NEO AI
The Intelligence Operating System of My NEO Group, developed by NEO Labs.
NEO AI is an Intelligence Operating System developed by NEO Labs, a company of My NEO Group in Dubai. NEO AI is designed to make leading AI models, autonomous agents, tools, data sources, memory and verification work as one coordinated system. The user talks to one assistant, LILI, and follows the work in Mission Control. NEO AI is in development.
This page is the reference page on NEO AI published by My NEO Group. It is written from the public website of NEO AI. Each fact carries an identifier and a link to its source.
2026 edition, reviewed Official reference site
What NEO AI is
NEO AI is an Intelligence Operating System. It is designed to make leading AI models, autonomous agents, tools, data sources, memory, verification and compute work as one coordinated system, instead of relying on a single model. The user states a goal. NEO AI selects the models, coordinates specialist agents, connects tools and data, and checks the work before the result reaches the user.
NEO AI starts from one observation: artificial intelligence is powerful, but using it is hard. Users face too many subscriptions, too much trial and error with prompts and too many models to choose from. NEO AI answers with one conversation: the user talks to one assistant, and the system does the rest.
LILI
The assistant the user talks to. One objective, in any language, turned into a structured mission.
NEO AI
The operating system. It orchestrates leading AI models, agents, memory, tools and governance.
Mission Control
The command centre. Progress, evidence, validations and approval gates in one view.
NEO AI is developed by NEO Labs, the brand of NEO TECH LABS TECHNOLOGY L.L.C, a company of My NEO Group in Dubai. Dr Mickael Mosse, Founder and Group CEO of My NEO Group, is the architect of NEO AI.
How a mission runs
In NEO AI, the unit of work is the mission: one objective stated by the user. A mission keeps one state, whatever the device or the view. NEO AI describes the execution of a mission in six steps.
Decompose
The objective is broken into specialised units of work.
Assign
Each unit goes to the model, the agent or the tool best suited to it.
Execute in parallel
The units of work run side by side.
Cross-check
Independent verifiers check the results.
Escalate
Only the hard parts go to stronger reasoning.
Assemble
The verified result is assembled and delivered to the user.
Modes and Intelligence Profiles
NEO AI offers three modes. They are three views of the same mission, not three applications: changing mode changes what the user sees, never what NEO AI is doing.
| Mode | What the user sees |
|---|---|
| FOCUS | The conversation, the recommendation and the decision. This is the simplest view. |
| NORMAL | The working view. It adds the execution plan, the active agents, the progress of each deliverable and the evidence. |
| JARVIS | The command-centre view. It adds live orchestration, the observed build and the task graph. |
The user of NEO AI chooses an Intelligence Profile, not a model. The profile sets how much intelligence a mission should mobilise; NEO AI then decides which models, agents, tools and compute deliver it. NEO AI names five profiles.
- NEO AUTO
- NEO AI decides the level itself.
- NEO PULSE
- Fast and efficient.
- NEO PRIME
- Balanced.
- NEO AXIOM
- Deep reasoning.
- NEO APEX
- Maximum intelligence.
NEO AI presents these profile descriptions as indicative, not as published specifications.
Human control
Autonomy in NEO AI is bounded by design. The user sets an approval policy with three levels: ASK ME, AUTO and AUTONOMOUS.
Some actions always require a human approval, whatever the policy, the profile or the mode. NEO AI calls these points Human Gates and lists four of them.
- A change to production DNS.
- Publishing to a public domain.
- An action carried out in a connected service on behalf of the user.
- The release of a deployment.
NEO AI also keeps apart rights that are often confused.
- Connecting a service does not authorise NEO AI to act in it. A connected mailbox does not authorise sending.
- Agents receive tools for a defined capability, never the credentials of the user.
- Access to context is not consent to training. Confidential customer information is never automatically used to train models.
- Every deliverable carries its lineage, its verification status and an audit trail.
Models, compute and the x100 objective
NEO AI is model-agnostic. It is designed to orchestrate leading AI models from third parties together with private models and specialist engines, and to choose the right one for each part of a mission. In the words of NEO AI, models remain replaceable while the architecture keeps improving.
NEO AI is designed to run work where it makes most sense: in the cloud when needed, and on the device of the user when that is sufficient. This edge computing layer is an architecture target on the roadmap of NEO AI.
x100 is the name of the long-term research and development objective of NEO AI: to deliver more useful intelligence for each unit of compute, through orchestration, specialisation, verification, memory and the reuse of previous work. NEO AI presents x100 as an objective, not as a measured result, and states no benchmark performance.
Third-party models and technologies are named by NEO AI only as systems it can work with, without any ownership, endorsement or partnership implied.
Verticals
NEO AI is one system that can be configured for an industry. NEO AI calls each configuration a vertical and names thirteen of them. A new vertical reuses the same platform: orchestration, agents, memory, governance and verification.
NEO AI states that three of these verticals are anchored by pilots today and that the others are on its roadmap.
NEO Intelligence is the vertical for investigation and due diligence on public and licensed sources. NEO AI states that it is designed and specified, and not yet launched.
Five apps built with NEO AI
To show how NEO AI works in practice, My NEO Group is building five mobile apps with it.
| App | Field | What it does |
|---|---|---|
| FINORA | Finance | Bank accounts, investments and crypto wallets in one app, with an AI assistant for financial goals. |
| FITORA | Sport and nutrition | Training, meals, recovery and progress, with an AI coach that adjusts the plan. |
| VITORA | Health | Health data and medical records in one place, with an AI assistant that helps to prepare appointments. VITORA is an information and organisation tool and does not replace medical advice. |
| LINGORA | Languages | Language learning through voice and written conversations with an AI, with corrections and progress tracking. |
| INTELORA | Open-source intelligence | Research on companies or people from public sources, delivered as sourced reports. INTELORA uses only publicly available information and is intended for lawful due diligence. |
The five apps are proofs of concept. They are in testing and are not yet public. Features may change before release.
