Casepoint vs Nextpoint: how they compare in 2026
Casepoint and Nextpoint are both cloud ediscovery platforms selling into different rooms. Casepoint's pitch is government grade authorization. Nextpoint's is a flat per user price with nothing charged for data. Casepoint publishes FedRAMP High and Moderate, Department of Defense Impact Levels 4, 5 and 6 with authorities to operate from the Defense Information Systems Agency, and SOC 1, 2 and 3 alongside ISO 27001:2022. Those government authorizations are independently assessed and listed by the authorizing bodies rather than asserted. Its agreement then carries an AI clause classifying inputs and outputs as confidential information, preventing cross tenant and cross matter retrieval, restricting personnel review, and enumerating five named limitations of the AI. Nextpoint answers on commercials and on candor. It itemizes the eight things it never charges for and publishes an instruction to treat generated summaries as starting points and verify them.
At a glance
All 15 axes, side by side
The same grid applied to every vendor in the index, graded from public sources. 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.
CaseAssist Active Learning learns from reviewers' decisions to put likely relevant documents first, and the vendor ties its claim of 95 percent faster identification to it. Generative features add chat over customer data, summaries and automatic classification. Without the AI, a full eDiscovery platform remains: legal hold, collection from Microsoft 365, Teams, Slack and Google Vault, processing of more than 600 file types, review, production, FOIA work and hosting at government authorization levels. The AI is bundled at no extra cost rather than sold on its own, and the vendor leads with security and unified workflow, with AI third.
The AI is real and peripheral. The main eDiscovery product page does not mention AI, machine learning or technology assisted review. Every feature it names is deterministic or classic text processing: deduplication, email threading, near duplicate detection, OCR, hit reports, coding, Bates stamping, production and privilege log templates, and pattern based auto redaction. Privilege review is done by people, and the log is generated from their coding. The AI consists of result ranking on models Nextpoint hosts and AI Transcript Summaries on Amazon's managed foundation model service. Without it, what customers pay for remains: unlimited data hosting at a flat per user rate from discovery through trial.
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.
Generative answers use retrieval augmented generation over the customer's own data, with source citations, so a reviewer can open the document behind an answer. The agreement names five limits of its AI: training data quality, errors from vague prompts, nuance and sarcasm in language, factual inaccuracy, and lack of common sense. No accuracy figure, recall or precision statistic, or validation protocol is published, and courts examine validation statistics for technology assisted review. The terms also bar customers from sharing benchmark results without written consent.
Accuracy is treated honestly at Nextpoint but not measured. Its trust page says every AI feature is tested extensively before release, and that summaries and results are starting points to trust but verify. No accuracy figure, test set, recall or precision statistic, error rate or evaluation is published behind the testing claim. Nothing says whether a transcript summary links back to the passages it rests on, which a reader would need to verify it, though the transcript itself is already in the customer's hands.
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.
Casepoint states that its technology should amplify human judgment, not replace it, and the agreement commits to human in the loop AI features. Every generative feature is optional, and the customer controls adoption. Reviewers get source citations on generated answers, and transparency logs are said to record every AI decision. Active learning is supervised by design, since reviewers' own calls train it. The agreement also tells customers not to rely solely on output for anything with material consequences. No threshold at which a feature declines to answer, no behavior when the data does not support a result, and no route for reporting a wrong output is described.
A person should review AI output, Nextpoint says, without saying where that happens. Its position is that the AI speeds review rather than replacing judgment, and that generated summaries are starting points to trust but verify. That is a caution, not a control. Nothing on the eDiscovery platform page, the trust and security page or the pricing page, or in the terms of use, describes what happens when a transcript does not support a summary. No confidence indicator, declining threshold, correction route or in product review step is described. The platform does have tamper resistant per document activity logs and six permission levels, but nothing published ties them to the AI.
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.
