Hanzo vs Smarsh: how they compare in 2026
Hanzo and Smarsh both preserve and collect collaboration data that ordinary ediscovery tools handle badly, such as Slack, Teams and chat, and both apply AI to narrow it. Hanzo is stronger on where its AI runs and what it stands behind. It states that Spotlight AI runs on local language models inside the customer's own tenant, explains why each item was flagged, and published a 95 percent recall figure at its 2023 launch. Its IBM Marketplace agreement adds an intellectual property indemnity and $5 million each of general and professional liability insurance. Smarsh is stronger where AI is central and evidenced in use. Its agents drive surveillance and investigation, and a commissioned Forrester study reports a 124 percent three year return with its method described. Its terms state plainly that the service does not guarantee legal compliance. Neither vendor's terms say whether client data trains its models.
Which one fits
Choose Hanzo if
- Your collaboration data cannot be sent to an outside AI provider. Hanzo states that Spotlight AI runs on local language models deployed at tenant level inside your environment, and it is switched on per matter, so review AI is optional and stays inside your perimeter.
- You need Slack, Teams and Google Workspace preserved as conversations, not flattened exports. Hanzo's Illuminate preserves Slack in place with holds by custodian, channel and timeframe, keeps threading, reactions, edits and deletions intact, and captures messages posted after a hold is set.
- You want the vendor insured against getting it wrong. Hanzo's IBM Marketplace agreement commits it to carry $5 million of commercial general liability and $5 million of professional liability insurance for the term and two years after, alongside an intellectual property indemnity and a twelve month liability cap with carve outs.
Choose Smarsh if
- You need one archive for every channel your regulated staff use. Smarsh captures email, instant messaging, collaboration tools, voice, social media and mobile messaging from carriers and apps including WhatsApp and Signal into a retained archive built for SEC Rule 17a-4 and FINRA supervision.
- You want an outcome study with a method, not only a claim. Smarsh publishes a Forrester Total Economic Impact study from September 2024, built from interviews with six financial services firms and risk adjusted, reporting a 124 percent return over three years and discovery time down 65 percent.
- Your investigators work across languages. Smarsh's AI Assistant summarizes and translates communications, and its Discovery Agent builds summaries, timelines and custodian maps for legal investigations so less material has to be exported to outside counsel.
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.
AI is optional and peripheral by Hanzo's own design. The Illuminate page describes Spotlight AI as an optional element that can be switched on within the platform. The FAQ describes two core products that work without it: Chronicle for website and social media preservation, and Illuminate for hold, collection and review of collaboration data. Hanzo's own 2023 account of how Spotlight AI began describes it as a feature added to a culling and preservation platform whose strength was culling large data sets.
The capabilities the vendor now leads with are driven by models, on a capture and archive platform that works without them. Two agents filter surveillance alerts and detect risk, the AI Assistant summarizes and translates, and the Discovery Agent builds summaries, timelines and custodian maps for investigations. Underneath sits a capture, retention, legal hold, search and export platform, sold for decades as a recordkeeping system of record, that functions fully without models.
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.
Hanzo publishes a measured figure and explains its outputs, but describes no test set and has not republished the figure. The November 2023 launch release states ninety five percent recall in identifying documents pertinent to a case, with no test set, method or date of measurement, and there is no newer figure for the current product. The FAQ and Spotlight AI page say results are ranked and tagged so reviewers know why something was flagged. Hanzo's engineering post describes facets written by generative AI that users can adjust and supply examples for, which grounds the review in the case description. The figure is three years old, and questions about primary authority do not apply to a review tool.
Performance figures are published without a method, and grounding is claimed rather than described. Releases and pages say the Noise Reduction Agent cuts false positives by 60 percent, the Intelligent Agent surfaces three to five times more real risk, and the Discovery Agent cuts investigation costs by up to 75 percent. No sample, baseline or test is described for any of them. The same FAQ grounds AI summaries and risk signals in traceable source data, but does not describe how a summary links back to the messages it relies on, or what happens when the AI misreads a conversation.
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.
