D
Data Safeguard
Data Safeguard sells privacy operations software for compliance, data protection and legal teams from Santa Clara, California. ID-PRIVACY combines consent management, sensitive data discovery, privacy impact assessments, individual rights requests, redaction and masking, compliance audits, program management and incident workflows. A data subject access request can draw records from business systems and document repositories.
ID-REDACT uses the company's Cognoscible Computing Engine to identify and remove sensitive content before a response goes out, and ID-MASK substitutes values for work such as nonproduction software testing.
ID-REDACT handles PDF, Word, PowerPoint, Excel and CSV files, and enterprise deployments can also process database, email and chat feeds. The self service workspace lets a user choose a processing mode and preview the redacted file and its certificate before saving them. Enterprise customers can run the software in their own data center or cloud so the data stays within their network, while the hosted Customer API and storefront process uploaded material in Data Safeguard's environment. Microsoft Marketplace distributes ID-REDACT, and Unizo describes helping Data Safeguard connect to enterprise systems.
The storefront sells redaction at $15 per document, $49 per month for 60 documents or $429 a year for 500 documents, with no page volume cap, and the Customer API starts at $99 a month. Enterprise options are quoted. SOC 2 reports and ISO information security and privacy management certificates can be requested under mutual nondisclosure.
Alternatives to Data Safeguard
Other products that do the same kind of work as Data Safeguard, and what each one does.
- SSaifrSaifr is a compliance AI business of FMR LLC, the parent of Fidelity Investments. It began in Fidelity Labs, Fidelity's in house software incubator, and operates through Fidelity Labs, LLC and Saifr.ai, LLC, with offices in Boston, New York, Westlake in Texas, Dublin, Chennai and Bangalore.
- PPhaselawPhaselaw is a document review and redaction platform from Phaselab, Inc. of New York, built for answering subject access requests, freedom of information and public records requests, and…
- RRed MarkerRed Marker helps enterprise legal, compliance and marketing teams review advertising for regulatory and brand risks.
- TTraactTraact is a web platform and managed service for corporate entity management, registered agent work, annual reports, business licenses and governance records.
- HHaastHaast reviews regulated marketing and communications against an organization's policies and risk tolerance. It applies a separate AI agent to each compliance task, flags concerns in draft content and suggests corrections.
- AAscentAIAscentAI is a regulatory lifecycle management platform for financial institutions. It is built on the idea that a firm should hold a register of the specific obligations that apply to it, rather than a library of rules at large.
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.
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.
The Cognoscible Computing Engine supplies detection and classification to the privacy platform. ID-REDACT identifies and removes sensitive values in files, databases, and streams, and ID-MASK substitutes values to limit exposure in other environments. The vendor describes recognizing a number from its surrounding context as part of that process. The wider product separately offers consent, assessment, incident response, and audit workflows.
Those workspaces organize privacy tasks and deadlines. The models do a core part of finding and treating the underlying data, while the privacy program also uses rule based workflows and records that persist independently of a particular model run.
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 automated output labels or removes data elements in records the customer supplies. The solutions page claims 94.5% accuracy for detecting, identifying, confirming, tagging and redacting information, and the Azure listing claims 99.54%. Neither figure comes with a test set, a sample breakdown, a false negative rate or a method that lets a customer compare the two. The workflow produces redaction logs and certificates for examining each operation. It produces no case law citations, and nothing validates statutory points a compliance team might draw from a result.
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.
Users can pick a processing mode and preview the redacted document and certificate before saving them, and the storefront also offers manual redaction. The API tiers run detection, tagging and redaction, with identification or confirmation added in the higher tiers. Enterprise processing can run on live streams or recurring batches. The help manual sets no confidence level that stops automated redaction and prescribes no approval step for every enterprise batch or subject access response.
The controls described are the choice of mode, the ability to examine an output and the operations that can run automatically.
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.
ID-REDACT is sold through Microsoft Marketplace, and a customer manual covers the storefront workflow. Unizo describes working with Data Safeguard to offer more enterprise data connections without dedicated integration engineering, which concerns the vendor's own product distribution rather than an end customer's result. Expected reductions in privacy work and separate accuracy claims for redaction appear on the site.
No named end user's fulfillment time before and after, volume of records tested or process for evaluating redaction errors in production is given. The Marketplace listing and support materials establish that the product is available and show how it operates.
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.
In the customer’s own data center or cloud, the enterprise deployment keeps documents and processing within that environment. The self service API and storefront process material in Data Safeguard’s cloud. The vendor states it does not claim rights to customer data. It also describes its engine as able to learn from a company’s existing data without specifying whether that means inference, tenant tuning, or updates to shared models.