Marriott International, Mayo Clinic, the West Virginia Office of Technology, Kramer Levin, Beveridge & Diamond, Larkin Hoffman and Lewis Roca are named as customers, with named people and titles. Lewis Roca, an Am Law 200 firm, is tied to a measured result: more than 90 percent less document review time from the analytics and AI. Other case studies report a 57 percent cut in discovery related costs at a Fortune 500 construction company, faster FOIA responses at a major defense agency, and a migration of up to a petabyte at a federal regulator. Casepoint states more than 250 government agencies and about 80 percent of federal agencies as customers. Most measured results name a category rather than a customer.
More than 600 law firms and legal teams are claimed as customers, and many are named, including McDermott, Foley & Mansfield, Jackson Walker, Seyfarth Shaw, Simmons Hanly Conroy, Clyde & Co, Williams Kastner, Yates Construction and Augusta University. Eight testimonials carry a name, job title and firm. One gives figures: a litigation support services manager at Foley & Mansfield describes migrating 180 databases, running close to 250, and much shorter training time than on the firm's previous platform. What is missing is a measured outcome. The claims of up to 50 percent lower cost and review time cut in half are not tied to any customer, and no testimonial is dated. A case studies library is also published.
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.
Confidential information, as the agreement defines it, includes attorney work product, attorney client privilege, personal data and health data. A separate AI clause makes AI inputs and outputs customer data and confidential information, limits their use to running the service, and commits AI features to enforce the customer's access and matter permissions. It also prevents retrieval across tenants or matters and logs administrative access. Casepoint staff may review AI inputs or outputs only for customer requested support or a security incident, with time limited, audited access. Remote support needs the customer's own enablement and multi factor authentication. Another clause gives Casepoint a transferable license over customer data and usage data to improve its products, and the AI clause carves out only AI inputs and outputs.
Nextpoint's commitments on confidentiality and training sit outside its contract. Its trust page says AI runs entirely inside its Amazon environment, case data never trains any model, and inputs and outputs are never shared with or kept by an outside model provider. The terms of use describe the service as a platform for privileged data and work product, confirm the customer owns its content, and limit Nextpoint's use to providing the service, with no license to improve products. Separation is documented: six permission levels down to the document, mandatory multi factor authentication, single sign on, limited staff access with approvals, and confidentiality agreements. What is missing: the training commitment is a website FAQ, not a contract term, the terms have no AI clause, no retention period is stated for prompts or output, and the master services agreement is not published.
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.
Customers are told not to rely solely on any output, including AI output, for anything with material consequences or that affects anyone's rights, under Casepoint's agreement. It makes the customer responsible for the completeness, accuracy and content of the results it relies on. Human in the loop is a stated commitment, and the vendor says technology should amplify human judgment. No bar or ethics authority is named, including ABA Formal Opinion 512, no jurisdiction limits are stated, and nothing addresses the duties of the reviewers who act on the output. The product has no consumer facing surface.
Nothing published addresses the advice line. The eDiscovery platform page, the pricing page, the trust and security page and the terms of use (updated June 2026) do not say that output is not legal advice, whether an attorney client relationship arises, or who is responsible for a summary a lawyer relies on. No bar authority, including ABA Formal Opinion 512, and no jurisdiction limit is named. The terms deny that information from the service is accurate or reliable, which allocates risk rather than drawing the line. Nextpoint does say its AI speeds review rather than replacing judgment and that output should be verified, a supervision statement without a disclaimer. The platform advises on nothing, though transcript summaries are work a lawyer acts on.
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.
TRUST, the AI framework, has five principles, and Casepoint's terms turn several of them into commitments. The terms limit how AI inputs and outputs may be used, restrict staff review of them, require AI features to respect the customer's access and matter permissions, and require logging of administrative access. Transparency logs are said to record every AI decision. Customers above a stated spend get an annual audit right over privacy and security controls. No fairness testing, testing regime or accountable owner is published, and nothing addresses whether review ranking can be uneven across custodians, languages or document types.
An AI governance position comes with a real mechanism at Nextpoint, and bias goes unmentioned. Its trust page answers five AI questions: whether case data trains models, where documents are processed, how this differs from a consumer chatbot, whether output can be trusted, and which compliance framework covers the AI. The answer to the last is that AI features sit inside Nextpoint's SOC 2 Type II scope and its annual IT risk assessment, so model behavior is inside an audited scope. An ungated document, Security, Compliance, Privacy and AI at Nextpoint, sets out its AI principles and safeguards. What is missing: no accountable owner is named, the stated testing before release has no published result, and nothing addresses bias or uneven output across custodians, document types or languages.
