AI Integration

DANTE

Drafting & Automation Natural-language Text Engine

A bespoke AI tool built on open-weights models for risk flagging and process improvement. DANTE helps organisations make practical decisions about when to seek external advice—improving outcomes and educating staff on correct processes before issues arise.

Book a Demo

Key Features

Architecture

Self-hosted, GDPR Compliant
Foundation

Qwen3.7 Max
Current Version

Lucan v2.4
Deployment

On-premise / Private Cloud

The Problem

Many organisations leave the seeking of advice until it is too late. By the time issues are identified, they have often escalated beyond easy resolution—leading to increased costs, operational disruption, and missed opportunities.

Nothing can replace good external advice, but DANTE provides a process and risk flagging layer that allows businesses to make practical decisions about when to seek that advice. Early identification of potential issues improves outcomes and reduces total cost of external engagements.

Cost of delayed identification

Early identification through automated monitoring typically reduces total resolution costs by 60-80%. Issues caught at the "escalation" stage cost 3-5x more to resolve than those flagged immediately.

Capabilities

What is DANTE

Analysis

Contract Review Pipeline

Automated scanning of incoming contracts to identify non-standard terms, missing clauses, and potential risk areas before human review begins.

85%

Time Saved on First Review

Monitoring

Compliance Flagging

Ongoing monitoring of communications and processes against defined compliance frameworks, flagging potential issues before they become problems.

Real-time

Risk Detection

Process

Onboarding Review

Structured review of employee and vendor onboarding processes to ensure correct procedures are followed and documentation is complete.

100%

Documentation Coverage

Output

Briefing Documents

Automated generation of formal briefing documents for internal stakeholders or external advisers, ensuring consistent and complete information transfer.

Formal

Documentation

Self-Hosted AI

Not a wrapper

Most "AI tools" for business are thin wrappers around commercial APIs—OpenAI, Anthropic, Google. They add a custom interface, but your data still flows through third-party servers, subject to their terms, their data handling practices, and their jurisdictional exposure.

DANTE is different. It runs entirely on infrastructure you control, using open-weights models that require no external API calls. Your documents, communications, and analysis results never leave your network perimeter.

The distinction matters because:

Data residency — Many clients require that privileged, confidential or personally identifiable material never be processed by third parties. API wrappers cannot satisfy this.
No usage-based billing — Commercial APIs charge per token. At scale, this becomes significant. Self-hosted means fixed infrastructure costs.
Model control — You choose when to update, which model version to run, and can maintain consistency over time.
Audit trail — Complete visibility into what was processed, when, and by whom. No more mysterious black-box in the cloud - critical for compliance.

API Wrapper

Your data → Third-party API → Their servers → Response
Data leaves your control

DANTE

Your data → Your infrastructure → Your model → Response
Data never leaves your network

Economics

Cost comparison at scale

Based on a 10-person team processing at standard token/day rate.

*Hardware cost amortised over 36 months plus power (~$40-60/mo). Rates as of June 2026.

The Moving Target

Falling token prices, exploding usage

Per-token inference costs have fallen roughly 10× per year from 2021 to 2025. GPT-4-equivalent capability that cost ~$20-30 per million tokens in late 2022 now costs around $0.40 per million. Gartner projects a further ~90% reduction by 2030.

Usage is exploding

Unit prices are collapsing while total bills rise, because consumption is exploding: reasoning models can burn ~100× more tokens internally than they emit, and agentic pipelines multiply tokens per task. Your "100M tokens/month" is far likelier to become 300M or 1B than to hold flat.

Of course even if price reductions are able to out-scale the explosion in token usage, the price is largely irrelevant to protected workloads, because residency and privacy requirements may bar you from using those endpoints for client data no matter how cheap they get.

Price per token falls ~10× annually, but agentic usage grows faster. Net spend continues rising.

Deep Dive

Contract review workflow

DANTE's contract review module processes incoming agreements through a multi-stage analysis pipeline, comparing each document against your standard playbook and flagging deviations that require human attention.