Its public agreement materials give no separate commitment on legal privilege, matter level permissions, or a time limit for keeping uploaded files and results in the hosted service.
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 platform is sold as software for privacy operations, with separate capabilities for consent records, privacy impact assessments, personal data requests, and compliance audits. Compliance teams use its rules and report functions to document their work across named privacy frameworks. The product guide illustrates reviewing a redacted result and its certificate before release. The vendor describes the technology as supporting the organization’s privacy and compliance process.
Its materials do not define advice boundaries for each jurisdiction or discuss how an attorney’s supervision and competence duties apply when the platform evaluates a request or prepares records for disclosure.
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.
Contextual classification, a rules engine for mapping privacy controls, and redaction certificates and audit records are part of the product, and the company says its products went through testing with customer feedback. Who approves model releases, which tests cover different types of personal data, and how performance compares across demographic groups, languages and file types are not stated. The 94.5% and 99.54% accuracy claims come with no method for evaluating those differences.
The compliance audit capability helps a customer document its own controls, but says little about how the models that produce privacy detections are governed.
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.
The privacy policy describes data minimization, deleting data promptly once it is no longer required, and the circumstances for disclosure to contracted processors. The enterprise product can run within the customer's infrastructure, while self service workflows upload files to Data Safeguard. The manual asks users to save redacted records and certificates to their own location. How long the hosted platform keeps input documents, certificates and audit logs is not specified.
No subprocessor roster, processor breach notice time or storage schedule covers the records used in hosted privacy workflows.
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.
A storefront redaction run produces a certificate that records the result. Support comes with different response levels by deployment: a 12 month SLA backed support term for enterprise plans, a three hour response for the Customer API and an eight hour email response for the storefront. There is no indemnity, no warranty of redaction accuracy and no statement of who bears the loss if private data stays visible in a released document.
Registration guidance refers to terms or a user agreement in the sign up process, but no liability cap or contractual remedy for errors is posted. Enterprise implementation is offered through professional services.
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.
Enterprise ID-REDACT accepts real time feeds from email, chat and web forms and runs scheduled batches against existing databases and data lakes. Named input sources include SharePoint, Documentum, Google Drive, OneDrive, Box and Dropbox. The platform processes selected documents and lets users save redacted files and certificates to a chosen location, and the self service API accepts document and historical data uploads.
Unizo describes helping Data Safeguard connect its platform to external enterprise sources, and Azure Marketplace provides a separate distribution route. The deployment guide explains when sources are integrated and when users upload files.
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.
The on premises deployment can process documents and records inside a customer’s own data center. The enterprise cloud and Marketplace options run inside customer controlled cloud environments. The hosted Customer API and storefront instead receive uploads for processing in Data Safeguard’s environment. The disclosures therefore establish different hosting and data flow options by purchase route. They do not name hosted processing regions or storage locations for originals and output, or give a region selection table for the hosted plans. The seller describes the infrastructure boundary for enterprise processing separately from the file upload route.
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.
SOC 2 Type 1 and Type 2 reports and ISO/IEC 27001 and 27701 management system certifications are named, and the product FAQ describes independent assessment and distinguishes the SOC reports from the ISO criteria. Buyers can request current documents under a mutual nondisclosure agreement. No audit period, report scope, certificate date or copy for direct review is posted. The security pages also describe enterprise deployment inside a buyer's environment, where the buyer's own controls apply, separately from the assurances that cover the service provider.
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.
Data Safeguard describes its Cognoscible Computing Engine as its own combination of models and deep learning, built internally, and says the privacy products use no third party products or services for processing. The Marketplace listing uses similar language and says the redaction models are hand coded. That establishes parts of the architecture and who builds it. No model version, outside model supplier or change notification process is given.
Azure Marketplace distribution and customer cloud deployment describe how the software is acquired or run, not a model supplier.
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.
On the storefront, redaction costs $15 per document, $49 a month for 60 documents or $429 a year for 500 documents, with no page volume cap. The product offering page lists the same annual price with 600 documents, so the two pages disagree on the annual allowance. The free trial covers one document a day for seven days. The hosted Customer API lists $99, $199 and $299 a month for its three processing tiers, with an allowance of ten API calls a month.
Enterprise on premises, customer cloud and Marketplace API deployments are quoted on use and volume, with 12 months of support included. Professional implementation is available for enterprise buyers, and no separate setup rate is listed.
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.
Privacy and compliance teams in financial services, healthcare, government, retail and technology are the stated buyers. The privacy product materials name GDPR, California privacy requirements, HIPAA and India's data protection framework, and an India privacy notice sits beside the main policy. Enterprise deployment options serve customers who need to keep processing in their own infrastructure. The storefront offers redaction of single documents, and the inquiry workflow addresses obligations to disclose personal information.
