Caseflood vs Intaker: how they compare in 2026

AI drives much of the work inside an intake operation that people run, and the company is explicit that the people stay on the calls. Caseflood's AI makes dialing decisions on proprietary dialing infrastructure, texts leads to get them on the phone, prepares reps with what the caller already said, offers real time guidance that a rep can refresh mid call, captures and qualifies case details, handles the follow through after the call, and answers leaders who ask what is happening in the intake team and what to do next. An optional personalized voicemail is offered separately. The FAQ answers the question of whether clients will talk to an AI voice agent with a plain no: the firm's reps hold the human conversations. The company also sells itself as much on expertise as on software, with a hands on engagement that audits the operation, configures the platform and coaches the team. Remove the models and a dialer, a call workspace, retainer tools and dashboards would still run.

Intaker is a scripted chat and CRM product with AI capability layered across it. The chat runs on more than 1,400 custom prompts and prebuilt scripts for different practice areas, which is written conversational content rather than generated conversation. The surrounding platform, lead management, texting, follow up sequences, virtual inbox workflows and reporting, is fixed automation. The stated AI capabilities are natural language processing, language detection and automated responses, and both the vendor and independent directories describe the intake automation as AI driven. Without the models, a working scripted chat widget, CRM and follow up engine remain.

Caseflood cites no law. The relevant accuracy questions are whether the guidance a rep reads during a call is sound, whether the case details the AI captures match what the caller said, and whether its qualification judgments sort leads correctly, because a lead wrongly screened out is a case the firm never signs. The site's own pitch counts cases turned down too soon and leads marked with the wrong status among the losses it fixes, which makes those error rates the natural measure of the product. Nothing is published on the accuracy of guidance, captured details, qualification or call summaries, and no test, sample or error rate is described. The one control mentioned is that reps can review and lock captured information before it moves on.

A chat running on 1,400 written prompts produces less free text than an adaptive conversational agent. Nothing addresses what the AI layer does: no accuracy figure for language detection, response selection or qualification, no error rate and no evaluation is published. Nothing says what happens when a prospective client's message matches no written prompt.

People hold the conversations and the firm writes the rules. Caseflood's FAQ says the firm's intake reps handle every human conversation while AI supports dialing, texting, preparation, guidance during calls and follow through. The firm defines the cases it wants, the matters that need attorney review, its referral relationships and how it wants to communicate. Reps can review and lock captured information, and firm judgment stays central to qualification and representation decisions. Leaders get a view of where cases are lost and coaching tools for reps. The design is silent on the AI's own outbound work. It texts leads to get them on the phone and handles follow through after calls. Nothing says what those messages may say, whether a rep approves them, how often a lead may be contacted, or when the AI must stop and hand the lead to a person.

Live Call Connect, a click to call widget, turns a chat lead into a live phone transfer to the firm, so the prospective client reaches a person while still engaged rather than after a record is handed off. The video feature puts a person in the conversation another way: an attorney records intro and outro videos that play in the chat, so a named person is visibly present in an automated conversation. There is no statement of what triggers a call transfer or whether only the caller can start one, and no threshold on qualification. Nothing says what the chat does outside business hours when no one can take a transfer, and there is no override or review path for a qualification decision.

Named firms with specific results, on the clients' own account. The home page features Schuerger Shunnarah Trial Attorneys, a firm of more than 200 employees across 22 states. It reports about 300 more cases signed a month from the same demand and 50 fewer cases lost after signing each month. Video from cofounder Robert Schuerger and two intake reps accompanies it, and one of the reps signed 23 cases in her first week. Lawrence & Associates, with more than 50 employees across two states, reports a rise from 60 to 140 cases signed a month, and Justin Lawrence says speed to lead fell from about ten minutes to two seconds. The Schuerger Shunnarah story also credits Caseflood with hiring and training ten reps, turning a lemon law division into referral income and advising on marketing vendors, so the results reflect the engagement as a whole, not the software alone. The figures are labeled client reported with individual results varying, and none is dated or tied to a stated measurement period.

Several independent software directories list the product, with provider data stated as verified and reviews moderated for authenticity. There is also a listing in a named legal technology publication, a published starting price and a corporate entity, Intaker, Inc. Directory testimonials are attributed to roles rather than named firms, with three user reviews on one major platform. No law firm is named in vendor material, and no case study, usage figure, funding announcement or outcome claim with a figure is published.

Caseflood handles the first conversation between an injured or accused person and a law firm, which is where a prospective client's confidential facts first enter the firm's records and where the duty of confidentiality to prospective clients attaches. No customer agreement, product privacy policy or data processing terms are published. The only privacy document, updated 17 September 2026, covers the marketing website: it discloses PostHog session recordings that can capture form text before submission and RB2B visitor identification, with an opt out and respect for Global Privacy Control. Nothing addresses who can see call recordings and captured case details, whether they are used to train models, how long they are kept or how firms are separated from one another. The notice is candid about the website; the product it sells has no equivalent.

