Licence utilisationAssigned seats and active usage
Applied research
InnosuisseInnovation Cheque support
AI activity · Firm viewIllustrative data
Connected view
AI spendCHF 0Selected period
Requests0Across connected activity
Licence utilisation0%72 active · 100 assigned
Unattributed0%Activity not yet linked
Grouped by matter
NameRequestsTask timeEst. costAttribution
MatterMatter 24-018Corporate
Requests5,820
Task time126h 40m
Est. costCHF 16,840
Attributed100%
Legal AI Operations
One operating layer for the AI your teams already use.
Across supported AI environments, Hebed brings together AI traffic, usage, cost and work context, with pseudonymisation where sensitive information requires it.
AI tools
The tools remain in place.
Copilots, legal AI products, models and agents continue to support research, drafting, review and workflows.
Hebed
Hebed adds the operating layer.
Hebed adds AI traffic, usage, cost and work context without replacing the tools themselves.
Four dimensionsThe operating model behind Legal AI Operations
PseudonymisationPseudonymise sensitive information before it is sent to an external AI model.
Scanning
Original prompt
Draft a note for Ms. Keller at Rue du Pont 5 regarding Matter 24-018.
Detect → pseudonymise
Pseudonymised prompt
Draft a note for [CLIENT] at [ADDRESS] regarding [MATTER].
Sensitive informationDetected before transmission
Personal dataName and address pseudonymised
Matter contextMatter reference pseudonymised
External LLMReady to send
AI TrafficMonitor active AI tools and the traffic moving through them.
How AI works, where it fails and why judgement matters.
A teaching session at HEG Genève followed by a seminar discussion on AI agents.
HEG Genève sessionSeminar setting and discussion
HEBED at HEG Genève: from the history of AI and the emergence of LLMs to hallucinations, sensitive information, safeguards and the role of critical judgement. The session explored the evolution of AI, how large language models emerged, how they produce outputs and the risks that sit underneath apparently confident answers. For students, the central question was not simply how to use AI, but how to keep enough critical distance to assess what it produces.AI agents seminar
Where AI agents add value, and where judgement remains essential.
The seminar used a created AI agent to show its practical value inside an organisation, including speed, support and workflow continuity, while also making the risks visible. The discussion covered access to information, delegated actions, errors that can propagate across steps, human oversight and the safeguards required when systems move beyond a single prompt and response.
Bringing product questions into responsible AI conversations.
At WSIS Forum 2026 in Geneva, HEBED presented the product and discussed how AI visibility, responsible adoption and operational safeguards translate into practical questions.
WSIS Forum 2026Product conversation in Geneva
HEBED presented the product at WSIS Forum 2026 in Geneva, using the setting to discuss AI visibility, responsible adoption and safeguards in practice.Field engagement
Responsible AI should stand up to practical scrutiny.
The discussion focused on practical AI adoption: what organisations can observe, where safeguards are needed and how controls can remain clear as tools and models evolve.
Academy
Bring Academy to your legal team.
Discuss a seminar, a Shadow AI workshop or a prompt-pseudonymisation session around the questions your team is already facing.
During his Master’s research at HES·SO, Heben Tekleab examined Shadow AI and how external LLM use can grow before an organisation has a clear view of that activity.
As AI reshapes legal workflows, usage, cost and sensitive information increasingly span multiple tools and processes. Visibility across that activity has not kept pace.
Product
The research became an operating model.
Pseudonymisation, AI traffic, usage and cost, and work context became the four dimensions of Legal AI Operations.
The work is developed in Geneva through applied research with HEG Genève, with support from an Innosuisse Innovation Cheque.
Hebed
Legal AI Operations, in practice.
Hebed operates across the AI tools legal teams already use, connecting pseudonymisation, AI traffic, usage, cost and work context.
Hebed AI is an early-stage MVP prototype provided for testing and demonstration purposes only. We are not a registered or licensed services provider in any jurisdiction.
