We opened applications for the Referent private beta in mid-July. We closed them on August 7 - just under four weeks later - with more than 450 applications from lawyers and law firms competing for 20 seats. That is over 22 applications per seat, for a product in private beta, from a profession with a reputation for adopting software slowly.
This post is what we learned from those applications: who applied, what their firms look like, what they say is actually broken in daily practice - and what happens next for the 430+ firms that did not make Cohort 1.
What happened in four weeks
The application was deliberately not a marketing form. The short version asked who you are and how your firm runs; the extended version went deeper into intake, tooling, admin hours and what you would never hand to an AI. Applications were read by hand, scored for operational fit, and the strongest fits went through a 20-minute interview before the founding team made final calls.
The demand curve did not flatten. Our single busiest day - 46 applications - came in week three, not week one. Interest compounded as lawyers shared it, which tells us this is not launch-spike curiosity; it is a standing need looking for an answer.
Who applied: the numbers
A few numbers from the applicant base, as of the August 7 close:
- 85% from the US and UK. Among applicants whose location we could verify, 62% practice in the United States and 23% in the United Kingdom. The largest US concentrations: New York, Miami, Los Angeles, Houston, San Francisco and Chicago.
- More than 8 in 10 are firms of five lawyers or fewer. Of applicants who told us their firm size, two-thirds are solo practitioners and another sixth run firms of two to five. This is exactly the segment where half of US private-practice lawyers actually work - and where AI adoption has historically been lowest.
- 42% applied from a firm-domain email address - established practices with their own infrastructure, not side projects.
- 37% are owners, partners or directors. We matched 114 applicants against public professional data; of those, more than a third hold the top decision-making role at their firm. The people applying are the people who sign.
Practice areas skewed toward high-repetition work: immigration, family law, estate planning, personal injury, employment, and small-business law - fields where the same operational motions repeat matter after matter, and where automation compounds fastest.
What applicants told us is broken
The extended applications and interview calls read like one story told 450 ways. Three patterns stood out:
Intake is the bottleneck, and it is manual almost everywhere. The most common description of new-client intake involved retyping the same facts from email into a spreadsheet, a document and sometimes a practice-management system - by the lawyer personally. For solo firms, the person doing intake at 9 pm is the same person due in court at 9 am.
The competition is chaos, not software. We expected to hear about switching costs from established platforms. Instead, the most common “current system” was a combination of inbox, calendar and spreadsheets. Many firms that do pay for practice-management software described using it as a filing cabinet - records go in, but the work of moving matters forward still happens by hand.
Lawyers want leverage, not replacement. Almost every interview drew the same line: automate the operational work - intake, follow-ups, deadlines, status updates - but nothing client-facing goes out without a lawyer’s sign-off. That line is Referent’s design principle, so we heard it as confirmation: AI-native means the firm runs on AI agents, while judgment stays with the lawyer.
What happens with Cohort 1 now
Invitations to the selected 20 firms are going out now. Onboarding starts August 10, and the first cohort runs as a four-week program: personal onboarding calls, first matters live in week one, weekly check-ins, and fixes shipped from cohort feedback throughout. We will publish what we learn - including the numbers on time saved per firm - as the cohort progresses.
Twenty is not a growth-hacking scarcity number. It is the count of firms we can onboard personally and give weekly attention while shipping fast. That attention is the product in beta.
Cohort 2: the waitlist is open
If your firm was not selected - or you are reading this after the close - the Cohort 2 waitlist is open, and it is not a polite rejection bin:
- Waitlisted firms get first access when Cohort 2 opens in September, in queue order, before anything opens publicly.
- Every applicant gets early-access perks during the open beta - including locked-in founding pricing that will not be offered again after launch.
- Completing the extended application moves you up the queue. It takes about three minutes, and it is how we spot strong operational fits early - strong fits go straight to interview.
- When a Cohort 1 spot frees up, we pull from the top of the waitlist. It happens; be at the top when it does.
Methodology note
All figures are first-party data from Referent beta applications received between mid-July and the August 7, 2026 close, counted after removing test entries and duplicates. Location figures are the share of applicants whose location we could verify from application data (189 of the base); firm-size figures are the share of applicants who self-reported firm size (248). Seniority figures come from matching 114 applicants against public professional-profile data (People Data Labs). Qualitative findings summarize extended-application answers and 40+ recorded interview calls; no individual applicant data is disclosed.
Frequently asked questions
Can I still get into Cohort 1?
Cohort 1 is full, but spots occasionally free up during onboarding. When one does, we pull from the top of the Cohort 2 waitlist - so the fastest path is to apply and complete the extended application, which moves you up the queue.
When does Cohort 2 open?
September 2026. Waitlisted firms get first access, in queue order, before we open anything publicly.
What do waitlisted firms get in the meantime?
Every applicant - not just the selected 20 - gets early-access perks during the open beta, including locked-in founding pricing. We email the details to the waitlist before launch.
How were the 20 firms selected?
Every application was read by hand. We scored operational fit - practice area, firm size, how the firm runs intake and follow-ups today - and interviewed the strongest fits on a 20-minute call before final selection by the founding team.
Why keep the cohort so small?
Because we onboard every firm personally. Twenty firms is the number we can give weekly attention to while shipping fixes from their feedback. Cohort 2 grows only as fast as that stays true.