SimpleLegal

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 on a named, shipped product. InvoiceAI is described in detail by the owner: machine learning models trained on millions of legal invoice charges, finding issues conventional billing rules miss including non working travel, block billing, vague descriptions, improperly billed administrative tasks, improper coding and wrong staff class, integrated with existing rules engines and continuing to learn from corrections. Real capability, not a roadmap. GRADE: the models are the engine of a core capability layered on a platform that would function without them. E-billing, matter intake and management, vendor management, budgets, accruals, reporting and the CounselGO portal all stand without any model, and the platform sold on exactly that from 2013 until the AI layer arrived in 2021. Recorded as a discrepancy rather than resolved: the owner's current marketing describes the product as AI native, which does not fit a platform that operated for eight years before its AI capability was introduced, and the grade follows the product history rather than the adjective.

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.

Grounding is real and the method is described with unusual specificity for this category, short of published measurement. The owner states InvoiceAI analyses historical and real time invoices, uses machine learning trained on millions of legal invoice charges, and integrates findings with existing rules engines so that machine learning and deterministic rules operate together rather than one replacing the other. The issue types it detects are enumerated rather than left general, which tells a buyer what the model is actually looking for. Every flag attaches to a specific invoice line the reviewer can open against their own billing guidelines, so verification is structural. Searched the product pages, the owner's InvoiceAI material and press coverage on 29 Aug 2026 and located no accuracy figure, no false positive rate, no test set, no evaluation methodology and no independent benchmark. The published figures are savings rather than accuracy: an average of six figures identified in travel related billed time across customers, which measures what was found rather than what was missed or wrongly flagged.

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.

A real published commitment with a documented feedback loop, short of thresholds. The owner describes InvoiceAI as handling the first pass review of incoming bills and setting up a framework that continuously learns as invoice corrections are refined in the system, which places the human at the correction point and makes their corrections the training signal. That is an oversight loop described as a mechanism rather than asserted as a principle, and it is more concrete than most records here manage. Approval workflows sit between a flag and any payment decision. Not located as of 29 Aug 2026: any threshold at which a flag is applied automatically versus routed for review, whether any reduction can take effect without human approval, and what the vendor commits to when a flag is wrong. The last matters because a wrong flag reduces a law firm's payment.

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.

Quantified figures without named customers. Scale is stated precisely by the owner and dated: more than 550 corporate legal departments and $5.2bn in annual legal spend processed as at May 2026, and processing volume is a meaningful proxy for production use in this category because it measures work actually flowing through the system rather than seats sold. Outcome figures exist but are unattributed: a published customer quote states 10 percent savings identified in legal spend, and the owner reported an average of six figures in travel related billed time identified across customers. Independent review presence exists on multiple platforms. Searched the product pages, the owner's material 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.

EVIDENCE FLOOR: see the build log warning on this record. Searched the product pages, the owner's InvoiceAI material, press coverage and three separate targeted searches for a trust centre, security page or certification listing on 29 Aug 2026 without reaching one. Nothing was located on how client confidences are handled: no statement on whether customer content may be used to train models, no retention or deletion terms, no segregation model, no treatment of privilege or work product. The question is sharper here than the absence alone suggests, because the owner's own description of InvoiceAI states the models are trained on millions of legal invoice charges and continue to learn from customer corrections, so training on customer derived data is the stated design and nothing published says whose data, whether it is pooled across customers, or whether a customer can decline. Invoice narratives describe what lawyers did on matters, which is privileged material. Rebuttable with one link and flagged as the highest priority recheck on this record.

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 corporate and the position is unstated. Users are in house legal departments, legal operations and finance counterparts, with outside counsel interacting through the CounselGO portal, and the product analyses billing rather than giving legal advice, so the advice line question arises less sharply than for a research or drafting tool. Searched the product pages, the owner's material and press coverage on 29 Aug 2026 and located no published position on advice versus tooling, no treatment of competence or supervision duties, and no jurisdiction limits. Recorded at C because the position is inferable from what the product is rather than published.

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.

EVIDENCE FLOOR: see the build log warning on this record. Searched the product pages, the owner's InvoiceAI material, press coverage and three targeted searches on 29 Aug 2026. No published AI governance framework, AI principles document, AI management certification, named owner of model governance, pre release testing regime or bias disclosure was located. The gap has a specific edge worth stating: InvoiceAI classifies whether billed work was appropriate, including judgements about whether work was performed by the correct staff class, and a model trained on historical invoice data will have learned historical staffing patterns. Nothing published addresses whether its judgements differ across firms, practice areas, staffing models or timekeeper seniority. Rebuttable with one link.

