D
Dioptra

Dioptra is an AI contract review add-in for Microsoft Word that drafts redlines on a company's own paper and on counterparty paper. Icertis acquired the Y Combinator backed startup in November 2025 and still sells Dioptra under its own name to in house legal teams, from solo general counsel up, and to law firms of every size. Wilson Sonsini launched AI document review with it in 2024.

A lawyer reviews an agreement against a playbook, standard forms, past executed agreements, example redlines or a plain prompt, and Dioptra returns tracked changes in Word with an issues list and a risk summary. Playbook distillation builds playbooks from a company's executed agreements, and market playbooks and a prompt library cover common contracts. A clause library inserts approved language with one click, term extraction pulls terms from executed contracts in bulk, and a chat assistant answers questions about a document.

Dioptra reports 95% accuracy on a company's own paper, 92% on counterparty paper and 97% on term extraction. Starter costs $450 a month or $4,320 a year with a 7 day free trial, Teams starts at $4,200 a year per seat, and Enterprise is custom, with demos booked through Icertis.

Capability grades

All 15 axes, graded from public sources on the date shown. Hover a grade to see what the letter means on that axis.

AA on AI CentralityThe artificial intelligence is the product. Remove the models and there is nothing left to sell.

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.

Every core function runs on models. Redlines are generated from a playbook, standard forms, past agreements, example redlines or a prompt, and playbooks are distilled from a company's executed contracts. Issues lists and risk summaries are written for each review, non market provisions are flagged, terms are extracted from contracts in bulk, and a chat assistant answers questions about a document. The clause library is the one manual tool, and without the models the add-in would have nothing to sell.

Source: Vendor Published
CC on Citation Accuracy and Hallucination DisclosureAccuracy is asserted without measurement, or grounding is claimed while output cites sources the reader cannot open and verify.

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.

The homepage claims more than 90% accuracy, broken down as 95% on a company's own paper, 92% on counterparty paper, 97% on term extraction and 97% on issue flagging. No test set, contract mix, method or date sits behind the figures. A Forbes piece from June 2024 quoted a user praising its accuracy on complex contract review, and Wilson Sonsini backed that accuracy when it launched AI document review with Dioptra in May 2024.

Redlines appear as tracked changes in the agreement itself, and how each change ties back to the playbook position that produced it is not described.

Source: Vendor Published
BB on Autonomy and Oversight ModelA written commitment that the models work alongside a supervising lawyer, with real review surfaces, short of the full control structure: commonly the threshold at which the system stops or what happens after it is wrong.

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.

Dioptra proposes and the lawyer decides. Redlines arrive as tracked changes in Word for the reviewer to accept, reject or edit, alongside an issues list and risk summary, and nothing is sent to a counterparty by the product. The terms make the customer solely responsible for using the outputs and judging their accuracy and appropriateness. Playbooks set the positions the AI drafts to, and Enterprise adds playbook building services and a usage dashboard. No threshold, confidence signal or limit on contract types where the AI should not draft is given.

Source: Vendor Published
BB on Operational and Outcome EvidenceReal deployment evidence with substance, short of full attribution or measurement: a named customer without figures, or figures without the named customer.

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.

David Wang, former chief information officer at Wilson Sonsini, calls Dioptra a Sputnik moment for partner level contract assessment, and the firm launched AI document review with Dioptra in May 2024. Vanessa, legal director and assistant general counsel at Collibra, says it saves her team many hours of redlining, and David M., a director at Fennemore, praises its precision and integration. Unnamed testimonials report a two hour review done in 30 minutes and a 45 minute review cut to 10 for a large firm associate in Virginia.

Others cite a seven hour drafting task cut to one hour at an Israeli firm and time saved of more than 80%. Nelson Mullins, Cornerstone, LawVu and CyberOne appear among the logos, and none of the figures carries a date or method.

Source: Vendor Published
BB on Privilege and Confidentiality PostureSubstantive published commitments on confidentiality and training use, short of the full picture: commonly silence on segregation between users or matters, or on what the underlying model provider may retain.

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.

Section 1.3 of the terms bars Dioptra from training its own AI models on customer materials or letting them train language model providers' models, and limits sharing to suppliers needed to run the service. Customer materials are confidential information, deleted within a reasonable period after termination, and the language model providers sit under zero data retention agreements. A self hosted deployment is available.

Usage data on AI interactions can still be used to improve the service, and matter level separation and privilege are not addressed.

Source: Vendor Published
BB on UPL and Professional Responsibility PostureA real position is published on advice versus tooling, short of full treatment: commonly a disclaimer without the supervision and competence dimension, or silence on jurisdiction limits.