The modes, constraints and review surface are published; the full control structure is not. Spotlight AI is optional and switched on per matter. It flags anomalies and relevant content for human review, with an explanation for each flag. The model generates the facets that drive relevance, but users can add to, adjust and clarify them with examples, which Hanzo describes as giving the user control over first pass review while computers take the laborious steps, and it says the explanations make every decision defensible. What is not published is the threshold at which content is excluded from review without a person, or any route back after a wrong relevance call beyond adjusting the facets.
Alert suppression is automated, and no human control over it is published. The Noise Reduction Agent and Intelligent Agent suppress alerts of low relevance so supervisory teams see fewer, and the vendor calls these autonomous systems that augment rather than replace human expertise. Nothing published says whether suppressed alerts can be reviewed or are sampled, what threshold governs suppression, or who approves the agents' configuration. That matters because supervisory review of communications is a regulatory duty for the vendor's financial services customers. Discovery Agent output is presented as input to investigators.
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.
Figures appear without a named customer. Hanzo's FAQ says first pass review moved from weeks to hours for some teams and that teams using Illuminate with Google Workspace data work up to fifty percent faster, both unattributed. Hanzo says it serves large corporations worldwide and is audited by several enterprise clients, none named. The site navigation links a case studies page.
Two kinds of evidence are published, a commissioned study with a described method and a named customer with figures, and neither joins both. Smarsh publishes a Forrester Consulting Total Economic Impact study (September 2024) that interviewed six representatives of financial services firms, built a composite organization and adjusted each benefit for risk. It reports a 124 percent ROI over three years, archive costs down 25 percent and electronic discovery time down 65 percent. The interviewees are anonymous and Smarsh chose them. A customer story for Securities America, a named broker dealer, reports messages flagged for review falling from 19 to 10 percent (8,000 a day) in three months. That 2020 result came from professional services tuning of policy rules, not the current AI agents, and no method is given.
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.
Hanzo publishes real commitments on separation, provider exposure and deletion, but no training statement and nothing on privilege as such. On model providers, the FAQ and About page say Spotlight AI runs on local language models deployed at tenant level, so data never leaves the customer's environment, which is the strongest possible answer on provider exposure if accurate. On separation, the Illuminate page describes controlled access for users, teams and third parties. On retention, the IBM Marketplace MSA transfers client content on request within thirty days of termination and then deletes it. On confidentiality and legal process, MSA section 4 gives prior notice of compelled disclosure. There is no statement on training: the MSA licenses client content solely to provide the service without naming training, and the agreement for direct purchases is unpublished. Privilege and work product are not addressed.
Confidentiality is contractual, and the AI data position is not addressed. The Services Agreement treats the client's data as the client's property and confidential information, with notice before compelled disclosure, and its use license is limited to support and improvement. No published term covers the AI side: which model providers, if any, see client data, or how privilege and work product are handled in data held for discovery.
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.
No advice line or supervision statement is published. The product is sold to corporate legal and compliance teams for preservation, collection and review. The IBM Marketplace MSA makes the client responsible for deciding whether the services are accurate or sufficient, and nothing says how the AI fits a supervising lawyer's duties or names a limit by jurisdiction. The site's own terms page says it is non binding and informational.
The product is positioned clearly as a compliance tool, not legal advice, and supervision is left out. The Services Agreement states that Smarsh does not guarantee that use of the services, or its advice or consulting, will ensure the client's legal compliance. The legal documents page says Smarsh materials are not legal advice and customers must consult an attorney. The audience is compliance, legal and records professionals in regulated firms. Nothing addresses how supervisors or counsel should treat alerts the AI suppressed, or AI summaries, in meeting their own obligations.
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.
Hanzo publishes transparency principles without a governance framework, testing regime or accountable owner. It states three core philosophies for Spotlight AI, client data security, decision transparency and practicality, and its engineering post describes how it writes prompts and designs facets. Hanzo's FAQ says the company follows NIST standards and the ISO 27001 framework, which cover security, not AI governance. No responsible AI framework, ISO 42001 or equivalent, test results before release or statement about uneven output is published.