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 D on the finding that three targeted searches reached no trust centre or security page. That finding was wrong: a trust centre exists at trust.onit.com on the Vanta platform with a dedicated controls section, publishing the owner's security practices and reachable without a sales conversation. The premise of the earlier note is withdrawn. What is now established: a published security posture exists at owner level with a controls disclosure and a self serve route to it. What is still not established, and why this is C rather than higher: the trust centre renders client side and its control detail was not retrieved in this pass, so no specific control is confirmed for this product. Searched the product pages, the owner's material and press coverage on 29 Aug 2026 and located no stated retention period, no deletion control, no encryption statement for this product, no hosting disclosure, no named subprocessor list and no incident or breach notification practice. The related concern recorded elsewhere on this record still stands and is sharpened by the gap: the owner states its invoice review models are trained on millions of legal invoice charges and continue to learn from customer corrections, so what is retained and for how long is a live question that the located material does not answer. Rebuttable in one step by reading the trust centre controls.

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 product pages, the owner's material, press coverage and third party review 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 matches Brightflag and is worth restating as a category observation rather than a vendor one: in legal spend management the AI's output reduces payments to a law firm that is not the vendor's customer and has no contractual relationship with the vendor, so the party bearing the direct financial consequence of a wrong output has no recourse route at all. Neither record in this category addresses it.

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.

Integration is claimed as a differentiator with categories named rather than connectors. The vendor states seamless integration with enterprise resource planning and finance systems ensures real time visibility into spend and performance, and third party material describes support for multiple systems and platforms. CounselGO is a real named component, though it is a vendor portal the product ships rather than an integration into a third party system. For an e-billing platform the finance system connection is the integration that matters most, and naming it as a category is a genuine positioning claim. Searched the product pages, the owner's material and third party review material on 29 Aug 2026 and located no integrations index page, no named connector, no API documentation, and no statement of which billing format standards are supported.

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.

EVIDENCE FLOOR: see the build log warning on this record. Searched the product pages, the owner's material, press coverage and three targeted searches on 29 Aug 2026. Nothing was located on the deployment model: no hosting provider, no named regions, no customer selectable residency, no tenancy model, and no statement of where processing happens. The product is evidently cloud delivered, which is inference from how it is sold and earns nothing on this axis. Rebuttable with one link.

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 D on the finding that no trust centre, security page or certification listing existed after three targeted searches. That finding was wrong. A trust centre does exist at trust.onit.com, operated by the owner on the Vanta platform, with a dedicated controls section and a stated commitment to protecting customer data through published practices and transparency. It is reachable without a sales conversation, which under the three tier test is a self serve route rather than a gate. That alone lifts this off a pure absence. The owner also publishes SOC 2 Type 2 attestation for at least one other product in its portfolio, a virtual data room, alongside AWS hosting, permission based user roles, two factor authentication and encryption, which evidences that the owner holds product level certifications and publishes them. Why the grade is C rather than higher. The trust centre renders its content client side and its certification list was not retrieved in this pass, so no certification is confirmed as covering this product specifically. No coverage period, audit scope, report date or auditing firm was located for this product. The SOC 2 located belongs to a sibling product and was not read across. Rebuttable in one step by retrieving the trust centre's compliance list, which remains the highest value check on this record.

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 vendor's own model layer is described and nothing beneath it is identified. Published: InvoiceAI as proprietary machine learning trained on millions of legal invoice charges, working alongside existing rules engines, and developed by the owner rather than licensed in, which tells a buyer the invoice review capability is not a thin wrapper on a third party model. The owner separately built or acquired several other AI products, so an in house model capability is evidenced. Searched the product pages, the owner's material and press coverage on 29 Aug 2026 and located no named foundation model or provider, no statement of where models run, no subprocessor list, and no commitment to notify customers when the supply chain changes. Note the currency question flagged elsewhere on this record: the detailed AI description dates from 2021 and current product pages describe AI capability without naming InvoiceAI, so what powers the product today is not confirmed.

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.

Checked the product pages, the owner's material and third party review material on 29 Aug 2026. No pricing page was located, no rate is published, no unit of charge is stated and no tier structure appears. Third party sources confirm the absence directly, one stating there is no public pricing as at May 2026 and another that plans are customisable and require a personalised quote. Every commercial path located terminates in a demo or quote request. Same pattern as the other record in this category, and the same irony applies: a platform sold on delivering visibility and control over legal spend does not publish what it costs.

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 coverage is described with substance and, unusually, the boundary is published by the owner rather than inferred. The buyer is stated as in house legal teams needing matter management, e-billing and spend visibility without full enterprise implementation complexity, and the owner positions this product explicitly against its own enterprise tier platform, so a prospect is told which of the two is meant for their situation. That is a vendor stating what its product is not for, which few records here do. Scale is quantified at more than 550 corporate legal departments. Practice scope is clear and consistent as spend, matters, vendors and reporting, with no claim to advisory or drafting capability. Not located as of 29 Aug 2026: an enumerated industry list, organisation size bands stated numerically, jurisdictional coverage, and any language support statement.