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.

The terms state that the services are not substitutes for the advice of an attorney and tell users to consult a licensed attorney, and they disclaim any warranty that output is accurate or correct. The product is sold to in house counsel and law firms, while paralegals, business operations staff and small business owners also use it. Supervision of nonlawyer users and jurisdiction limits for the market playbooks are not addressed.

Source: Vendor Published
DD on AI Governance and Bias DisclosureNo governance position published for a system whose output affects legal outcomes.

AI Governance and Bias Disclosure

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

No responsible AI policy, named owner for the models, release testing regime or review of uneven results across contract types, paper or jurisdictions is offered. The accuracy figures on the homepage and the prompting office hours included with Starter are the only material on model quality. The guided ROI programs in Teams and Enterprise measure time saved rather than the quality of the redlines.

Source: Operator Verified
BB on AI Safety and Data StewardshipSubstantive published policy covering most of the ground, short of the full set: commonly no named subprocessor list or no stated incident practice.

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.

Data is encrypted at rest with Amazon S3 server side encryption and in transit with TLS 1.2. Audit logs from AWS CloudWatch and GuardDuty with rules based alerting are kept for up to two years, and third party vulnerability assessments run at least once a year. Language model providers work under zero data retention agreements, and every other subprocessor under a data processing agreement. Customers keep control of their data and outputs, and customer materials are deleted within a reasonable period after termination, with backup copies kept confidential. No named subprocessor list or incident notice timetable is given.

Source: Vendor Published
BB on AI Liability and RecourseA real published position on liability, short of the full picture: commonly a stated indemnity without scope or caps.

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.

Each party's liability is capped at the fees paid under the order in the 12 months before the event, with gross negligence, willful misconduct, fraud and indemnities outside the cap, and consequential damages are waived. Dioptra defends and indemnifies customers against third party claims that the service infringes intellectual property, capped at $1,000 for free trials. Output carries no accuracy warranty. New York law governs, with small claims under $10,000 open to non appearance arbitration.

Source: Vendor Published
CC on Practice Systems Integration DepthIntegrations are listed as logos or marked as coming, with no documentation an implementer could use.

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.

The Word add-in is the working surface, generating tracked changes and inserting clauses inside the agreement. Teams and Enterprise add Bring Your Own Repository for a company's contract store, and connections to CLM, CRM and procure to pay systems are said to take days rather than weeks. None of those systems is named, and nothing explains what moves between them and Dioptra. Enterprise single sign on is the only identity connection listed.

Source: Vendor Published
BB on Deployment Model and Data ResidencyDeployment model is stated clearly with partial residency detail, or residency is offered without the processing location being addressed, or the tenancy model is stated on its own with no residency detail published.

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.

Dioptra runs as a hosted service on AWS by default, with storage encrypted in Amazon S3, and a self hosted deployment is offered for customers who need one. The review runs on private language models under zero data retention agreements. No hosting region, residency choice or location for language model processing is named, and what changes in the self hosted option is not described.

Source: Vendor Published
BB on Security Certifications and Trust CenterCertification is real and stated, short of accessible evidence: a named standard without scope, date, or a way to obtain the report.

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.

A SOC 2 Type II report covers the service, and the terms commit to showing customers proof of security measures once a year on request, met by a copy or extracts of that report. AICPA SOC and GDPR badges and annual third party vulnerability assessments sit beside it. No trust center, audit period or way to obtain the report before signing is offered.

Source: Vendor Published
CC on Model Supply Chain DisclosureThe vendor refers to advanced or proprietary models without identifying what sits underneath.

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.

Private language models power the review, and the terms let Dioptra send customer materials to third party large language models and other AI subprocessors, all under zero data retention agreements, with data processing agreements for every subprocessor. The providers and models are not named, and nothing says when customers hear of a change.

Source: Vendor Published
BB on Commercial TransparencyReal pricing is published for part of the range, with enterprise tiers withheld, or the unit and structure are stated without the figure.

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.

Starter costs $450 a month or $4,320 a year for one user, with unlimited reviews, the Word add-in, market playbooks, a prompt library and prompting office hours, and a 7 day free trial signs up without a call. Teams starts at $4,200 a year per seat with a 15 day guided ROI program, a shared prompt library, a 30 day onboarding plan and Bring Your Own Repository. Enterprise is custom, with a 30 day ROI program, playbook building services, single sign on, dedicated support and a usage dashboard, and both larger plans book demos through an Icertis calendar.

Fees are paid annually upfront, orders renew unless canceled 30 days before the term ends, and renewal prices can rise on 30 days' notice.