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.
Casepoint commits to written notice within seventy two hours of a validated security incident, with updates on cause, mitigation, impact and remediation. On termination it exports all customer data in a common format within a seven day window, and states that copies may persist on backup media. Access controls include role based access, zero trust networking, encryption in transit and at rest, and no standing remote support access. Staff face ten year criminal background checks and OFAC screening. Disaster recovery plans are tested every year. The agreement promises a subprocessor list, and none is published.
Nextpoint publishes detailed, current practice across most of the ground. Access control is strongest: six permission levels down to the document, mandatory multi factor authentication, single sign on through Okta or Azure AD, IP restricted AWS access with lockout, and staff access limited to a subset of employees with approvals and prompt revocation. Customers can export at any time, deletion follows a Data Archive Form, and retired media are wiped to DoD 5220.22-M or NIST 800-88 or destroyed. The incident process covers daily log review, notice to affected users as soon as possible without harming the investigation, cooperation duties, and the customer's sole right to decide onward breach notice. A public status page reports incidents. Staff get background checks and annual training, and storage is redundant with eleven nines of stated durability. What is missing: no subprocessor list beyond Amazon, no retention period during the term, and no breach notice deadline.
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.
The general cap is the fees paid under the relevant order in the six months before a claim. A higher cap of five times annual fees, plus direct damages, applies to confidentiality breaches, data breaches, gross negligence and willful misconduct. Personal injury, fraud and anything that cannot be excluded by law are carved out. Casepoint defends and indemnifies against claims that the software infringes a copyright in the United States or misappropriates a trade secret, with named remedies. The software is warranted to perform substantially as documented for sixty days, services are warranted as workmanlike, and uptime is committed at 99.5 percent with service credits. No warranty attaches to AI output, the agreement puts accuracy on the customer, and no insurance position is published.
Liability is handled by a standard limitation package, and nothing runs toward the customer. The terms of use (updated June 2026) cap liability at fees paid in the twelve months before the event, exclude indirect and consequential damages including lost data, and set a one year limit to bring claims, under Illinois law with venue in Chicago. The service is provided as is, with no promise that information from it, a generated summary included, is accurate or reliable. The customer releases Nextpoint, with a California Civil Code 1542 waiver, and indemnifies it; there is no vendor indemnity. What is missing: no carve outs from the cap, no warranty on output, no insurance position and no AI term. The Subscription Agreement and Master Services Agreement the terms refer to are not published.
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.
Microsoft 365, Microsoft Teams, Slack and Google Vault are collected in place without manual export, through connections into the systems where data lives, with more than 600 file types in processing. An API automates collection from those platforms. The MCP Server connects a customer's own approved AI systems to the platform. No iManage or NetDocuments integration is named, nothing covers matter management, ebilling or court filing, and no public API documentation is published. The terms restrict use of APIs Casepoint does not support.
Working connections are named with what they do. Evidence can be pulled from OneDrive, Dropbox and Google Drive without manual export. Single sign on runs through Okta or Azure Active Directory, with custom SSO through Client Success, and multi factor authentication is mandatory. No connector to iManage or NetDocuments, matter management, electronic billing, court filing, Word or Outlook is listed. An API exists, mentioned only as an invitation to discuss custom integrations, with no public documentation.
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.
Deployment comes three ways, hosted by Casepoint with 99.5 percent uptime, installed by the customer or a third party under named conditions, and in separately certified government environments. SOC 1, SOC 2, SOC 3, ISO 9001, ISO 27001 and NIST apply to the commercial offering, and GovRAMP, FedRAMP Moderate and High and Department of Defense Impact Levels 4 to 6 to certain environments. Residency is committed for export controlled data, which stays in environments Casepoint controls in the United States, with no replication abroad and access only by American persons. No region list for the commercial offering, tenancy model, or split between where data is stored and where it is processed is published. A Canada government offering appears in the navigation with no residency statement.