The vendor material carries no confidentiality statement, encryption reference, data handling description or treatment of privilege or prospective client confidences. Two way business texting means prospective client communications travel by SMS, an unencrypted channel outside the firm's control, and the platform centralizes those communications alongside chat transcripts and lead records in a vendor held CRM. Nothing distinguishes the status of a chat transcript from a prospective client who never becomes a client.

Firm judgment is written into the product design, and the professional responsibility questions around it are not addressed. Caseflood's FAQ says the firm decides which matters need attorney review and that firm judgment stays central to qualification and representation decisions. Because reps rather than an AI voice hold every conversation, the prospect is always speaking with a person, which removes a common disclosure problem for intake tools. The questions around the conversation remain open. Nothing states that Caseflood gives no legal advice, although its live guidance tells reps what to ask and say to a prospective client. The AI texts leads directly, and nothing says whether those messages identify themselves as automated or how they comply with consent rules for automated texts and calls. Nothing addresses call recording consent across the 22 states one client operates in, or the attorney advertising and solicitation rules that apply to rapid outbound contact. No customer agreement exists to allocate any of this.

The chat qualifies prospective clients against criteria for specific matter types using more than 1,400 written prompts, so the system has a substantive exchange about a person's legal problem and reaches a qualification outcome before any lawyer is involved. Nothing says the chat does not give legal advice, describes what automated responses may say when a prospective client asks a substantive question, or tells the person whether they are talking to software or a human, though the video feature presents a named attorney inside an automated conversation. No bar guidance is engaged.

No AI governance position, accountable owner, testing before release, or finding on how the qualification guidance treats different callers or case types is published. The company describes configuring the platform with each firm and refining it over the first two months, with results observed against starting goals in the third. That is an operating cadence rather than a governance program. The question is concrete here. An AI that decides whom to dial first, suggests which leads qualify and flags cases turned down too soon can treat callers unevenly by language, accent, age or type of injury, and nothing says whether anyone checks. The engineering cofounder's stated conviction that sales should remain human is a design principle, but no policy, review or measurement follows from it.

No AI policy, model card, bias or fairness testing, evaluation method, accuracy monitoring, drift statement, named governance body, ISO 42001 or EU AI Act position is published. The chat qualifies leads automatically, so it decides which inquiries reach a firm, and nothing indicates whether qualification outcomes have been examined across different groups of people. No evaluation of language detection is published, though a misdetection at first contact affects whether a prospective client who does not speak English can proceed at all.

Nothing is published on how call recordings, texts and captured case details are encrypted, who at Caseflood or its engineering team can access them, how long they are kept, how they are deleted when a firm leaves, which subprocessors handle them, or how and when a firm would be told of a breach. The platform processes injury details, medical history and criminal allegations for firms with tens of thousands of leads a year, and none of those stewardship commitments is in writing. The website privacy notice shows the company can write clear disclosures, naming its analytics and visitor identification providers and explaining how to opt out, but it covers only the marketing site. Caseflood's FAQ says integrations pass only what should go downstream, which is the single stewardship statement about the product.

Nothing states whether chat transcripts, text message threads, lead records or qualification outcomes are used to train or improve models, no retention period is published, and no deletion right is described. The platform is a CRM as well as a chat widget, so it holds a firm's prospective client base durably by design, and two way texting adds a communications archive on top of it. The vendor pages and directory listings carry no privacy policy or data processing statement.

No customer agreement is published, so nothing states Caseflood's liability, warranties, service levels or recourse. The FAQ describes the relationship as both software and an ongoing engagement, with Caseflood diagnosing the operation, configuring the tools, executing changes and assessing results. The case study credits it with hiring and training reps and advising on marketing spend. That makes the allocation of risk more consequential than for software alone. If guidance leads a rep to turn away a viable case, or automated texts breach calling rules, a firm cannot tell from the public record whether Caseflood shares any of the consequence. Every route to a contract runs through a free intake audit call, so the terms exist only in private.

No position on liability for AI output, warranty, service levels or remedy is published. Automated follow up sequences send personalized email and text reminders to leads on the firm's behalf, so the vendor's system sends outbound messages to consumers, and texting prospective clients engages consumer protection and messaging rules regardless of AI. Nothing published addresses responsibility for the content or sending of those messages, or for an inquiry wrongly screened out, which the firm would not see.

Integration is described as a service performed at setup rather than a set of named connections. Caseflood's FAQ says it connects the intake workflow with the firm's lead sources, case systems, documents and partner processes, and that the mapping is done during the first week of the engagement, including what should be passed downstream and which tools reps need inside the call workspace. Reps work in that workspace with everything the caller has already said, and signing the retainer happens inside the product. For plaintiff firms, whose leads arrive from many vendors and whose signed cases move into case management systems such as Filevine, Litify or CasePeer, the connections are the product's spine, and none is named. There is no integrations page, no list of supported lead sources, CRMs or electronic signature tools, and no statement of what moves in which direction. A firm learns what connects only during the audit.

