K
Kalinda
Kalinda is an AI intake agent for personal injury law firms from a San Francisco company founded by Sayan and Sohil Bhatia, two former AI researchers, and part of Y Combinator's Summer 2025 batch. It answers inbound injury calls within seconds and calls back web form, referral, missed and after hours leads by voice and text. It runs the firm's intake questions, qualifies the case against the firm's rules and sends the retainer for signature.
The company launched with a different product, which extracted and summarized medical, financial and employment records to qualify claimants for class action and mass tort firms; its public website now presents only the intake agent.
Capability grades
All 15 axes, graded from public sources on the date shown. Hover a grade to see what the letter means on that axis.
AI Centrality
How much of the product is actually AI. Whether the machine learning is the mechanism the buyer is paying for or a feature layered onto conventional software, and whether the vendor is specific about which is which.
Kalinda is an AI agent for personal injury firms that answers an injured caller within seconds, calls back every web form, referral and missed call by voice and text, runs the firm's own intake questions, qualifies the case against the firm's rules and sends the retainer for signature, all in one conversation. The home page describes it as putting the firm's best intake specialist on every lead. Its animated demo follows a single call from the agent's opening, through the accident, injury, treatment and insurance facts, to a retainer on the caller's phone and a line reading retainer signed and matter created in the CRM.
No separate case management, dashboard or workflow product is sold around it; without the agent, nothing remains to buy. The company's accelerator profile describes an earlier product, AI review of medical and product records for class action and mass tort firms, which is not what the current site sells.
Citation Accuracy and Hallucination Disclosure
Whether the vendor publishes measured accuracy on citations and assertions, grounds output to primary sources, and says plainly what its system does when it does not know. Legal has a documented public record of fabricated citations reaching filed briefs, so an untested claim of accuracy is not evidence.
In the home page demo, after hearing that a caller was rear ended, has neck and back pain, was checked by an ambulance and has a police report, the agent says that all helps the case and tells the caller they have a strong claim, before any lawyer has seen the facts. Nothing is published on how often the agent's qualification judgments are wrong, how accurately it records what callers say, what its view of a claim's strength rests on, or how a firm can audit it.
No accuracy figure, test, sample of reviewed calls or error rate appears anywhere. Nothing says whether the strong claim line reflects the product's actual behavior or a scripted illustration.
Autonomy and Oversight Model
What the system decides on its own, what a lawyer must approve, and whether the vendor documents where the review point sits. A tool that drafts under review and a tool that files without one are different products and different risks.
No oversight structure is published. The agent acts for a firm with the public and ends the call by sending a contract: Kalinda's home page says it runs the firm's exact intake, qualifies the case against the firm's rules and signs the retainer. The demo shows it moving from the caller's first words to a retainer on their phone with no one from the firm involved. The public site does not describe which cases must go to an attorney before a retainer is sent, what the agent may not say, how many times it may call or text a lead, when it stops, or how the firm reviews a call afterward.
The retainer is the point at which a prospective client becomes a client, and nothing published says whether a lawyer makes that decision.
Operational and Outcome Evidence
Named, dated evidence that the product works in production at real firms or legal departments. Case studies with figures and identified customers count. Unattributed testimonials and launch announcements do not.
No production evidence for the intake agent is published. The public site names no customer, quotes no lawyer, reports no answer rate, contact time or signed case figure, and carries no case study; the firm in its demo, Reyes Law, is part of an illustrated call. The accelerator profile, from the Summer 2025 batch, says the company had processed 600,000 pages of records and was live with two of the largest plaintiff firms in the United States, unnamed, but that describes an earlier product that reviewed medical and product records for class action and mass tort qualification, not a voice agent answering intake calls.
Privilege and Confidentiality Posture
How client confidences are handled: attorney client privilege and work product treatment, segregation of one client matter from another, whether client data trains any model, and what the vendor commits to in writing rather than in marketing.
Nothing is published on how confidences are handled. The product hears an injured person describe their accident, injuries, treatment and insurance before any engagement exists, the information a lawyer's duty to prospective clients protects. The site carries no privacy policy, terms of service, data processing terms or security statement, and no footer link to any of them. Nothing says who at Kalinda or its providers can hear recordings or read transcripts, how long they are kept, whether any model is trained on them, whether one firm's callers are separated from another's, or what happens to a recording when the firm stops using the service. The site does load a visitor identification script for its own marketing, with no matching disclosure.
UPL and Professional Responsibility Posture
Whether the vendor is clear that it supplies a tool rather than legal advice, who its audience is, and how it addresses unauthorized practice of law, competence and supervision duties, and jurisdiction limits. ABA Formal Opinion 512 is the reference point. Where the advice line is not the duty a product raises, the axis is read through the nearest professional duty it does raise: judicial conduct rules and the reviewing duty for products sold only to courts, and the duty to bill for time actually spent for products that draft time entries.