Infrastructure at Nextpoint is described in detail, with residency left unaddressed. It runs on Amazon Web Services, with redundant storage across facilities, eleven nines of stated durability, checksums, survival of two facility losses, Virtual Private Cloud isolation, CloudWatch monitoring and IP restricted staff access. Its own office holds client data in a keyed office in a keyed building, with server access by password over a VPN. What is missing: no region, residency option or tenancy model is stated, and nothing separates where data is stored from where AI runs, beyond AI running in the same Amazon environment. GovCloud appears once, about Amazon's physical controls, not as an option. The Advanced plan's custom account headquarters feature is not defined.
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.
Casepoint holds fourteen certifications and authorizations, each explained on its security page. They are FedRAMP High and Moderate, Department of Defense Impact Levels 4, 5 and 6 with DISA authorities to operate, GovRAMP, SOC 1 Type II, SOC 2 Type II, SOC 3, the SIG questionnaire, ISO 9001:2015, ISO 27001:2022, NIST 800-53, NIST 800-171 and the Data Privacy Framework between the EU and the United States. FedRAMP and DISA authorizations are assessed by third parties and listed by the authorizing bodies, so they can be confirmed without a sales call. The SOC 2 Type II is described as annual, covering all five Trust Services Criteria over twelve months. The agreement restates the set and says which apply to the commercial offering and which to certain environments. No auditor is named for the SOC or ISO work, and no report date is shown. Reports are available to customers on request.
SOC 2 Type II, described as an annual third party audit covering security, availability, confidentiality and privacy, is held by Nextpoint. SOC 3 reports are available on request, and SOC 2 reports after a signed nondisclosure agreement. It separates its own certification from Amazon's SOC 1, SOC 2, SOC 3, ISO 27001 and FedRAMP for the underlying cloud, without claiming those as its own. A public status page, an annual IT risk assessment and an ungated overview document support it. What is missing: no auditor, report date or coverage period is given, and the SOC 2 report needs a signed agreement.
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.
Generative answers use retrieval augmented generation over the customer's own data. The agreement names machine learning, technology assisted review and large language models as the technologies in use. It commits Casepoint to license third party software properly, patch it, and keep a list of material third party components and subprocessors; that list is not published. No model, provider, version or inference location is named, and no notice is promised when a model changes. The separate MCP Server lets a customer connect its own approved AI systems.
Generative AI runs on Amazon Bedrock, Amazon's managed foundation model service, which Nextpoint names. Ranking runs on models Nextpoint hosts in its own private cloud, and transcript summaries run on Bedrock in the same environment. Naming the service means its terms can be read. Bedrock does not train on customer data, does not store or log prompts or completions, and does not share inputs or outputs with model providers. What is missing: which Bedrock model writes a summary, any model version, and a promise to notify customers when the model changes, which Bedrock makes easy.
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.
The pricing page is titled Transparent Pricing and gives no price. It does publish the shape of the deal: a modular platform priced to the applications a customer takes, with no charge for products it does not use. The purchasable modules are named across the site, from eDiscovery and Legal Hold to FOIA, Casepoint AI and the MCP Server. The agreement adds billing mechanics. Fees are the greater of actual usage or the agreed cap, metered on data storage, processing and users. Terms run one year and renew automatically unless either side gives sixty days' notice. No rate, per gigabyte or per user figure, or starting price is published. The only route is a demo request.
Nextpoint publishes the full shape of its pricing and no number. Three plans are named with their features. Essential covers smaller review, production and deposition work, Advanced adds analytics, auto redaction for each document, production templates, advanced permissions and activity reporting, and Apex is a custom enterprise plan. Each user costs a flat monthly rate. The page lists what is never charged: storage, processing, hosting, OCR, deduplication, productions, exports and sharing, compared with per gigabyte fees elsewhere. Annual prepayment discounts and volume pricing are mentioned. No figure appears at any tier, and every plan routes to a call. One discrepancy: the marketing says unlimited data, while the terms of use say plans may be user based or data based, with hosting and processing limits set in an unpublished master services agreement.
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.
