D

Diliforce

Diliforce adds AI analysis to a governance system that already manages contracts, board meetings, legal entities and due diligence materials. The Assistant summarizes agreements, compares versions, translates clauses and explains potential risks. Metadata extraction proposes tags and key terms during individual or bulk imports, while SmartIndex can build a folder structure from a sample file or photograph. These functions reduce preparation work across the suite, but the repository, meeting cycle, reminders and signature connections operate independently.

Generative features require an additional account subscription and administrator permission. The product remains a corporate governance platform with substantial AI assistance rather than a standalone legal reasoning service.

The Assistant analyzes selected corporate files and can compare two versions through an AI Analysis List. Metadata extraction is promoted as producing accurate tags, and the Assistant offers explanations of complex clauses and potential compliance issues. Its responsible use guidance warns that generated content may be inaccurate, incomplete or outdated and requires independent verification of important information. No measured error rate, reproducible accuracy test or passage citation method is specified.

Selecting the source files gives users material against which to inspect an answer, but does not establish that each assertion links to an exact supporting passage or current legal authority.

Metadata extraction produces suggestions that a user must review and apply for each agreement, including files submitted through mass import. Users can discard suggestions; changing a proposed value requires applying it and then editing the metadata card. The account administrator controls access to AI features and can separately enable or disable optional Claude models. An AI Analysis List lets each user select an accessible file for analysis or two for comparison.

The service terms exclude automated decisions with legal or similarly significant effects under GDPR Article 22. Confidence thresholds, automatic escalation for uncertain answers and correction handling beyond these review controls are unspecified.

RELEX Solutions uses the governance portal, entity register and data rooms to maintain materials for financing and subsidiary administration. Its head of legal, Tuomo Sirkka, describes replacing scattered network folders with a common repository and reusing due diligence materials for successive funding rounds, including the €500 million round in 2022. Legal, finance and administrative staff share the system, and external law firms receive materials through its data rooms.

SIGG and Kiron Interactive also describe contract or entity management use. These accounts establish operational adoption, but do not measure AI accuracy or isolate the Assistant's contribution to time savings or transaction results.

The service agreement prohibits model training on prompts, files uploaded for AI interactions and generated outputs. Mutual confidentiality duties extend to employees and subcontractors, while the DPA restricts personal data processing to customer instructions. Administrators assign category permissions and can use access groups for particular entities or business units. AI Analysis Lists accept only attachments the user may access.

Optional Claude processing requires administrator activation and uses Anthropic infrastructure. Providers may temporarily retain prompts for abuse prevention under their own terms. These protections address confidential corporate materials, without promising preservation of legal privilege or a separate conflicts management process for outside counsel.

Responsible use guidance tells users to inspect AI output, independently verify important information and avoid treating generated responses as legal, financial or other professional advice. That boundary applies even though the Assistant is marketed for governance guidance, board responsibilities, ESG issues and contract risk explanations. Users remain responsible for deciding whether a proposed interpretation or summary is suitable for the underlying business decision.

The terms also prohibit protected health information, GDPR special category data and specified financial information. No jurisdictional professional conduct mapping, lawyer qualification check or detailed allocation of supervision duties between corporate counsel and board administrators is specified.

Diliforce's responsible use instructions acknowledge incomplete, inaccurate and outdated AI answers and require human verification before reliance. Administrators control AI access and optional provider activation, and the terms rule out automated decisions with legal or similarly significant effects under GDPR Article 22. Those are practical use boundaries, but they do not describe a model governance program. A named owner for AI risk, release testing criteria, bias assessment across languages or jurisdictions and published findings from such evaluations are unspecified.

Board oversight and corporate governance features organize the customer's affairs; they do not establish how the vendor evaluates its own models.

Conversation attachments are removed after the interaction at Diliforce's layer, while model providers may retain prompts and responses for abuse prevention. Account and profile information remains for the subscription plus twelve months; usage logs and support communications normally remain for twelve months. Security guidance describes seven day database recovery and file soft deletion. The DPA requires return or destruction after termination, subject to legal retention and scheduled backup deletion, and written breach notice without undue delay.

