Logikcull

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.

MEMBERSHIP: the AI bar is cleared, and it was worth checking rather than assuming. For most of its life this product's value proposition was workflow automation rather than models. The owner announced in March 2026 that it was bringing ASK generative AI to the platform alongside European hosting and new pricing, so a shipped generative feature now exists rather than a roadmap. GRADE: the models are the engine of a core capability layered on a platform that would function without them. Culling Intelligence automatically categorises ingested data into predefined filters including document type and personally identifiable information, and automated PII detection, similar document detection, email threading and audio and video transcription are all model driven and central to what makes self service viable. But upload, processing, review, redaction, legal hold and production are workflow automation that works without any model, and that workflow is what the product sold on for two decades before the generative layer arrived.

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.

Accuracy is asserted without vendor published measurement and the newest capability is undescribed. The generative feature, ASK, was announced in March 2026 and searched on the vendor site, the owner's news releases and third party review material on 29 Aug 2026 without locating any description of how it grounds an answer, whether output cites the documents it rests on, or any accuracy figure. On the machine learning side a PII redaction accuracy figure above 99 percent circulates in third party review material but was not located on vendor property, so it is recorded here as unverified rather than relied on. No test set, no evaluation methodology, no independent benchmark participation and no hallucination disclosure was located. This axis will be worth rebuilding once the generative feature has settled and the vendor documents it.

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.

Autonomy is present and the control structure is implied by workflow rather than published. The product's design places the human at the decisive points: legal teams upload their own data, apply their own review tags, run their own redactions and generate their own productions, with Culling Intelligence sorting and categorising rather than deciding. That is a genuine allocation of judgement and it is visible in how the product is described. But it is a description of a workflow rather than a published oversight position. Searched the vendor site, the owner's news releases and third party material on 29 Aug 2026 and located no statement of what ASK or Culling Intelligence do unaided, no threshold at which anything escalates, no review surface described as such, and no statement of what the vendor commits to when an output is wrong. The absence matters most for the newly added generative feature, where no oversight material exists at all.

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.

Independent review volume is substantial and vendor deployment evidence is not. Third party review platforms carry a large verified body of user feedback, including 269 verified reviews at 4.6 on one platform and further verified sets elsewhere, which is real evidence about adoption and satisfaction from an independent source and is unusual in this index for its volume. Quantified outcome claims circulate in third party review material, including discovery cost reduction of up to 70 percent and handling three times the matters without new headcount, but were not located on vendor property and are recorded as unverified. Searched the vendor site, the owner's news releases and the review platforms on 29 Aug 2026 and located no named customer paired with figures and a date, and no case study with an assessable method.

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.

CORRECTED 29 Aug 2026 during the trust portal sweep. Previously graded C on inherited certification with product scope unconfirmed and no product level controls located. The product's own security page has since been reached and the position is materially stronger. Confidentiality controls are published and specific: AES-256 encryption at rest with keys held in AWS Key Management Service and rotated regularly, TLS 1.2 or higher in transit with continuous monitoring, granular role based permissions enforced with single sign on and two factor authentication, real time malware scanning of uploads, stated egress controls preventing unauthorised data leaving the application, and data residency in United States and European data centres described as SOC 2, HIPAA and ISO 27001 compliant. Certification scope covering this product is now confirmed rather than assumed. Two gaps hold this off an A and both are the limbs this axis turns on. No statement was located on whether customer content may be used to train models, which matters because a generative feature was added to this product in March 2026 with no accompanying data handling statement. And no retention or deletion terms were located, which matters for a product whose users upload entire custodian collections. Attorney client privilege and work product are not addressed directly either.

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.

The audience is unusually broad for this index and no position is published. Stated users span corporate legal departments, state and local government agencies, law firms, educational institutions and non profits, and the product is explicitly designed to be usable by paralegals and IT staff as well as lawyers, with the self service proposition resting on people who are not litigation support specialists running discovery themselves. That is a wider non lawyer surface than any other ediscovery record here. The product processes and produces documents rather than giving advice, so the advice line question arises less sharply. Searched the vendor site, the owner's news releases and third party material on 29 Aug 2026 and located no published position on advice versus tooling, no treatment of competence or supervision duties, and no jurisdiction limits despite FOIA, public records and data subject access request workflows all being jurisdiction specific.

AI Governance and Bias Disclosure

Published governance over model behaviour: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.

Searched the vendor site, the owner's news releases and announcements, and third party material on 29 Aug 2026. No governance position for model behaviour was located at product level: no AI principles or framework specific to this platform, no named owner of model governance, no pre release testing regime, no AI management certification, and nothing on uneven output across matter types, parties or populations. The owner publishes a three pillar AI pledge, which was recorded on its own record and is not restated as a product level framework here. The absence has a specific edge for this product: automated PII detection and redaction determines what gets withheld from a production, and differential performance across name formats, languages or document types would be invisible to a self service user with no litigation support specialist to catch it.

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.