By market Casepoint serves federal civilian, federal defense, state, local and education, Canada, and corporate enterprise. By team it serves corporate legal, government legal, FOIA and public records, audit and investigations, and IT and security, with a separate law firm page. Use cases include legal hold, investigations, litigation, eDiscovery, data subject access requests, third party subpoena responses, FOIA requests and congressional inquiries. It processes more than 600 file types from named sources. No excluded practice area, unsupported data source or volume limit is named on the platform, pricing or security pages or in the site navigation, and the Canada offering is not described.
Coverage at Nextpoint is documented three ways, with no stated limits. By buyer: law firms, corporations, government agencies and education, each with a page. By specialty: construction litigation, insurance defense, mass tort and IP litigation. By use: data collection, internal investigations, FOIA requests and transcript management. File types are listed, including PST, MSG, EML and MBOX email, Office files, images, audio, video, archives, source code, CAD files and building information models, with text extraction and OCR. It states it handles millions of documents and multi party MDL matters, reusing processed data across hundreds of related cases. What is missing: no excluded practice, size limit or unsupported jurisdiction is named; the only limit stated is that specialized formats may need specific workflows.
The 12 legal signals, side by side
Recorded rather than graded. These are the questions a practitioner has to answer before a tool touches a client matter, and the answers are taken from public material only.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
Casepoint will use AI inputs and outputs only to provide, secure and support the service, and not for product improvement, unrelated analytics or AI training without the customer's approval. AI inputs include prompts, questions, customer data and metadata sent to AI features. Two other statements bear on it. The AI strategy page goes further than the contract and says customer data is never used to train models. And another clause gives Casepoint a transferable license over customer data and usage data to improve its products, with only AI inputs and outputs carved out.
Nextpoint's trust page says its AI does not use case data to train models: features run inside its Amazon environment, are never used to train or improve any AI model, and inputs and outputs are never shared with or kept by an outside model provider. It contrasts this with consumer tools that may train on inputs. The commitment is not in the terms of use (updated June 2026), which have no AI clause, so it lives on a website page Nextpoint can change.
The terms support it indirectly, since Nextpoint gets no license to improve its products and may use content only to provide the service. The master services agreement, where a binding commitment would sit, is not published.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
The agreement treats AI inputs and outputs as customer data and confidential information, so prompts and results belong to the customer like anything else it uploads. On termination Casepoint exports all customer data in a common format, may delete its copies after delivery or seven days from notice, and states that backup copies may persist. No retention period during the term, separate lifecycle for prompts and outputs, or zero retention setting for AI use is published.
Retention is answered for the model provider and left open for Nextpoint. The trust page commits that inputs and outputs are never shared with or kept by an outside model provider. It does not say how long Nextpoint keeps a prompt or summary, and offers no retention setting or zero retention option. Customers can export at any time, and deletion follows a Data Archive Form that archives, deletes or exports the database. Activity logs are kept permanently unless the database is archived or deleted.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
The agreement commits that AI features will enforce the customer's access controls and matter and workspace permissions, prevent retrieval across tenants or matters, and log administrative access to AI inputs and outputs. Casepoint is itself the system of record, so the AI works inside the platform's own permission model. Support access follows the same rules: no standing access, enablement only by the customer's administrators, access limited to the relevant matter, named staff with multi factor authentication, and audit logs for the customer.
Neither how the boundary is enforced technically nor how the commercial and government environments are kept apart is described.
A permission model works at matter level. Six permission levels, from view only to dashboard administrator, control access by case, folder and document for internal teams and outside parties. Multi factor authentication is mandatory, single sign on runs through Okta or Azure Active Directory, staff access is limited and approved, and every document's activity is logged in tamper resistant records. Nothing says the AI features respect those permissions, or whether a summary or ranked result could cross a wall.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
The confidentiality clause allows disclosure under law, court order, subpoena or other compulsory process on three conditions. The receiving party must notify the other immediately and share the relevant documents on request. It must resist and limit the disclosure, including by lawful objection, protective orders and redaction. And it must cooperate with the other side's efforts to protect the information. A matching clause requires a government customer under FOIA or an open records law to give Casepoint notice and a chance to object before releasing its confidential information. No transparency report or disclosure statistics are published.