Last updated: November 23, 2025
Information We Collect
We collect only minimal information necessary for profile creation, including:
Name
Email address
Basic profile information you voluntarily provide (e.g., organization details or team role)
Non-sensitive usage data (device type, browser type, analytics)
We do not collect: bank account details, payment information, or other sensitive data.
How We Use Your Information
We use your information solely to enable your account, display your profile (if you choose), and provide AI governance services on the Hebed AI platform. We do not execute any financial transactions or broker deals of any kind.
Data Sharing & Security
Your information is not sold or rented. We may share minimal required data with third-party service providers (hosting, analytics) who assist us in operating the platform.
As this is a prototype, features may change and data may be altered or removed at any time. Please avoid uploading confidential or sensitive data.
Your Rights
You may request access, correction, or deletion of your information by contacting us at contact@hebedai-office.com
By using Hebed AI, you acknowledge the prototype nature of the platform and accept that the service may change or discontinue at any time.
Hebed AI is an early MVP prototype provided for testing and demonstration purposes only.
Last updated: November 23, 2025
Important Disclaimers
We are not a registered or licensed entity in the UAE.
We do not provide financial, investment, fundraising, brokerage, or advisory services of any kind.
No financial transactions or regulated activities occur on this platform.
All content is experimental and should not be relied upon for any business decision.
By using the platform, you acknowledge that all features are pre-release and may be changed or removed at any time.
Platform Purpose
Hebed AI is designed as a demonstration platform for AI governance and compliance monitoring. Any decisions or activities that may result from using the platform occur entirely outside of our platform and are the sole responsibility of the parties involved.
User Acknowledgment
By accessing and using Hebed AI, you acknowledge and agree that:
This is a prototype platform for demonstration purposes only
Features may change, be removed, or become unavailable without notice
The platform may be discontinued at any time
You will not rely on any information provided for business decisions
You will conduct your own due diligence for any activities
Hebed AI uses a minimal set of cookies and similar technologies as part of its early-stage MVP prototype. This Cookies Policy explains what cookies are, how we use them, and the limited choices you have regarding their use.
Last updated: November 23, 2025
What Are Cookies?
Cookies are small text files stored on your device when you access a website. They help websites function properly, remember basic preferences, and collect non-sensitive usage analytics. Since Hebed AI is a prototype, our cookie usage is intentionally limited and may change as the platform evolves.
Types of Cookies We Use
1. Essential Cookies
These cookies are required for the basic operation of the platform, such as enabling page loading, ensuring security, and allowing you to log in or create a simple profile. These cannot be disabled as the platform may not work without them.
2. Functional Cookies (Minimal Use)
These may store lightweight user preferences (e.g., UI settings) to improve your prototype experience.
3. Analytics Cookies (Non-Sensitive)
We may use basic analytics tools to understand general usage patterns such as:
Device type
Browser type
Pages visited
Session duration
These analytics do not track sensitive data, financial information, investment activity, or personal behaviour across other websites.
Cookies We Do Not Use
Hebed AI does not use:
Advertising cookies
Behavioral tracking cookies
Cookies collecting payment, financial, or sensitive investment information
We also do not engage in cross-site tracking.
How We Use Cookie-Related Information
Cookie data is used solely to:
Ensure the platform functions as intended
Improve user experience during testing
Collect basic prototype analytics
Because Hebed AI is an MVP prototype, cookie behavior and technology may change or be removed at any time.
Managing Cookies
You can manage or delete cookies through your browser settings. However, disabling essential cookies may prevent the platform from operating correctly.
Prototype Disclaimer
By using Hebed AI, you acknowledge that:
The platform is an early MVP for testing and demonstration only
Features and cookie usage may change without notice
Data stored in cookies may be temporary or removed at any stage of development
Applied research
OpenAI
Claude
Mistral
Harvey
Legora
Copilot
Deepjudge
Omnilex
Perplexity
PPrivate models
