For personal injury law firms
Your leads answered,qualified, and booked.While you work.
Tallyrun builds production AI agents for personal-injury law firms and keeps them running afterwards. The first one, Tally, handles intake and qualification, and it is live on this page right now. Go say hi.
CASE FILE · builds as they talk
- Incident
- ·
- Injury
- ·
- Treatment
- ·
- Triage
- ·
Case file emailed to the attorney. Elapsed: 94 seconds, at 1:47 a.m.
The real system, running live on this page. Watch the full walkthrough
Agents we build
The conversations your team cannot get to
Shaped to your firm
Tally isn't rigid software. We tune it to how your practice actually works: your qualifying questions, your red flags, your after-hours handoffs. By the end it is shaped like your firm.
- Your qualifying rules, asked in your order
- After-hours and overflow, routed your way
- Case files formatted for your inbox and CRM
- Booking that matches your calendar
- Conflicts and case types you refuse, screened out before they reach you
Engineered, not prompted.
An agent that talks to your clients is production software. We build it that way.
- Rules decide, the model talks
- Deterministic rules own triage, scoring, and compliance. The language model owns the conversation. Your logic never hallucinates.
- No single point of failure
- Provider failover is built in. A rate-limited API degrades the wording, never the outcome. No lead is dropped because a vendor had a bad night.
- Adversarially tested
- Prompt injection, hostile users, ramblers, people who type one word per message: we attack our own agents before your clients ever meet them. Every claim in a case file is traceable to what was said.
- Documented and owned by you
- Clear logic, clear handoffs, written down. Someone new to your firm should be able to read how intake works and follow it.
A qualified lead lands where you already work: Slack · email · Cal.com, or a signed webhook into any other CRM
How we work
You will know the price before anything gets built.
You watch the agent work on real cases before it goes anywhere near your clients.
- 01
A 30-minute call
We map where inquiries leak out today: the missed calls, and the leads nobody gets around to qualifying.
- 02
A fixed-scope proposal
One document covering what the agent does and what it costs, with a date it ships on. Nothing is billed by the hour.
- 03
Build and prove
A first working version in your hands within days. You watch it handle real cases in a controlled run before it goes near your clients.
- 04
Run and improve
Optional monthly care: monitoring, fixes, and improvements as your process changes. Or take the keys, it is documented either way.
One engineer.
No layers.
I'm Mihail Kenarov, an engineer in Eindhoven, and Tallyrun is just me. I write the conversation, the qualification rules, the integrations and the monitoring, then I watch the thing run once it is live on a firm's site.
I spent a few years building internal software inside other companies, which is where I learned that most intake problems are not really AI problems. They are a firm getting back to someone at nine in the morning who wrote in at eleven the night before, by which point that person has called two other firms. Tally exists to close that particular gap. The honest way to judge it is to open the demo and try to break it.
When you book a call, you talk to me. When we build, I write the code.
Mihail Kenarov · founder & engineer · Eindhoven, NL
Common questions
Asked before booking
Something else on your mind? Email mihail@tallyrun.ai and the person who builds the system will answer.
What is Tallyrun?
Tallyrun is an AI automation studio based in Eindhoven, the Netherlands. We design, build, and run production AI agents for personal-injury law firms: intake, lead qualification, and follow-up. Our first agent, Tally, answers, qualifies, and books incoming leads. A live version runs on this site, so you can test the product before you ever talk to us.
What does Tally, the intake agent, do?
Tally picks up a website inquiry the moment it arrives, including at two in the morning on a Sunday. It gathers the facts of the case, applies your firm's own qualification rules, books qualified leads into your calendar, and delivers a structured case file to your inbox. Deterministic rules make the triage and scoring decisions while the language model only carries the conversation, so your intake logic never hallucinates. It was built first for personal-injury firms, where the first firm to respond usually signs the client.
Is Tallyrun legit?
Fair question; check for yourself. The intake agent we sell is deployed on this website, and you can put it through its paces right now. Our terms of service and privacy policy are published, our contact details are public, and you speak to the founder himself on every call. If you want to verify something before booking, email mihail@tallyrun.ai and he will answer you directly.
Who is behind Tallyrun?
Mihail Kenarov, an engineer based in Eindhoven, the Netherlands. Tallyrun is a one-person studio on purpose: the person who scopes your build is also the one writing the code and picking up the phone when something needs changing, so nothing gets handed off to an account manager who has never seen the system. His background is in enterprise software, including an internal ERP agent at the Dutch software company DigitalWorkmate and implementation work at Abagar AD in Bulgaria. He can be reached at mihail@tallyrun.ai without going through anyone.
Is my clients' data safe?
Client conversations are handled under our published privacy policy. Tally runs on Anthropic's Claude API, which does not train on the data it processes, and case files are delivered into systems your firm already controls: your inbox and your calendar. We operate from the Netherlands under EU privacy law, and the privacy policy lists every sub-processor we use.
How do we get started?
With a 30-minute call. We go through how inquiries reach your firm today and where they stall on the way to a signed client. You then get a fixed-scope proposal, one document that says what the agent does, what it costs, and when it ships. A first working version is in your hands within days, and you watch it handle real cases in a controlled run before it goes near your clients. Pricing is fixed and agreed before we start.
Start with the conversation you keep missing.
Bring us the calls, forms, or follow-ups that slip through. Thirty minutes, no obligation.