Access groups, MFA and individual permissions restrict visibility. A full subprocessor list is supplied on request, and provider retention periods are not fixed in the agreement.

ContractZen's agreement expressly excludes liability for damage caused by AI features or their use, including inaccurate, incomplete or misleading output. The service is supplied as is without performance, fitness or noninfringement warranties. Its general aggregate liability ceiling is 50% of fees paid during the six months before receipt of a claim, subject to liabilities that law does not allow it to exclude, including gross negligence and intent.

Customers indemnify the provider against third party claims connected to their use or contractual breach. No reciprocal AI output indemnity or accuracy warranty is offered to cover a mistaken governance recommendation or contract summary.

Electronic signature connections carry agreements from Diliforce to Assently, Adobe Sign, Docusign, SignHero or Visma Sign and bring signed copies back into the workspace. Assently packages are purchased in the in-app store; other providers require an existing account connected there, and Adobe requires its Enterprise plan. Starting from an existing agreement or meeting preserves that item's signature status and returned signed file.

Starting from the dashboard cannot return the result to an existing item. Signers cannot be changed after sending without cancellation and restart. Microsoft 365 and Google provide SSO, while Teams supports board work and a custom plan offers API access.

The governance service runs on Microsoft Azure, with security guidance naming the Netherlands for hosting and EU geography for data at rest. The default Assistant accesses models through Microsoft Foundry, but its privacy guidance permits processing in other Microsoft regions, including outside the EEA, under transfer safeguards. Optional Claude runs on Anthropic's own infrastructure and may also process outside the EEA.

This makes storage location narrower than the permitted processing footprint. Customer files receive a separate encryption key held in Azure Key Vault. A customer hosted deployment, selectable inference region and country list covering every optional service are not specified.

Security measures include AES-256 protection for attachments and databases, TLS connections, customer specific file encryption keys, MFA and restricted user access. The engineering process follows a secure development lifecycle and addresses OWASP risks; security tickets require validation before closure. Azure supplies redundant storage, monitoring and disaster recovery infrastructure. The ISO, SOC and other compliance standards named in the security guidance belong to Microsoft Azure, rather than constituting an independent attestation of ContractZen or Diliforce's application.

No vendor specific certification report, auditor, assessment period or attested application scope is identified. Technical controls and the hosting provider's credentials therefore remain separate assurances.

Microsoft Foundry supplies the default model access layer, with Azure OpenAI identified in the AI privacy guidance. Administrators may enable Anthropic Claude, which runs on Anthropic infrastructure even though Foundry supplies access, identity and billing. Claude is disabled by default and has separate processing terms. The agreement prohibits training on AI inputs and outputs but permits providers to handle them under their own abuse prevention rules.

Exact model versions, the routing between available models and a model specific change notice commitment are not stated. General contract amendment notice does not specify advance notice for each model or provider substitution.

Standard costs $9.50 per user each month when billed annually and includes contract management, the board portal and SmartIndex. Advanced costs $27 per user monthly on annual billing, with at least five users, and adds entity management, access groups, detailed audit logs and included data rooms. AI is a separate account subscription starting at $15 monthly, with administrator allocation of access. Signature providers charge separately, and larger custom plans require a quote.

The terms normally renew yearly unless canceled fourteen days before expiry. Migration, training and assistance can be separately agreed, so the full implementation cost is not fixed by the standard subscription rates.

The suite serves corporate legal teams, company secretaries, directors and finance staff maintaining governance information. Its main workflows cover board preparation and minutes, legal entities and ownership, agreement administration, funding rounds and audit preparation. Standard targets smaller organizations, Advanced adds controls for midsize companies, and custom arrangements support larger groups. Named customers include RELEX Solutions, SIGG and Kiron Interactive, with use across more than thirty-five countries.

Multilingual analysis does not establish coverage of a defined body of national law. The service terms also exclude health information, GDPR special category data and specified financial account information, which limits the material these teams may upload.