Nothing is published on the advice line, and the product speaks to the public about their claims on a firm's behalf. In the home page demo, the agent answers with the firm's name, tells an injured caller that the facts help their case and that they have a strong claim, and texts a retainer to sign during the same call. That is an assessment of a legal claim delivered by software to someone who is not yet a client. No page says what the agent may and may not tell a caller.
Nor does any page say that it is not a lawyer, that it is an AI rather than a staff member, or that an attorney reviews the matter before the firm takes it on. Nothing addresses consent for automated calls and texts to web form and referral leads, the solicitation and advertising rules for contacting prospective clients, or recording consent. Those obligations fall on the firm, and the vendor says nothing about helping meet them.
AI Governance and Bias Disclosure
Published governance over model behavior: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.
No governance position is published. The founders' letter, from two brothers who describe themselves as former AI researchers, says AI should protect the human heart of this work rather than replace it, and that no one who reaches out after being hurt should be met by a voicemail or left waiting in a queue. That is a statement of purpose, not a mechanism. Nothing names who is accountable for what the agent says to callers, what is tested before a change to its voice, script or qualification logic reaches firms, or how calls are reviewed for quality.
The agent decides in real time which callers qualify and which get a retainer, and nothing published addresses whether it treats people differently by language, accent, age or how they describe their injuries.
AI Safety and Data Stewardship
Retention, deletion, access control, and what happens to prompts and documents after they are processed. Whether the vendor states its subprocessors and its incident practice, or leaves the buyer to assume.
Nothing is published on retention, deletion, access or incident handling for call recordings, transcripts and the intake data the agent collects. The public site carries no privacy policy, subprocessor list, security page or trust statement, and nothing published describes encryption, access controls, backup, breach notification or what happens to a firm's data when it leaves. The agent handles injury details, medical treatment and insurance information, often from people calling in distress hours after an accident.
The one data practice visible on the site concerns the site itself, a visitor identification script from a sales tracking vendor, which is disclosed nowhere.
AI Liability and Recourse
What the vendor stands behind contractually when its output is wrong. Indemnities, caps, carve outs, insurance, and whether any of it is published or only reachable through a negotiated agreement.
No terms of service or customer agreement is published, so nothing states who bears the loss when the agent is wrong. The failure modes are easy to name from the product's own description: a caller told they have a strong claim who does not, a viable case wrongly turned away by the qualification rules, a retainer sent to someone the firm had a conflict with or should not have taken on, or automated calls and texts that breach consent rules and draw a claim against the firm.
Nothing published says whether Kalinda warrants anything, carries insurance, indemnifies the firm, caps its own liability or offers any service commitment. The only route to the contract is booking a call from the home page, so the allocation of risk exists only in private.
Practice Systems Integration Depth
How deeply the product reaches into the systems legal work already lives in: document management such as iManage and NetDocuments, Word and Outlook, contract lifecycle management, matter management, e-billing, and court filing systems.
No integration is named or documented. Kalinda's home page says the agent works across inbound calls, callbacks to web forms, referrals, missed calls, after hours and overflow, by voice and text. Its demo ends with a line reading retainer signed and matter created in the CRM, so the product is meant to write into a firm's systems, but it does not say which. No case management or intake system such as Litify, Filevine, Lead Docket, SmartAdvocate or CasePeer is named, no phone system or call forwarding setup is described, and no electronic signature provider is identified for the retainers it sends.
There is no integrations page, API reference or setup guide, and what Kalinda may change in a firm's systems is not stated.
Deployment Model and Data Residency
Where the software runs and where the data sits. Multi tenant cloud, single tenant, private deployment, on premises, and whether region of residence is a published option or an enterprise conversation.
Nothing is published on where the software runs, where call recordings, transcripts and intake data are stored, or how firms are separated from one another. The product is plainly cloud delivered, a phone number a firm forwards its calls to, and the company lists San Francisco as its base on its accelerator profile. No hosting provider, region, tenancy model, private deployment or data location commitment is described.
Nothing says whether a caller's account of their accident is stored in the United States, whether it passes through providers abroad for speech or language processing, or whether one firm's callers' data sits in the same environment as other firms'.
Security Certifications and Trust Center
Independent attestation a buyer can pull without a sales call: SOC 2, ISO 27001, penetration test summaries, a trust center with current reports and named scope rather than a badge image.
No independent security attestation is published. There is no SOC 2 report, ISO certification, HIPAA statement, penetration test, trust center or security page, and no badge or claim of compliance appears anywhere on the site, including in the page footer. There is also no stated route to request security documentation short of booking a demo call. The company was founded by two former AI researchers and backed by Y Combinator from its Summer 2025 batch, and nothing about its security, current controls or plans for a SOC 2 audit is published for a product that answers a firm's phone line and collects injury and medical details from the public.
Model Supply Chain Disclosure
Which models sit underneath, whose they are, where they run, and whether the vendor commits to telling customers when that changes. A legal buyer inherits every dependency it cannot see.