Source: Vendor Published
BB on Firm and Practice CoverageSegment and practice coverage is described with substance, short of the boundaries: what is supported is clear, what is not is left open.

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.

Plans run from solo general counsel and solo practitioners on Starter to teams and enterprises. In house testimonials come from legal directors, deputy and assistant general counsel, chief commercial counsel and paralegals. Law firm testimonials come from partners, associates and innovation officers at large California, Illinois and Virginia firms, an Israeli firm and a New York boutique, and business operations staff and small business owners use it too.

The product handles both a company's own paper and counterparty paper. Contract types and jurisdictions outside commercial agreements are not listed.

Source: Vendor Published
Sources on file

4 public documents

The public pages on file for Dioptra, with the recorded signals each one supports and the date it was last read. Open any of them and check the reading against the record.

  • dioptra.ai6 signals

    Prompt and Output Retention, Primary Law Corpus Provenance, Good Law Verification and 3 more

    Read Nov 20, 2026

  • Client Data in Training, Third Party Request and Subpoena Notice, Bar Guidance Alignment and 1 more

    Read Nov 20, 2026

  • Fabricated Citation Record

    Read Nov 20, 2026

  • Ethical Walls and Matter Segregation

    Read Nov 20, 2026

Pricing

$450/month for one user

USD, as published, never converted
  • Starter costs $450 a month, or $4,320 a year, for one user.
  • It comes with a 7 day free trial and unlimited contract reviews.
  • Teams starts at $4,200 a year per seat.
  • Enterprise pricing is custom.

Starter is for solo general counsel and solo practitioners at $450 a month or $4,320 billed annually, with redlines from playbooks, forms and past agreements, the Word add-in, unlimited reviews, market playbooks, a prompt library and prompting office hours. Teams starts at $4,200 a year per seat and adds a 15 day guided ROI program, a team prompt library, a 30 day onboarding plan and Bring Your Own Repository. Enterprise is custom and adds a 30 day ROI program, playbook building services, enterprise single sign on, dedicated support and a usage dashboard. Teams and Enterprise are sold after a demo.

Implementation: No setup fee is listed. Teams includes a 30 day customized onboarding plan, and Enterprise includes playbook building services.

Note: Fees are paid annually upfront unless the order says otherwise and are due within 30 days of invoice. Orders renew unless either side gives notice 30 days before the term ends, and renewal fees can rise on 30 days' notice. A free trial converts to a paid subscription unless the customer opts out, and fees are not refundable except as the terms or the order provide.

Legal Signals

What each signal means

A signal records what public sources say on the date shown. It is not a grade and it is not a recommendation. Where a signal reads Not addressed, it means the index did not locate the material in public sources on that date, which is a statement about disclosure rather than about the product.

Confidentiality and Privilege

Client Data in Training

Can material a lawyer puts into this product be used to train a model?

Never, in the contract

The published terms prohibit training on customer content. Not a policy page, the agreement.

Section 1.3 of the terms, updated November 23, 2025, bars Dioptra from training its own AI models on customer materials and from letting them train language model providers' models. Section 1.5 still lets Dioptra use usage data about AI interactions to improve the service.

Source: Vendor Publishedallow Customer Materials to be used to train large language model provider modelsAs of Nov 20, 2026Evidence

Prompt and Output Retention

How long does the product keep what a lawyer typed, and can that be set to zero?

Disclosed without a period

Retention is acknowledged in public materials with no stated period.

Language model providers keep nothing under zero data retention agreements, and audit logs are kept for up to two years. Inside Dioptra, customer materials are deleted within a reasonable period after termination, with no set period, and backup copies are kept confidential.

Source: Vendor PublishedWe have Zero Data Retention Agreements with our LLM providersAs of Nov 20, 2026Evidence

Ethical Walls and Matter Segregation

Does retrieval respect the firm’s ethical walls, or can the model read across them?

Not addressed

No located public material addresses walls or matter level segregation.

Teams share a prompt library and Enterprise adds single sign on, but no control keeps one matter's or client's contracts apart from another's within an account. Bring Your Own Repository connects a company's contract store without a described permission model.

Source: Operator VerifiedAs of Nov 20, 2026Evidence

Third Party Request and Subpoena Notice

If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?

Notice committed

Terms commit to notice where lawfully permitted. No transparency report located.

A party compelled to disclose the other's confidential information must give advance notice where the law allows and possible, and in any case within 5 business days of the disclosure, and must support efforts to limit it. Customer materials count as the customer's confidential information.