Transparency features are described, and no governance framework is. The innovations FAQ says AI runs within preserved, governed environments with full audit trails, traceable source data, chain of custody and access by role, so outputs are transparent and reproducible. Press coverage quotes the vendor on auditability aligned with regulator expectations. No accountable owner, testing regime before release or disclosure of uneven performance across languages or communication styles is published, although the misconduct detection agent is pitched on reading slang, jargon and multilingual exchanges. Smarsh also sells AI governance products to financial firms; those are products, not a policy for its own AI.
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.
Published policy covers most of the ground, within the agreement's scope. On retention, IBM Marketplace MSA section 7.3 gives thirty days after termination to transfer client content in the format set by the offer, then deletion unless legally prohibited. On access, MSA section 5.4 commits to commercially reasonable security against unauthorized access, the Illuminate page describes controlled access, and Hanzo's FAQ states annual SOC 2 Type 2 audits with a twelve month evidence period. On incidents, section 5.4 commits to prompt notice of any breach that may expose unencrypted client confidential information, saying what is at risk, with immediate steps to stop continued access. No subprocessor list is published; Hanzo says the AI runs inside the customer's tenant. The MSA governs IBM Marketplace purchases, and the agreement for direct purchases is unpublished.
The client sets retention, capture data is held only briefly before deletion, client data is deleted after termination, and subprocessors for mobile capture are listed by name and location. The trust page states encryption in transit and at rest, SSO and MFA, and regular penetration testing. The agreement and the trust page give no incident notification timeline. An Information Security Addendum is available on request.
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.
What Hanzo stands behind is published and specific, and includes insurance. In the IBM Marketplace MSA of June 2023, section 9.1 gives a defense and indemnity for third party intellectual property claims, with listed exclusions and remedies to procure, replace, or terminate and refund. Section 6.1 caps each party's liability at fees received in the preceding twelve months and excludes consequential loss, carving out gross negligence, willful misconduct, breach of confidentiality and the IP indemnity. Section 5.2 warrants material conformity with documentation, with remedial services, and section 5.5 makes the client responsible for accuracy and sufficiency. Section 10.2 commits Hanzo to carry five million dollars of commercial general liability and five million dollars of professional liability insurance for the term and two years after, with a certificate on request. The agreement covers purchases through the IBM Marketplace. The agreement for direct purchases is not published, and the site's terms page says it is not binding.
A detailed liability position is published with an indemnity and a cap, and AI output is not covered. The Services Agreement gives a vendor indemnity against claims that use of the services infringes a U.S. patent, trademark or copyright, and caps Smarsh's liability at fees received in the prior twelve months. Service level credits are the remedy for availability failures. No term addresses AI summaries or suppressed alerts, and no insurance position is 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.
The connectors are the product, and their depth is described. Hanzo describes preserving Slack in place, with holds by custodian, channel and timeframe and live syncing of new messages, and capturing Teams and Google Workspace data with threading, reactions, edits and deletions intact. The site navigation has a Data Connectors page. The 2023 launch says Spotlight AI uses IBM's watsonx.ai studio. What a customer must configure, and any connection to a review platform or document management system, are not described.
The Services Agreement and pages describe capture from email, mobile carriers such as Verizon and AT&T, apps such as WhatsApp and Signal, and Microsoft 365 Copilot, among other channels. The platform offers Audit, Identity and Review Alert APIs and exports to outside counsel tools. Product documentation sits in the Smarsh Central support portal.
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 of the AI layer is stated clearly; the platform's residency is not. The FAQ and About page say Spotlight AI runs on local language models deployed at tenant level inside the customer's environment, so data does not leave the perimeter. The platform itself is hosted in the cloud, with no region, hosting provider or residency choice stated.
Hosting model and region are stated per service in the agreement; the AI processing location is not. Professional Archive and Web Archive run in an environment Smarsh manages in the United States, Cloud Capture runs in a multitenant AWS environment in the United States, and mobile capture stores data in the United States unless agreed otherwise. Where the AI agents process data, and whether other regions are offered for them, is not stated in the documents read.
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.
Certification is real, dated and audited by a CPA firm, but the report is available only on request. Hanzo's FAQ says SOC 2 Type 2 certification was achieved in October 2019 and is renewed annually with a twelve month evidence period, audited by a CPA firm to AICPA standards, alongside NIST standards, the ISO 27001 framework and audits by enterprise clients. IBM Marketplace MSA section 5.4 commits to provide the latest audit report on written request as confidential information. The auditor is not named and no trust center is published.
