering roles at Microsoft, Facebook and Instagram. Six months after they left their previous roles, Wordsmith has already landed high-profile clients such as Trustpilot and is in partnership with at least one major law firm, DLA Piper.
This early traction caught the attention of global venture capital firm Index Ventures, which led a $5 million seed investment into Wordsmith alongside General Catalyst and Gareth Williams, founder and former CEO of Scottish tech unicorn Skyscanner.
That the Scottish startup secured the backing of two venture capital firms that have co-invested in Facebook, Slack, Sonos, Airbnb, Stripe, and Snap speaks not only to Wordsmith's early promise, but also to its founder's pedigree. Prior to TravelPerk, McNairn founded a travel management startup called Dorsai Travel, which he sold to Skyscanner just nine months after launch, where he became Skyscanner's head of product. He then joined second-hand shopping app LetGo, another unicorn, before joining TravelPerk.
Additionally, McNairn is a qualified lawyer, but left that profession after a few years to become a software engineer.
Legally Friendly
The legaltech space is hot: in the past six months alone, several “Copilots for lawyers” have emerged, such as Harvey AI in the US and Luminance in the UK. Other legaltech startups such as Definely and Lawhive in the UK have also raised decent seed and series A rounds of funding, as have Alexi (Canada) and Leya AI (Sweden).
While these companies approach the legal sector from different angles and geographical centers, they have one thing in common: they’re all riding the wave of generative AI.
Like other paper-heavy industries, legal professionals are looking for ways to automate repetitive, labor-intensive tasks so they can focus on more strategic tasks. This is where Wordsmith steps in, offering a generative AI platform that it calls “lawyer-in-the-loop.”
While Harvey AI is aimed at the lawyers themselves, Wordsmith is aimed at employees within a firm, with legal teams configuring the platform behind the scenes and connecting to their own data sources, allowing lawyers to act on an as-needed basis.
McNairn compares it to something like TravelPerk, which offers small businesses a self-service business travel management platform where managers can define policies and approval processes. Employees then make all of their own bookings within those parameters.
“At TravelPerk, [we made] “We went from selling the travel team a slightly better tool to speed things up, to letting other departments essentially book themselves,” McNairn told TechCrunch. “And then the travel team just managed and checked and made sure it was aligned correctly. This shift from building tools just for the sake of functionality, to building tools to help other departments work more efficiently, is a big change in the way we work.”
Firms can configure Wordsmith in two main ways: as an autopilot for simple matters that don’t require expert oversight, and as a copilot where a lawyer always gives approval before a formal answer is provided.
A typical workflow might include a sales rep needing to review a new contract, or a procurement department trying to close a deal and needing access to information like the company's security posture. These are pretty standard questions, the answers are unlikely to vary much. Anyone can query Wordsmith and get the information they need.
Wordsmith Query Image Credit: Wordsmith
Another potential use case is someone issuing a Personal Information Access Request (SAR) to a company. Companies in certain jurisdictions have a legal obligation to respond to requests for access to personal data. In this case, Wordsmith can be configured to accept the submission and connect to the company's ticketing system, responding with either the information requested, or a template response outlining timescales and next steps, as dictated by the company's internal guidelines and processes.
Model Behavior
Wordsmith uses a combination of underlying large-scale language models (LLMs), including OpenAI's GPT-4 and Anthropic's Claude.
“We use the right ones for the job,” McNairn said. “Some are good at analyzing things like the logic within legal contracts, some are very accurate at helping you reword things. Claude is good at streamlining problems, and OpenAI (GPT-4) is great across the board with all the different facets.”
Enterprises have expressed some trepidation about adopting generative AI, but McNairn said the company is addressing that in a number of ways, including allowing companies to stipulate that their data will not leave the EU. It has also committed to not training its AI with enterprise data. Wordsmith configures a “private instance” for enterprises, meaning it will connect to the data wherever it is (e.g. Google Drive or Notion) and use the enterprise's own data to improve responses, but this data will not be used to train other enterprises' models.
“We use a technique called RAG (Search Augmentation Generation),” McNairn says, “so we're not training on their data, we're only using it when we need to. We remember it, use it to enrich the answer, and then respond to them.”
high frequency
While strengthening in-house legal teams will be Wordsmith's main goal initially, the company is also looking to partner with law firms, as seen in its early partnership with DLA Piper, in which DLA, a $1 billion global legal giant, is partnering with Wordsmith to co-develop an AI agent that it hopes to distribute to its own clients.
So, in effect, they're putting their own technical knowledge into refining Wordsmith for a very specific area of law, something that could potentially be sold as a new type of legal service, perhaps at a lower fee.
“This kind of leveraging of corporate knowledge is more frequent and less costly than paying thousands of dollars an hour,” McNairn says. [also] It’s a much better way to show they’re progressive and looking to adopt AI.”
This business model is especially suited to small to mid-sized law firms that can leverage Wordsmith to land bigger jobs and take on more clients.
McNairns said the service is still in the early design stages with DLA, but Wordsmith will likely commercialize it soon. “We're not there yet,” he said.
With $5 million in the bank, McNairn says Wordsmith intends to accelerate hiring in both Scotland and the U.S. The company currently has nine employees, several of whom are based in London, and is on the way up the ranks, but McNairn says he wants Edinburgh to be the company's centre of gravity.
“It's about the ecosystem, which I'm very passionate about,” he said. “I've been involved with three unicorns so far and I just want to build something cool in Scotland.”