Source: Vendor Publishedgive the Disclosing Party, advance notice prior to disclosureAs of Nov 20, 2026Evidence
Accuracy and Authority

Primary Law Corpus Provenance

Where does the law in this product come from, and does the vendor have the right to use it?

Not addressed

No located public material identifies the corpus behind the product’s answers.

Reviews draw on a company's own playbooks, forms and executed agreements plus out of the box market playbooks, with no named source of law behind the market positions. Playbook distillation builds a company's own positions from its past contracts.

Source: Operator VerifiedAs of Nov 20, 2026Evidence

Good Law Verification

Does the product tell you when the authority it just cited has been overruled?

Not addressed

No located public material addresses whether authority is checked for subsequent history.

Contract review cites no case law or statutes, so no check of later treatment applies. Market positions come from out of the box playbooks rather than cited authority.

Source: Operator VerifiedAs of Nov 20, 2026Evidence

Refusal and Uncertainty Behavior

What does the product do when the answer is not in the corpus?

Not addressed

No located public material addresses what the product does when it cannot ground an answer.

Nothing describes what the add-in does when a playbook has no position on a clause or when it is unsure of a redline. Every proposed change lands as a tracked change for the reviewer, with no confidence marker.

Source: Operator VerifiedAs of Nov 20, 2026Evidence

Fabricated Citation Record

Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?

None located

No court order, opinion or disciplinary record addressing fabricated or hallucinated legal citations produced by this product has been located as of the date shown. This is a statement about the public record on that one subject, not a finding about the product, and this signal is not a litigation history.

The AI Hallucination Cases database kept by Damien Charlotin has no entry naming Dioptra or Icertis. The product redlines contracts rather than drafting court filings.

Source: Operator VerifiedAs of Nov 20, 2026Evidence
Professional Responsibility

Bar Guidance Alignment

Has the vendor engaged in public with the ethics opinions its buyers are bound by?

Not addressed

No located public material engages with bar or ethics guidance.

The terms say the services do not replace an attorney's advice and tell users to consult a licensed attorney, with no reference to bar rules or ethics opinions on lawyers' use of AI. Nothing addresses a firm's duties when paralegals or business staff run reviews.

Source: Operator VerifiedAs of Nov 20, 2026Evidence

Billing and Fee Posture

Does the vendor address what happens to the bill when the work takes an hour instead of six?

Savings claims only

Public materials claim time savings without addressing billing or disclosure, and the product sits inside a fee relationship between a lawyer and a client where those savings would change the bill.

Law firms are among the buyers, and the homepage promises reviews three times faster, citing a two hour review done in 30 minutes and a 45 minute review cut to 10 for a large firm associate. Nothing covers how that time appears on a client bill, and the guided ROI programs measure savings for the buyer.

Source: Vendor PublishedAs of Nov 20, 2026Evidence

Outside Counsel Guideline Readiness

Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?

On request only

The material exists behind a sales conversation or an executed agreement.

Customers can ask once a year for proof of security measures, met by a copy or extracts of the SOC 2 Type 2 report. Subprocessors and language model providers are covered by data processing and zero retention agreements but are not listed, and no client facing AI disclosure pack is offered.

Source: Vendor PublishedAs of Nov 20, 2026Evidence

Court Disclosure Support

If a judge’s standing order requires an AI disclosure, can the product produce one?

Not addressed

No located public material addresses court disclosure or verification certification.

No record of AI involvement in a reviewed agreement is produced for disclosure to a court or a counterparty. Redlines arrive as ordinary tracked changes in Word.

Source: Operator VerifiedAs of Nov 20, 2026Evidence
Contact

Correct a record, or ask how something was graded

Every grade and every signal on this index is drawn from public sources and dated. If a record is wrong, out of date, or missing an artifact the index did not locate, send the source and it will be reviewed and the record redated. Vendors are welcome to submit documentation. Nothing on this index is for sale, including a listing, a placement, or a grade.

AI Legal Index

The AI Legal Index is an independent index that tracks changes to AI vendors in legal. It holds 617 vendors across 9 categories, each graded on the same 15 capability axes and recorded against 12 legal signals, from public sources. No vendor pays for inclusion, placement, or rating.

Index Status
Last index update
November 20, 2026
The AI Legal Index is an editorial reference. It is not a regulatory body, not a law firm, and nothing published here is legal advice or a recommendation to retain or avoid a vendor. Records are verified against published sources, bar guidance and public court records. Where a record reads not addressed, the material was not located in public sources on the date shown. See the Methodology page for evaluation standards and limitations.
© 2026 AI Legal Index
3801 N Capital of Texas Hwy, Ste E240 · Austin, TX 78746