What gets flagged:

DANTE draws on a comprehensive database of contract clauses and risk indicators, and learns from past reviews to improve its accuracy over time. Key issues it identifies include:

Non-standard liability caps
Unusual indemnification
IP assignment concerns
Termination provisions
Data handling requirements
Missing clauses

flowchart TB
    A["Document Ingestion
PDF, Word, email"]
    B["Structure Analysis
Clause identification"]
    C["Playbook Comparison
Risk scoring"]:::accent
    D["Flag Generation
Priority ranking"]
    E["Output
Report + redlines"]

    A --> B --> C --> D --> E

    classDef accent fill:#7A2E22,color:#FFFFFF,stroke:none

Risk Analysis

How DANTE scores risk

You set your risk tolerance and industry requirements, and DANTE scores each identified issue based on severity, potential impact, and likelihood. This allows your team to prioritise which issues to review first and determine when to escalate to external advisers.

Risk Scoring Factors:

Severity — How critical is the issue?
Impact — What is the potential financial/operational impact?
Likelihood — How likely is this issue to cause a problem?

Risk scores are used to prioritise issues for human review and determine when to escalate to external advisers. Thresholds can be adjusted during implementation and refined over time based on feedback and outcomes.

Typical distribution across contract portfolio. Thresholds configurable per organisation.

Implementation

How we work together

Discovery & Assessment

We analyse your current processes, identify pain points, and define the scope of automation. This includes mapping existing workflows, reviewing sample documents, and identifying where AI integration creates genuine leverage.

Duration: 2-3 weeks Lewis+Behr Led

Playbook Development

We configure DANTE for your specific requirements—defining compliance frameworks, building your contract playbook, establishing risk thresholds, and training the system on your document templates and organisational standards.

Duration: 3-4 weeks Lewis+Behr Led

Infrastructure Setup

DANTE is deployed in your environment—on-premise servers or your private cloud. We configure integrations with your existing systems (document management, email, communication tools) and establish secure data pipelines.

Duration: 1-2 weeks Joint (IT + Lewis+Behr)

Pilot & Validation

We run DANTE against historical documents to validate accuracy, refine thresholds, and demonstrate ROI before full rollout. This phase typically processes 50-100 sample documents to ensure the system meets your requirements.

Duration: 2-3 weeks Joint

Training & Rollout

Your team is trained on using DANTE effectively—understanding outputs, interpreting risk flags, and knowing when to escalate. We provide documentation, conduct hands-on sessions, and support the transition to live operations.

Duration: 1-2 weeks Lewis+Behr Led

Ongoing Optimisation

Continuous refinement based on real-world usage. We remain available for iterative improvement—adjusting thresholds, adding new modules, expanding coverage to new document types, and optimising performance based on feedback.

Ongoing support available Optional Retainer

Typical Implementation: 10-14 weeks

Security

Data protection & compliance

Data Sovereignty

Your data never leaves your infrastructure. DANTE runs entirely within your network perimeter—on-premise servers or your private cloud instance. No data is transmitted to external services or third-party APIs.

GDPR Compliance

Designed for EU data protection requirements from the ground up. Full data subject access request support, audit logging, and configurable data retention policies. No cross-border data transfers required.

Audit Trail

Complete logging of all system activities—document processing, risk assessments, user actions, and configuration changes. Exportable audit reports for compliance reviews and regulatory requirements.

Access Control

Role-based access control with integration into your existing identity provider (Active Directory, Okta, Azure AD). Granular permissions for document access, system configuration, and administrative functions.

Encryption

Data encrypted at rest (AES-256) and in transit (TLS 1.3). Optional integration with your existing key management system. No plaintext storage of sensitive document content.

SOC 2 Ready

Architecture designed to support SOC 2 Type II compliance requirements. Documentation and controls available for inclusion in your organisation's compliance programme.

Looking Ahead

The regulatory trajectory is toward more oversight, not less. The EU AI Act is a floor, not a ceiling — member states may add requirements, and sectoral regulators (financial services, healthcare, legal) are developing their own guidance.