No specialized package for law firms, eligibility by customer size or list of privacy controls that vary across deployments is described.
6 public documents
The public pages on file for Data Safeguard, 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.
-
datasafeguard.ai/faqs4 signals
Ethical Walls and Matter Segregation, Good Law Verification, Bar Guidance Alignment and 1 more
Read Nov 19, 2026
-
Prompt and Output Retention, Third Party Request and Subpoena Notice
Read Nov 19, 2026
-
datasafeguard.ai/help-manual2 signals
Refusal and Uncertainty Behavior, Court Disclosure Support
Read Nov 19, 2026
-
datasafeguard.ai/solutions2 signals
Client Data in Training, Fabricated Citation Record
Read Nov 19, 2026
-
datasafeguard.ai/data-privacy1 signal
Primary Law Corpus Provenance
Read Nov 19, 2026
-
Billing and Fee Posture
Read Nov 19, 2026
$15 per document
- The storefront sells redaction at $15 per document, $49 a month for 60 documents or $429 a year for 500 documents, with no page volume cap. The Customer API starts at $99 a month. Enterprise pricing is quoted.
Document transactions and subscriptions are advertised for the storefront. Hosted API subscriptions have three tiers based on the processing steps included.
Implementation: Professional services implement enterprise deployments. No separate setup rate is listed.
Note: The storefront, where the plans are bought, lists 500 documents a year for $429, and the product offering page lists 600 for the same price. The seven day trial allows one document per day, and its link opens a contact form rather than a checkout. Support is an eight hour email response on the storefront, three hours on the API and a 12 month SLA term on enterprise plans.
Legal Signals
What each signal meansA 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.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
A published agreement or policy exists and none of it addresses the question either way, or the document that would answer it could not be read and the summary names the retrieval limit. The summary states which shape the silence takes: an improvement right granted that never names training, or no improvement right granted at all.
The privacy solution states that it "Does not claim rights to customer data", while the home page describes the engine as able to learn from a company's existing data. No agreement or DPA is posted, the product offering page says nothing on training, and neither statement says whether customer content trains, tunes or personalizes a model, or whether any learning reaches models shared across customers.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
Retention is acknowledged in public materials with no stated period.
The privacy policy promises to delete data promptly once it is no longer required. The hosted product sets no time window for keeping original files, redacted output, certificates or audit logs. The manual asks users to save their results to their own location.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
No located public material addresses walls or matter level segregation.
Enterprise processing can stay inside the customer's own data center or cloud, within its secured network. No technical separation between the matters or requests of one legal team is described.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Published terms or policy address disclosure to authorities or in response to legal process, and no commitment or reservation regarding customer notice is located anywhere. The vendor has told the customer that data can leave and has said nothing about whether the customer hears of it.
The privacy policy permits disclosure when the law requires it. It does not promise to tell the customer about a third party legal request before the disclosure is made.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
No located public material identifies the corpus behind the product’s answers.
The detection engine works on customers' documents, databases and business records, and the rules engine maps controls to named privacy laws. No legal authority collection or license for one sits behind the privacy product.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
No located public material addresses whether authority is checked for subsequent history.
The product applies privacy rules and classifies data elements in customer records. It cites no case law, and no citator or check on the status of authority is part of it.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
No located public material addresses what the product does when it cannot ground an answer.
A user can choose a processing mode, preview the redacted result and fall back to manual redaction. No condition is described under which the model withholds an uncertain detection or routes it to a reviewer.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
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 maintained by Damien Charlotin lists no case involving fabricated citations generated by Data Safeguard. Its redaction functions classify and remove content in supplied material and produce no legal citations.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
No located public material engages with bar or ethics guidance.
The platform maps privacy controls to GDPR, California privacy law, HIPAA and India's framework for privacy teams. It names no bar guidance governing lawyers' use of AI.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
The product sits inside a lawyer to client fee relationship and no located public material addresses billing, fee or disclosure treatment, with no savings claim published either.
Data Safeguard bills per document, per month or per year on the storefront and by subscription for the API and enterprise editions. The buyers are privacy and compliance teams, and nothing in the product records AI work on a legal fee invoice.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
The material exists behind a sales conversation or an executed agreement.
SOC 2 reports and ISO 27001 and 27701 certificates go to prospective customers under a mutual nondisclosure agreement. No complete subprocessor list is offered for the hosted storefront or API.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Some elements of the record are available, short of a document level export.
Each redaction run produces a certificate and an activity log, and users can save the certificate with the redacted file. No single export records the model version, the underlying material and the human verification together, as a court disclosure would need.