The trust page shows ISO 27001 certification under ANAB accreditation and SOC audits by third parties. The Services Agreement commits Smarsh to annual independent audits under ISO 27001 or SSAE 18 and to give clients its latest ISO 27001 and SSAE 18 reports and a penetration test summary as standard audit documentation. The trust page and the agreement do not state the certification scope, auditor or report periods. Smarsh also runs a trust portal at trust.smarsh.com, hosted on Vanta.
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 supply chain is partly disclosed. Hanzo's FAQ says Spotlight AI runs on local language models deployed at tenant level inside the customer's environment, which says where inference runs and that no outside model provider sees content. The 2023 launch says the solution uses IBM's data science and machine learning studio, now part of watsonx.ai, which names the platform provider. The engineering post describes generative models writing case facets. No specific model is named, and there is no commitment to notify customers of changes.
The vendor describes AI models ready for production at global institutions, press coverage quotes it on large language models adapted to the domain and built by its own team, and its APIs support customers' own models. No model, provider or version is named in the agreement or on the product pages, and nothing commits to notice when the models behind surveillance decisions change.
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 agreement states the unit and structure but no figure. The IBM Marketplace MSA prices by data ingestion, storage or other metrics set in the offer, with overage charges above volume limits, on annual subscriptions that renew automatically for the shorter of the prior term or one year unless thirty days' notice is given. Fees can rise each year by the greater of three percent or CPI, and billing runs through IBM. No figure, tier or pricing page is published on Hanzo's site.
The unit and structure of pricing are published, and the figures are not. The Services Agreement licenses its archives by connection and by site, with a minimum commitment, overage fees, a capped renewal uplift and extra fees for long retention, all recorded in the pricing structure. The Intelligent Agent was announced as a separately priced addition. No price or rate is published, and buying runs through sales.
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 coverage are described in substance, with some limits stated. The buyer is corporate legal and compliance, with industry pages for government, healthcare and corporations and use cases for legal hold, investigations, website preservation, ediscovery, early case assessment, DSAR and PCAOB evidence. Data coverage is Slack, Teams, Google Workspace, email, websites and social media. The product's scope, dynamic and collaborative data rather than all of ediscovery, is a stated limit; no jurisdiction or matter type is named as unsupported.
The vendor serves financial services firms from broker dealers and RIAs to banks, along with public sector bodies handling FOIA requests, energy and utilities companies under FERC and NERC oversight, and life sciences. Small and midsized firms and enterprises have separate offerings, and SEC Rule 17a-4 and FINRA supervision are named among the rules supported. The agreement sets use limits, such as capturing only employees' communications. Nothing states which languages, channels or misconduct types the AI agents handle less well.
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?
The published agreement does not mention training either way. IBM Marketplace MSA section 8.2 licenses client content to Hanzo solely as needed to provide the subscription service, and section 8.1 permits collecting and using aggregated, anonymized analytics that could not identify the client. A license limited to providing the service does not permit training, but no clause names model training. The agreement for direct purchases is not published and was not available to read.
Hanzo's FAQ and About page say Spotlight AI runs on local models inside the customer's tenant, which covers where data goes, not whether it trains models.
The published Services Agreement grants a use right over client data bounded to support and improvement of the services, and never names training. Section 4.2 licenses Smarsh to access and use client data as necessary to provide support and improve the services on the client's behalf. No published term addresses whether client communications train or adapt the models behind Smarsh's AI agents.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
A specific period is published for purchases through the IBM Marketplace. MSA section 7.3 says that on request within thirty days of termination Hanzo transfers client content in the format and process set by the offer, and after that deletes all client data it holds unless legally prohibited; section 7.3 also requires return or destruction of confidential information. Nothing states a configurable window for AI facets, tags or outputs during the term, and the agreement for direct purchases is unpublished.