For frontier models specifically, watch for: evolving FLOP thresholds for systemic risk classification; narrowing of the open-source exemption as capabilities increase; and potential requirements for model provenance and supply-chain transparency. Organisations deploying AI should build compliance infrastructure now rather than retrofitting later.

The self-hosting advantage: On-premise deployment gives you direct control over model selection, versioning, and audit trails — critical for demonstrating compliance when regulators ask what model processed what data, and when.

Modules

Available capabilities

Core

Contract Review

Automated first-pass review of incoming contracts against standard templates and risk frameworks. Identifies deviations, missing clauses, and high-risk terms.

Playbook comparison
Risk scoring
Redline generation
Summary reports

Core

Compliance Monitoring

Ongoing scanning of communications and processes for potential compliance issues. Configurable rules engine for industry-specific requirements.

Email scanning
Policy enforcement
Regulatory mapping
Alert workflows

Core

Briefing Generation

Structured output generation for internal stakeholders and external advisers. Ensures consistent, complete information transfer.

Templated outputs
Context aggregation
PDF export
Version tracking

Add-on

Intellectual Property

Review of IP-related documentation, licensing terms, and ownership structures. Track IP assets and identify potential conflicts.

License analysis
Assignment review
IP register integration
Conflict detection

Add-on

Privacy & Data

GDPR and privacy compliance monitoring, DPA review, and data handling assessment. Track data flows and processor relationships.

DPA analysis
Data mapping
Transfer assessments
Processor management

Add-on

Employment

Employment documentation review, onboarding compliance, and policy consistency checks. Support for multi-jurisdiction requirements.

Contract review
Policy checking
Onboarding flows
Termination review

Add-on

Procurement

Vendor agreement review, supplier risk assessment, and procurement compliance. Integration with vendor management systems.

Vendor due diligence
MSA review
SOW analysis
Renewal tracking

Add-on

Real Estate

Lease review, property documentation analysis, and portfolio management support. Track obligations and key dates.

Lease abstraction
Obligation tracking
Renewal alerts
Cost analysis

Custom

Bespoke Modules

Industry-specific or organisation-specific modules developed to your requirements. Built on DANTE's core engine with custom rules and workflows.

Custom playbooks
Industry rules
Workflow integration
Dedicated support

FAQ

Common questions

How accurate is DANTE compared to human review?

DANTE is designed to augment, not replace, human review. For standard contract review, our testing shows 92-97% accuracy in identifying deviations from playbook terms. The system is configured to err on the side of caution—flagging potential issues for human review rather than missing them. False positive rates are typically 8-12% and decrease with continued use as the system learns your preferences.

What happens to my data?

Your data never leaves your infrastructure. DANTE runs entirely on your servers or private cloud—there's no data transmission to external APIs or third-party services. Documents are processed locally, and you maintain complete control over storage, retention, and deletion policies.

How long does implementation take?

Typical implementation takes 10-14 weeks from kickoff to go-live. This includes discovery (2-3 weeks), playbook development (3-4 weeks), infrastructure setup (1-2 weeks), pilot validation (2-3 weeks), and training/rollout (1-2 weeks). Timeline can vary based on scope and integration requirements.

What are the ongoing costs?

After implementation, ongoing costs are primarily infrastructure (compute, storage) and optional support retainer. Infrastructure costs are typically $500-2,000/month depending on scale. There are no per-document or per-user licensing fees. Optional support and enhancement retainers are available for organisations wanting ongoing optimisation and new module development.

Can DANTE work with our existing systems?

Yes. DANTE integrates with common enterprise systems including Microsoft 365, Google Workspace, SharePoint, Slack, Teams, and most document management systems. A REST API is available for custom integrations. We assess integration requirements during the discovery phase and include necessary connectors in the implementation.

Is this suitable for regulated industries?

DANTE is designed with regulated industries in mind. The self-hosted architecture addresses data sovereignty requirements, and the audit trail supports compliance obligations. We have experience deploying in financial services, healthcare, and government contexts. Specific regulatory requirements are addressed during the discovery and playbook development phases.