Nothing published addresses third party demands for customer data. The terms of use (updated June 2026), the trust and security page, the eDiscovery platform page and the pricing page say nothing about subpoenas, court orders, warrants or government requests: no notice to the customer, no protective order, no plan if notice is prohibited, and no transparency report. On a security breach, the customer alone decides whether and how to notify others, so the notice question is handled in that setting only.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Documents the customer collects from its own systems are the material, so there is no vendor assembled body of law, content license or data supplier to name. The agreement states that the customer owns all customer data, with Casepoint holding only a limited license to provide the service. Collection sources are named, including Microsoft 365, Teams, Slack and Google Vault.
The question does not arise for this product. Nextpoint works on evidence customers collect and upload from their own systems, so there is no vendor assembled body of law, content license or data supplier. The terms of use say the customer, or the party that entrusted the data, owns all user content, and Nextpoint's access is limited to providing the service. No training corpus is described either, since generative features run on Amazon's managed foundation model service.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
Primary law sits outside the platform, which works on the customer's own collected documents, so there is no cited authority to check and no citator is claimed. The related accuracy question for this product is validation of technology assisted review, and no recall, precision or elusion figure is published.
No citator applies, and none is claimed. The platform works on a customer's own documents and deposition transcripts rather than primary law. The accuracy question that does apply, whether a summary fairly represents the transcript, is unmeasured.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Nothing on the AI strategy, platform or security pages or in the full terms describes what the generative features do when the data does not support an answer. There is no abstention path, no confidence or grounding indicator for the user, and no statement of what a chat query shows when it finds nothing. The agreement tells customers that output has limits and should not be relied on alone, which allocates risk rather than describing behavior. Active learning ranks documents by predicted relevance, and nothing says how uncertainty in that ranking is shown.
Users are told by Nextpoint to check output, with nothing on what the system does when unsure. It says AI features are tested before release and that summaries are starting points to trust but verify. No abstention path, behavior when a transcript does not support a summary, confidence indicator or declining threshold is described on the trust and security page, the eDiscovery platform page or the pricing page, or in the terms of use.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
The AI Hallucination Cases database maintained by Damien Charlotin records no case naming Casepoint, and no court order, opinion or disciplinary record naming the product appears in several independent 2026 sanctions trackers or trade coverage. Casepoint's generative features answer questions about collected documents rather than citing law, so a fabricated citation would not usually start here.
The AI Hallucination Cases database maintained by Damien Charlotin, several independent 2026 sanctions trackers and trade coverage record no court order, opinion or disciplinary record naming Nextpoint. Nextpoint's generative output summarizes a transcript the customer already holds rather than citing law, so the nearer risk is a summary that misstates testimony.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Casepoint's marketing promises defensible results, processes and chain of custody, and the agreement supports that with audit logging. No rule, standard or opinion is cited on the platform, AI strategy, pricing or security pages or in the full terms: not the Federal Rules of Civil Procedure, Rule 502(d), Sedona Conference commentary, ABA Formal Opinion 512 or any state bar guidance.
Passing software costs through to clients with the invoice generator is fully compliant with ethical guidelines, according to Nextpoint's pricing page, which points to a Billing for eDiscovery eGuide, an ungated PDF. No rule, opinion or bar authority is cited, so a firm cannot check the claim. Nothing on the eDiscovery platform page, the pricing page or the trust and security page, or in the terms of use, names ABA Formal Opinion 512, state bar guidance, the Federal Rules of Civil Procedure or Sedona Conference commentary.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Casepoint publishes cost and speed claims: a 57 percent cut in discovery related costs, relevant data found up to 95 percent faster, time to insight cut by up to 83 percent, FOIA processing up to 75 percent faster, and lower outside counsel spend. The pricing, platform and AI strategy pages and the full terms describe no record of AI assisted work by matter for fee purposes and no guidance on how AI assisted review should show on a client's bill.
One named law firm customer says it looks forward to guidance on billing and cost recapture, which the vendor has not published.