CORRECTED 29 Aug 2026 during the trust portal sweep. Previously graded C on general inherited assurance with little product specific detail. That was wrong: this product publishes its own security page, which was not reached in the original pass, and it is specific. Published there: deployment on Amazon Web Services with multiple availability zones and daily backups; encryption in transit using TLS 1.2 or higher with continuous monitoring; encryption at rest using AES-256 with keys managed and stored in AWS Key Management Service and rotated regularly, which is key management described at the mechanism level rather than asserted; all HTTPS traffic protected against distributed denial of service attacks, zero day exploits and intrusion attempts, with egress controls stated to prevent unauthorised data leaving the application; real time virus and malware scanning of every uploaded file; granular role based permissions with single sign on and two factor authentication; and 24/7 system health monitoring on a public status page. Data centres are stated to be SOC 2, HIPAA and ISO 27001 compliant and located in the United States and Europe. A trust centre with a documentation request route is published. Not located as of 29 Aug 2026, and the reason this is B rather than A: no stated retention period or deletion control for uploaded data, generative prompts or outputs, no named subprocessor list, and no incident or breach notification practice.

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.

Searched the vendor site, the owner's news releases and third party material on 29 Aug 2026. No published indemnity, liability cap, carve out, warranty on output or insurance position was located, and no customer terms of service was located on the surfaces reached. Recorded as a pure absence. The shape is worth naming for a self service product specifically: the buyer here is often a small legal team or a paralegal running discovery without litigation support expertise, which is precisely the buyer least equipped to negotiate terms or to absorb the consequences of an automated redaction that misses personally identifiable information in a production. Rebuttable with one link.

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.

Real integrations exist, are named individually, and are the ones this product's proposition depends on. Named direct collection connectors: Google Vault, Slack, Microsoft 365 and Box, with the vendor stating these let a user pull data straight from the source rather than waiting for a vendor to collect it. Onna was added in March 2026 specifically for collaboration platform data from Slack and Teams, which addresses the fastest growing evidence source in modern matters. Legal hold integrates with those same systems for preservation and acknowledgment tracking. For a self service product the collection integration is the integration that matters, because it is what removes the service bureau from the workflow. Not located as of 29 Aug 2026: an integrations index page, per integration documentation of what moves and what an administrator configures, and any document management or practice management connector.

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.

Cloud delivery with a real and recently added residency option. European cloud hosting was announced in March 2026 as part of a stated investment in the platform, alongside the generative feature and new pricing, and the vendor frames it as expanded global infrastructure for customers worldwide. That is a genuine residency choice rather than a geography footnote, and it is the second ediscovery record here to offer one. Not located as of 29 Aug 2026: the named cloud provider, an enumerated region list, the tenancy model, any statement of where processing happens as distinct from storage, and whether the generative feature processes in the same region as the data it reads. That last question is live for any customer choosing European hosting specifically.

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.

CORRECTED 29 Aug 2026 during the trust portal sweep. Previously graded C on the basis that certification existed only at owner level and this product's inclusion in scope was unconfirmed. That is now resolved in the product's favour. The vendor states on its own why we exist page that its offerings include both platforms and that it remains committed to the confidentiality and integrity of customer data with ISO 27001 and SOC 2 Type 2 certifications, and repeats the same certification statement on this product's own trial page. This product additionally publishes its own security page, which states that customer files reside in SOC 2, HIPAA and ISO 27001 compliant data centres in the United States and Europe. A third party marketplace listing independently records ISO 27001, SOC 2 Type 2, EU US Data Privacy Framework compliance and GDPR support for this product specifically. A trust centre is published with a stated route to request security documentation, and a public status page runs at status.logikcull.com. Product scope is therefore confirmed and the earlier open question is closed. Short of an A because no coverage period, audit scope, report date or auditing firm was located, and because the certification statements are marketing page assertions rather than certificates on the trust centre reached in this pass.

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.

The model layer is identified by lineage without any component being named. The owner announced it was bringing its own AI capabilities to this platform, describing the investment as delivering the industry's most advanced AI to the self service product, which tells a buyer the generative feature draws on the owner's engine rather than an unrelated stack, and the owner's own models trace to acquired data science capability. Searched the vendor site, the owner's news releases and third party material on 29 Aug 2026 and located no named large language model or provider behind ASK, no statement of where generative processing runs, no subprocessor list, and no commitment to notify customers when the supply chain changes. Note the interaction with residency: European hosting was announced in the same release as the generative feature, and nothing published states whether the two operate in the same jurisdiction.

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 best commercial position in the ediscovery category and one of the better ones on the index, because self service and pricing are incompatible. Published by the vendor and owner: a self service model requiring no service bureau, a free trial as a genuine entry point, no per seat fee so a whole firm works on one subscription, and new pricing options announced in March 2026. The pricing model itself is the differentiator the product was built on, replacing per gigabyte vendor processing charges with a subscription, and independent verified user reviews consistently characterise the pricing as predictable, transparent and pay as you go, with free trials useful for evaluation. Rates circulating in third party review material start around $250 a month scaling with data volume, but those figures were not located on vendor property in this pass and are recorded as unverified rather than relied on. Short of an A because the rate itself was not confirmed on the vendor's own pages here.

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.

Segment and use case coverage is described with substance, and unusually the boundary is stated by the owner rather than inferred. Segments named: corporate legal departments, state and local government agencies, law firms, educational institutions and non profits, explicitly including small teams and solo practitioners. Use cases are enumerated concretely: early case assessment, litigation, FOIA and public records requests, internal investigations, subpoena responses and data subject access requests, which is a broader public records and privacy footprint than the enterprise ediscovery records here. The limit is published in the owner's own two platform framing, which positions this product as self service discovery for firms of all sizes and the enterprise platform for the most complex needs, so a buyer is told which product is meant for which problem. Not located: jurisdictional or language coverage, and any data volume ceiling stated by the vendor rather than by reviewers.