The client sets retention. Archived data is kept for periods the client sets, up to seven years by default and longer for a fee, and capture data is held for a temporary period the client configures, up to 30 days, before deletion. Deletion after termination is provided for, though one section says as soon as practicable and another allows up to six months. Retention of AI summaries and prompts is not addressed separately.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Hanzo claims separation without documenting a permission model. The Illuminate page states controlled access for users, teams and third parties such as outside counsel, and Hanzo's FAQ says Spotlight AI is deployed at tenant level inside the customer's environment, which separates one customer from another. Nothing describes walls between matters within a tenant or how the relevance engine respects them.
Access by role is described, and separation of matters or investigations is not documented. The agreement lets clients set user roles with different access levels, and the innovations FAQ applies role controls around AI outputs. How access is walled between investigations or legal holds is not documented in the materials read.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
The published agreement commits to notice for purchases through the IBM Marketplace. MSA section 4.2 allows disclosure of confidential information required by law, legal process or regulation, provided the receiving party gives the disclosing party reasonable prior written notice so it can contest the disclosure, and limits disclosure to what is required; client content is confidential information under section 4.1. The agreement for direct purchases is unpublished, and the site's terms page is non binding. No transparency report is published.
The Services Agreement commits to reasonable notice before compelled disclosure of confidential information, including client data, where feasible and legally permitted, so the client can contest the order, and to cooperate at the client's expense.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
No legal corpus is named behind the product's output, and the product is not built on one: Spotlight AI assesses relevance across the customer's own collected collaboration data against facets generated from the case description, citing no law.
The AI works over the client's own captured communications, and no external legal source is described. The innovations and trust pages and the Services Agreement were searched on 18 September 2026.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
Nothing published addresses whether authority is checked for subsequent history, and the product does not retrieve or cite primary law; its output is preserved collaboration data and relevance assessments.
The product does not cite legal authority, so no check of later history arises, and none is described.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Hanzo documents how a reviewer can see why an item was flagged, rather than what the model does when it cannot decide. Its FAQ says results are ranked and tagged with reasons, the Spotlight AI page describes transparent explanations for every decision, and the engineering post describes facets a user can clarify with examples for edge cases. No abstention path or confidence threshold for uncertain relevance calls is described.
No abstention path, confidence score or grounding indicator is described for AI summaries or risk signals. The agents suppress alerts judged of low relevance rather than flagging uncertain ones. Searches on 18 September 2026 covered the 2026 press releases, the innovations pages, the trust page and the agreement.
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 court order, opinion or disciplinary record naming Hanzo, Illuminate, Chronicle or Spotlight AI, and no court finding naming them appears in the public record. A preservation and review tool that cites no authority has little exposure here.
The AI Hallucination Cases database maintained by Damien Charlotin and trade press reporting were searched on 18 September 2026. Neither held a court record addressing fabricated or hallucinated content in output from Smarsh products. Litigation history of other kinds is not covered here.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
No public material names an ethics opinion, bar rule or professional responsibility framework. Hanzo's material covers the defensibility of preservation and review and PCAOB audit evidence, which are procedural standards, not bar guidance on lawyers' use of AI.
The vendor cites securities recordkeeping and supervision rules, which are regulatory requirements, not bar guidance. No bar ethics opinion or court rule on AI is named, as of 18 September 2026.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
The buyer is an in house corporate legal or compliance team that bills no client, so the product sits outside a fee relationship between lawyer and client. The published savings are operational: weeks of first pass review cut to hours, and smaller collections. Nothing addresses how AI assisted review is recorded or disclosed on a bill, and law firms are not a named buyer segment.
The product is bought by regulated firms for their own compliance and investigations, where no client is billed for the work. Savings are claimed for the buyer's own costs, including lower outside counsel spend and investigation costs.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
No subprocessor or model provider list is published, and Hanzo's answer on model providers is that there is none: its FAQ says Spotlight AI runs on local language models deployed inside the customer's tenant, so data never leaves the customer's environment, a statement a firm could forward to a client. The 2023 launch names IBM's watsonx.ai as the studio the solution uses. There is no published register of processors and no DPA on the site, and the agreement for direct purchases is not published.