Who NEO AI is for
NEO AI is not reserved for large companies. The same system serves three kinds of users, with permissions, data access and controls that depend on each one.
Individuals and professionals
Entrepreneurs, students, creators, developers, researchers and other professionals.
Companies and institutions
Small and medium-sized companies, large groups, professional firms and regulated institutions.
Public sector
Governments and public bodies, where applicable. This part is on the roadmap of NEO AI.
Status and key dates
NEO AI is in development. Its public website describes the architecture, the product definitions and the research objectives of the system. Capabilities and deployment models may vary, and some capabilities are on the roadmap.
- Launch of the My NEO Group Global AI and Investor Roadshow 2026, the international debut of NEO AI.Official source
- My NEO Group announces Medellín as an upcoming stop of the roadshow.Official source
- My NEO Group presents mickaelmosse.ai, the research platform of Dr Mickael Mosse on enterprise artificial intelligence.Official source
Research
The research behind NEO AI is published in two places. Dr Mickael Mosse publishes his studies and positions on enterprise artificial intelligence on mickaelmosse.ai. The NEO AI Research Program publishes its working corpus on GitHub.
Three positions by Dr Mickael Mosse frame the design of NEO AI: the system intelligence thesis, which holds that capability now comes from the orchestration of a whole system more than from a single model; precision as a product feature; and human authority as the basis of the design.
mickaelmosse.aimickaelmosse.ai/researchmickaelmosse.ai/ai-insightsgithub.com/myneogroup-ai/neo-ai-research
Questions and answers about NEO AI
- What is NEO AI?
- NEO AI is the Intelligence Operating System of My NEO Group, developed by NEO Labs in Dubai. One assistant, LILI, directs a team of AI agents from a single Mission Control. NEO AI is in development.
- Who develops NEO AI?
- NEO AI is developed by NEO Labs, the brand of NEO TECH LABS TECHNOLOGY L.L.C, a company of My NEO Group in Dubai. Dr Mickael Mosse, Founder and Group CEO of My NEO Group, is the architect of NEO AI.
- How does NEO AI work?
- The user states a goal in natural language. NEO AI turns the goal into a mission, selects the AI models, coordinates specialist agents, connects tools and data, and checks the work before the result reaches the user. The user talks to one assistant and does not have to choose a model.
- What is LILI?
- LILI is the assistant of NEO AI, the one the user talks to. The user gives LILI one objective, in any language, and LILI turns it into a structured mission carried out by a team of AI agents.
- What is Mission Control?
- Mission Control is the command centre of NEO AI. It shows the progress of a mission, the evidence, the validations and the points where a human approval is required, in one view.
- Which AI models does NEO AI use?
- NEO AI is model-agnostic. It is designed to orchestrate leading AI models from third parties together with private models and specialist engines, and to choose the right one for each part of a mission. Third-party models are named by NEO AI only as systems it can orchestrate, without any ownership or partnership implied.
- Can NEO AI act without human approval?
- The user of NEO AI sets an approval policy. Some actions always require a human approval, called a Human Gate, whatever the policy: for example publishing to a public domain or releasing a deployment.
- What is NEO x100?
- x100 is the name of the long-term research and development objective of NEO AI: to deliver more useful intelligence for each unit of compute. NEO AI presents x100 as an objective, not as a measured result.
- Is NEO AI available today?
- NEO AI is in development. Capabilities and deployment models may vary, and some capabilities are on the roadmap. My NEO Group presented NEO AI internationally in June 2026, at the launch of its Global AI and Investor Roadshow 2026. NEO AI states that three of its verticals are anchored by pilots today.
- Which apps are built with NEO AI?
- My NEO Group is building five mobile apps with NEO AI to show how it works in practice: FINORA for finance, FITORA for sport and nutrition, VITORA for health, LINGORA for languages and INTELORA for open-source intelligence. The five apps are proofs of concept, currently in testing, and are not yet public.
- Who is NEO AI designed for?
- NEO AI is designed for entrepreneurs, students and professionals as well as for companies and institutions. The same system serves individuals and organisations, with permissions and controls that depend on the kind of user.
- Where is the official information on NEO AI?
- The official website of NEO AI is neoai.myneogroup.com. NEO Labs presents NEO AI on myneo.tech, and Dr Mickael Mosse publishes his research on enterprise artificial intelligence on mickaelmosse.ai.
Official websites and accounts
Official websites and accounts of NEO AI.
| Website or account | What it is |
|---|---|
| neoai.myneogroup.com | Official website of NEO AI |
| myneo.tech | NEO Labs, the company that develops NEO AI |
| myneogroup.com/creations | The five apps built with NEO AI |
| mickaelmosse.ai | Research and insights by Dr Mickael Mosse |
| GitHub: myneogroup-ai | GitHub organisation of NEO AI |
| GitHub: neo-ai-research | Corpus of the NEO AI Research Program |
| GitHub: myneogroup-ai | Official account |
| GitHub: NEO AI Research Program corpus | Official account |
Announcements by My NEO Group
Posts published by My NEO Group on its official accounts about NEO AI.
- Official sourceLinkedIn - My NEO Group October 2026Insight post on AI agents and the move from systems that answer to systems that act.
- Official sourceLinkedIn - My NEO Group July 2026AI Insight by Dr Mickael Mosse on programmable trust in digital infrastructure.
- Official sourceLinkedIn - My NEO Group July 2026AI research note by Dr Mickael Mosse on autonomous banking.
- Official sourceLinkedIn - My NEO Group June 2026Launch of the My NEO Group Global AI and Investor Roadshow 2026.