Kalinda speaks with callers in a natural voice, understands their answers, texts them, and fills and sends retainers. No model family, model provider, speech or voice vendor, telephony carrier, electronic signature provider, hosting arrangement or inference location is named on the public site, and nothing commits to notice when any of them changes. The founders describe themselves as former AI researchers, and the accelerator profile shows their earlier product extracted and summarized medical records with AI.
A caller's account of their injuries passes through models that are not named. The site presents the agent as AI throughout.
Commercial Transparency
Whether a buyer can learn what this costs without entering a sales process: published rates, the unit being charged, what sits behind an enterprise tier, and what implementation adds.
No pricing is published at any level: no price, tier, unit of charge, minimum term or setup fee, and no pricing page. Every call to action on the home page leads to seeing the agent take a live intake call, in effect a demo booking. Nothing says whether Kalinda charges per minute, per call, per lead or per signed case.
Firm and Practice Coverage
Who the product is actually built for. AmLaw, midlaw, small firm and solo, in house departments, government and courts, and which practice areas are supported rather than merely claimed.
The site is written entirely for personal injury firms, from its title, AI intake that signs your next case, to its demo of a rear end collision call. It covers every way a lead arrives: live inbound calls answered in seconds, callbacks to web forms and referrals, missed calls, after hours and overflow, by voice and text. The founders' letter frames the mission around injured people reaching out for help. The company's earlier work, described on its accelerator profile, was for class action and mass tort firms.
Kalinda does not say which states' rules its calling and retainer workflows follow, which languages it speaks, whether it qualifies case types beyond vehicle accidents such as premises, medical malpractice or workers' compensation, or what size of firm or lead volume it suits.
1 public document
The public pages on file for Kalinda, with the recorded signals each one supports and the date it was last read. Open any of them and check the reading against the record.
-
kalinda.ai11 signals
Client Data in Training, Prompt and Output Retention, Ethical Walls and Matter Segregation and 8 more
Read Sep 27, 2026
Legal Signals
What each signal meansA signal records what public sources say on the date shown. It is not a grade and it is not a recommendation. Where a signal reads Not addressed, it means the index did not locate the material in public sources on that date, which is a statement about disclosure rather than about the product.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
No customer agreement, terms of service or equivalent contract is published on any surface located, and no policy page states a position on training. Nothing is granted and nothing is withheld, so a client has no term to hold the firm to. Where a policy page does state a position, the row takes the matching policy value instead and the summary records that no agreement exists.
No customer agreement, terms or policy page is published, and nothing on the public site states whether call recordings or intake data are used to train any model. Nothing is granted and nothing is withheld.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
No located public material states how long prompts and outputs are retained.
Nothing published states how long call recordings, transcripts or intake answers are kept.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
No located public material addresses walls or matter level segregation.
Nothing published addresses who at a firm can hear or read a caller's intake, or how conflicts are checked before the agent sends a retainer.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
No located term or policy addresses third party requests for customer data.
No term or policy addresses what happens when a court, insurer or government demands call recordings or intake records. No terms or privacy policy is published.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Sources are identified without stating the license or rights basis.
The agent works from the firm's own intake. The vendor says it runs the firm's exact intake and qualifies each case against the firm's rules, so no external legal corpus sits behind its decisions and no licensing question arises. What is not stated is how the agent reaches a view such as a caller having a strong claim when the firm's rules do not settle it.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
No located public material addresses whether authority is checked for subsequent history.
The product cites no legal authority, and nothing published addresses later history.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
No located public material addresses what the product does when it cannot ground an answer.
Nothing published describes what the agent does when a case does not fit the firm's rules, a caller asks something it should not answer, or it is unsure whether a case qualifies.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
No court order, opinion or disciplinary record addressing fabricated or hallucinated legal citations produced by this product has been located as of the date shown. This is a statement about the public record on that one subject, not a finding about the product, and this signal is not a litigation history.
The AI Hallucination Cases database maintained by Damien Charlotin records no case naming Kalinda. This is a statement about the public record rather than a finding about the product. The product does not generate legal citations.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
No located public material engages with bar or ethics guidance.
Nothing published engages with bar or ethics guidance. The agent speaks to injured people, assesses their claims and sends retainers, work that touches the rules on solicitation, advertising and a lawyer's responsibility for nonlawyer assistance.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
The product does not touch a fee between a lawyer and a client. It operates before an engagement exists, or it is bought by a team that bills no client for the work. Savings claims aimed at the buyer’s own cost are recorded in the summary and do not make the row a savings claim, because no client bill is in the loop.
The product works before an engagement exists and bills no client for its work. It handles intake up to the signed retainer, and the retainer it sends is the firm's own agreement. Nothing published claims savings for clients.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
No located public material supports a client side disclosure obligation.
No subprocessor or model provider list and no security or disclosure material is published.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
No located public material addresses court disclosure or verification certification.
Nothing published addresses a record of the agent's work that a firm could produce if its intake practices were questioned.