Only mobile capture subprocessors are listed in the agreement, among them TeleMessage, Microsoft Azure and AWS. No AI or model provider is named, and the DPA is available on request rather than published. The Services Agreement, trust page and legal documents index were searched on 18 September 2026.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Some parts of a disclosure record exist, but no export of an AI use record is described. The Spotlight AI page says every relevance decision carries a transparent explanation so it is defensible, and the platform captures collaboration data in a format Hanzo says holds up to legal scrutiny, with edits and deletions intact. Nothing says a record of the model used, the facets applied and the human check can be exported for a court.
Parts of a defensible record exist, short of an AI disclosure record. The archive supports legal holds and exports, and the vendor describes chain of custody and reproducible AI outputs. Nothing records which model produced a summary or risk signal, or who reviewed it.
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
- Bar Guidance Alignment
- Outside Counsel Guideline Readiness
In summary
Hanzo
Hanzo is an ediscovery and preservation platform for dynamic and collaborative data, sold to corporate legal and compliance teams by Hanzo Archives, Inc., with its North American headquarters in Portland, Oregon. Chronicle preserves websites and social media, and Illuminate handles legal hold, collection and review across Slack, Microsoft Teams, Google Workspace and email. Spotlight AI, an optional layer, ranks and tags content by relevance with an explanation for each flag and runs on local language models inside the customer's tenant. According to the AI Legal Index, Hanzo's IBM Marketplace agreement adds an intellectual property indemnity and $5 million each of general and professional liability insurance.
Smarsh
Smarsh Inc. of Portland, Oregon sells communications capture, archiving, surveillance and discovery to regulated organizations. Its buyers are mainly financial services firms, and also public sector bodies, energy and utilities and life sciences companies. It retains email, messaging, collaboration, mobile, voice and social content for recordkeeping rules such as SEC Rule 17a-4. Its AI agents filter surveillance alerts, detect misconduct in context across languages, and summarize and map evidence for investigations. According to the AI Legal Index, its Services Agreement licenses per connection, lets clients set retention and states that the service does not guarantee legal compliance. Its data processing addendum is available on request.
Questions buyers ask
Hanzo vs Smarsh: which is better for Slack and Teams discovery?
They are close, and they are built for different jobs. Hanzo is built for preserving and reviewing collaboration data in place for legal holds, with AI that runs inside the customer's tenant. Smarsh is a broader archive for regulated firms, capturing every channel for recordkeeping and supervision, with AI centered on surveillance. The choice turns on whether the need is litigation holds or regulatory supervision. From the AI Legal Index, based on each vendor's own published materials as of September 25, 2026. No vendor pays for placement.
Where does Hanzo's Spotlight AI run?
Hanzo states that Spotlight AI runs on local language models deployed at tenant level inside the customer's environment, so data never leaves it, and its 2023 launch said the solution leverages IBM's data science studio, now part of watsonx.ai. From the AI Legal Index, based on each vendor's own published materials as of September 25, 2026. No vendor pays for placement.
What contract terms come with Smarsh?
Smarsh's Services Agreement licenses Professional Archive per connection, lets clients set retention periods up to seven years by default, commits to notice before compelled disclosure, gives an intellectual property indemnity, caps liability at the prior twelve months of fees, and states that the services do not guarantee legal compliance. It lists subprocessors only for mobile capture, and its data processing addendum is available on request. From the AI Legal Index, based on each vendor's own published materials as of September 25, 2026. No vendor pays for placement.
How accurate is Hanzo's AI?
Hanzo's November 2023 launch release states 95 percent recall in identifying documents pertinent to a case, with no test set, method or later measurement published. Its product pages state that results are ranked and tagged with the reason each item was flagged, and users can refine the case facets that drive relevancy with examples. Smarsh publishes figures such as 60 percent fewer false positives from its agents, also without a method. From the AI Legal Index, based on each vendor's own published materials as of September 25, 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.
Two points to weigh. Hanzo's published master services agreement governs purchases through the IBM Marketplace. The agreement for direct purchases is not published, so its liability, notice and deletion terms are confirmed for that channel only. Its 95 percent recall figure dates from the 2023 launch and has no published test set or later measurement. Smarsh's Forrester study was commissioned by Smarsh, its interviewees were chosen by Smarsh and are anonymous, and its figures describe the platform rather than the newer AI agents. Neither vendor reviewed this page.