Nextpoint publishes tools for allocating software cost to clients, and none of them address AI assisted work. A custom invoice generator lets a firm allocate software costs to clients, calculates a reasonable monthly amount per project and produces a client invoice. Nextpoint's pricing page says passing these costs to clients is fully compliant with ethical guidelines, without citing an authority, and offers a Billing for eDiscovery eGuide.
None of it answers what happens to the bill when work that took six hours takes one. No per matter record of AI assisted work is described, and no guidance on fee or client disclosure for generated summaries is published.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
The agreement commits Casepoint to provide third party audits and its SIG questionnaire on request, to keep a list of material third party components and subprocessors, and to give customers above a stated spend a yearly audit right. A data processing agreement is referenced beside the security page. The fourteen certifications appear on the public security page, and the FedRAMP and Department of Defense authorizations can be confirmed with the authorizing bodies. No subprocessor list is published, despite the commitment to keep one, and no model provider is named.
Nextpoint publishes a pack without a form. The trust page offers Security, Compliance, Privacy and AI at Nextpoint as a direct download, alongside a companion security overview. Together they cover role based access with six named permission levels, the privacy policy, and compliance, including a plain statement that Nextpoint complies with the CCPA for California residents and is not currently GDPR compliant. The public pages answer the usual AI clause questions: case data never trains any model, processing stays in Nextpoint's Amazon environment, and inputs and outputs are never shared with or kept by an outside model provider. No subprocessor list exists beyond Amazon.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Transparency logs are said to document every AI decision, generated answers cite sources in the customer's data, administrative access to AI inputs and outputs is logged, and support access logs are made available to the customer. Chain of custody from preservation to production is the platform's central design. No model is named or versioned, so a filing could not say which system produced an output. Nothing says whether the transparency log can be exported for a standing order or certification, and no template exists for disclosing AI use to a court.
A strong activity record exists, and nothing ties it to the AI. Every document's view, edit and markup history is logged with timestamps in tamper resistant logs, kept permanently unless the database is archived or deleted, which Nextpoint presents as a defensible record. What is missing for AI: no model is named or versioned, so a filing could not say which system wrote a summary. Nothing says whether the log captures AI use or who verified output, and no export, template or guidance for disclosing AI use to a court exists.
The questions both sides leave open
Derived from the records above rather than written, so it cannot favor either vendor. Take these into both conversations and ask each side the same question.
- Primary Law Corpus Provenance
- Good Law Verification
- Refusal and Uncertainty Behavior
- Bar Guidance Alignment
Which one fits
Choose Casepoint if
- Your matters reach into government. Casepoint publishes FedRAMP High and Moderate, Department of Defense Impact Levels 4, 5 and 6 with authorities to operate from the Defense Information Systems Agency, GovRAMP, SOC 1, 2 and 3, ISO 27001:2022 and ISO 9001:2015, and the government authorizations are independently assessed and listed by the authorizing bodies, so a buyer can confirm the status without a sales conversation.
- You want the AI governed by the contract rather than a page. Casepoint's agreement classifies AI inputs and outputs as both customer data and confidential information, limits their use to providing, securing and supporting the service, commits that AI features enforce the customer's own matter and workspace permissions and prevent cross tenant and cross matter retrieval, restricts personnel review to customer requested support or a security incident under least privilege time bounded access, and enumerates five named limitations of the AI itself.
- Collection is where chains of custody break. Casepoint collects in place directly from Microsoft 365, Teams, Slack and Google Vault without manual export, processes more than 600 file types, and runs hold, collection, review, production and FOIA workflows on one system of record, with Lewis Roca named alongside a stated reduction of more than 90 percent in document review time.
Choose Nextpoint if
- The bill should not scale with the data. Nextpoint charges a flat monthly rate per user and itemizes what is never charged: data storage, processing, hosting, OCR, deduplication, productions, exports and sharing, set against the per gigabyte hosting norm, across three named plans whose feature splits are published even though no figure is.
- You want to know which service the generative feature runs on. Nextpoint states that search and ranking run on models it hosts in its own private cloud and that AI transcript summaries run on Amazon Bedrock inside the same environment, with three specific protections published against it: no training on customer data, no storage or logging of prompts and completions, and no sharing of inputs or outputs with model providers.