Clio and Filevine are named for practice and case management, Salesforce as the general CRM, Calendly for scheduling, and Google Analytics and Google My Business for traffic and listing attribution. Plaintiff and high volume consumer firms commonly run Filevine, and many consumer legal inquiries start from a local listing rather than the firm's own site. No integration depth is described for any named system, nothing says whether lead data flows one way or both ways, and no API documentation is published.

No hosting provider, region, tenancy model or deployment option is published. The product is plainly cloud delivered, a dialer and call workspace that reps log into, configured with each firm during the first week of an engagement. Beyond that the record is empty. A firm cannot learn where its call recordings and prospect data are stored, whether its data sits alongside other firms' in shared infrastructure, whether any processing happens outside the United States, or whether a dedicated environment is available to a large firm. The company page describes its engineering cofounder's background building contact center systems for banks, telecom operators and insurers, the kind of buyers who ask these questions first, which makes the silence more notable rather than less. For clients operating in as many as 22 states, where the data lives is the question most likely to come from a firm's own clients, and the public record gives it no answer.

No hosting provider is named, no region or data residency commitment is published, and no deployment options are described beyond the product being cloud based and embedded on a firm website. The platform holds prospective client chat transcripts, text message history and lead records, and routes text messaging through a carrier layer that is not identified, so neither the location of the data nor the parties handling the messaging is stated.

No certification, audit report, penetration test or trust center appears anywhere on the current site, and the compliance claims made in a 2025 launch announcement are not repeated there. The company is well funded, announcing an additional $3.2 million round on its home page with Y Combinator among its backers, and sells to plaintiff firms whose intake calls routinely include medical details. No SOC 2 report or HIPAA posture is offered, and there is no published route to request one short of the intake audit call. A security review of Caseflood starts from nothing.

No certification, attestation, trust center or security page is published. The vendor material names no SOC 2 of either type, ISO 27001, named auditor, penetration testing partner or encryption statement, and refers to no gated security page. The platform holds consumer chat transcripts and text message history for law firms whose own obligations attach to that material.

No model, model provider, speech or voice provider, inference location or change notification is named. The company page says the platform is built to improve as AI models advance and that its team includes deep learning researchers, and the home page calls the dialing infrastructure proprietary, but nothing says which models write the texts, generate the in call guidance, transcribe the calls or produce the optional personalized voicemail. The website privacy notice shows the company is willing to name providers, identifying PostHog and RB2B for its marketing site; no equivalent list exists for the product, where prospects' injury and case details are what the models process. A firm cannot tell whether a lead's account of an accident is sent to an outside model provider, under what terms, or whether that changes when the models advance.

No foundation model provider, model family or version is named, no distinction is drawn between proprietary and third party models, and no subprocessor list is published. Two way business texting runs through a messaging carrier or platform and Live Call Connect routes through telephony infrastructure, and neither is identified, so the third parties that carry prospective clients' messages and calls are not named. The vendor names six integration partners and no processor.

No price, unit of charge, tier, setup fee or contract length is published. Every route in, from the header to each section of the home page, leads to a free 30 minute intake audit with the chief executive. The audit form asks for team size, monthly lead volume and monthly signed cases, which suggests pricing scales with the operation, but the site never says so. The offer is framed in outcomes instead, promising millions to tens of millions in additional attorney fee revenue. The FAQ describes the product as software plus an ongoing engagement, so the cost of the platform cannot be separated from the cost of the consulting.

Intaker does not publish a price, plan or unit of charge. Several independent software directories report a starting price of $80 per month and record that no free trial is available, and Intaker has not confirmed that figure. Nothing describes a tier structure, what the entry price includes or what raises it, or whether pricing scales by seats, conversations, leads or messaging volume, for a product that bundles chat, CRM and two way texting.

The buyer is described with precision, and the limits are implied rather than stated. Caseflood addresses firms already investing in leads and a dedicated intake team, and the FAQ opens by asking exactly that. Its tools are built for reps, intake managers and firm leaders, and its two named clients are plaintiff firms of more than 50 and more than 200 employees working across two and 22 states. The audit form lists the practices it expects to see: personal injury, workers' compensation, mass torts, medical malpractice, Social Security disability, employment, consumer law, criminal defense and immigration. Its team page cites former Morgan & Morgan intake managers and an adviser whose referral platform serves more than 700 plaintiff firms, which reinforces the plaintiff focus. Nothing says who should not buy it. The audit form's smallest options, no intake staff and under 250 leads a month, suggest smaller firms are welcome, but no minimum volume or poorly suited practice is stated.

The prompt library holds more than 1,400 custom prompts with prebuilt scripts for different practice areas, so the product ships with qualification scripts for specific matter types. No practice area is named and no jurisdiction is stated, and firm coverage is described only as law firms of all sizes.