- The matter should not change systems at trial. Nextpoint carries the same evidence from processing and review into deposition transcript and video review, designations, chronologies, visual timelines and a presentation mode, with six permission levels reaching individual documents, multi factor authentication mandatory on every user and device, and single sign on through Okta or Azure Active Directory.
In summary
Casepoint
Casepoint is a unified data discovery platform on a single system of record. It covers legal hold, in place collection from cloud sources, processing, AI assisted analysis, review, production, investigations and public records response, with CaseAssist included rather than sold as an add on. According to the AI Legal Index, Casepoint publishes FedRAMP High and Department of Defense Impact Levels 4, 5 and 6, and names Lewis Roca alongside a stated reduction of more than 90 percent in document review time. Its customer agreement carries an AI clause that prevents cross matter retrieval and restricts personnel review. No accuracy statistic, model provider or price is published.
Nextpoint
Nextpoint is a cloud native litigation platform covering discovery through trial. It is sold on a flat per user subscription with nothing charged for storage, processing, OCR, deduplication, productions or sharing. The same evidence carries from processing and review into deposition designations, chronologies, timelines and courtroom presentation. According to the AI Legal Index, machine learning is a narrow layer rather than the pitch. Search and ranking run on models Nextpoint hosts itself, and generative transcript summaries run on Amazon Bedrock inside the same environment, with three published protections on what that service may do. No accuracy figure, residency statement or price is published.
Questions buyers ask
Casepoint vs Nextpoint: which is better for ediscovery?
They sell into different rooms. Casepoint publishes government grade authorizations that can be verified independently and an AI clause inside its customer agreement. Nextpoint publishes a commercial model instead, charging a flat rate per user with nothing charged for data, and is frank that generated output is a starting point to be verified. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What does each cost?
Neither publishes a figure. Nextpoint publishes the whole shape, naming three plans, stating the unit as a flat monthly rate per user, and itemizing the eight things it never charges for against the per gigabyte industry norm. Casepoint publishes a page titled transparent pricing that carries no price, describing tailored pricing across named modules, with the billing mechanics set out in its agreement including usage based fees calculated on storage, processing and user counts. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
Who runs the AI?
Nextpoint names the service: search and ranking on models it hosts itself, and generative transcript summaries on Amazon Bedrock within the same environment, though which foundation model inside that service produces a summary is not stated. Casepoint names the architecture but not the provider, describing retrieval augmented generation over the customer's own data with source citations and transparency logs recording each AI decision, with no model, provider or version identified. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What happens to privileged material?
Casepoint addresses it expressly, defining confidential information to include attorney work product and attorney client privileged material, and committing in the same agreement that AI features enforce matter and workspace permissions and prevent cross matter retrieval. Nextpoint describes its platform as holding privileged data and work product, confirms the customer owns it, limits its own use to providing the service, and offers six permission levels down to the document, without naming privilege in a contractual commitment. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What do Casepoint and Nextpoint both leave unpublished?
Neither publishes an accuracy or validation statistic for its AI, which in this category means neither publishes the recall and precision figures a court would ask about. Neither names the auditor or the report date behind its attestations. Neither names an integration with a document management system such as iManage or NetDocuments. And neither publishes anything on uneven output across custodians, document types or languages, which is the live question for a system that decides what a human reviews first. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
Neither vendor paid for inclusion, placement or a grade, and neither reviewed this page before it published. Everything on it comes from public material on the dates shown. How the index grades.
One clause in Casepoint's terms bears directly on what a buyer can learn. It forbids the customer from disclosing software performance benchmark results to any third party without written consent, which forecloses independent testing. No accuracy, recall or validation statistic is published anywhere. That gap is conspicuous in ediscovery specifically, where technology assisted review validation figures are what courts examine. On Nextpoint, the marketing describes unlimited data throughout. Its terms of use state that the service is offered under user based or data based plan types, and that each tier determines the data hosting and processing limits, as detailed in a master services agreement that is not published. A reader of the pricing page alone would not know that data limits exist. Neither vendor reviewed this page.