AI engineering · Enterprise software · Cloud platforms
We design, build and operate AI systems, enterprise software and cloud platforms for organisations in Malaysia and Europe, linked to the systems they already depend on.
Nine nodes in a three by three grid, joined by one path that forms the FromNine logo: it runs from business context through intelligence, software, cloud and operations to a single system in production.
01 Size and principles
We concentrate on AI, cloud and modern engineering, with the size and delivery discipline of a large company behind us. That gives us the people to staff a mixed programme, and teams that stay with it from the first design to daily operation.
2,400+people who design, build and operate systems for public authorities and companies.
600+government projects for public authorities in EU Member States, some completed and others still running.
400+projects for private companies
Public sectorDesigned around what you run
We work with your existing platforms, data and processes, and replace only what has to go.
Built to be maintained
Tested, documented and handed over, so your own staff can still change it in years to come.
A person decides where it counts
Steps with consequences wait for a person, and every automated action leaves a record you can check.
and public authorities in EU Member States
02 Services
AI is only as useful as what surrounds it: software people like to use, data it can trust, integrations that hold and a platform that keeps it running. We deliver all of these together, so the system works as one and a single team is answerable for the whole.
Generative AI that keeps working in production: grounded in your own data, measured before go-live and monitored while it runs.
AI agents that stay inside the permissions you set and wait for approval before steps with consequences.
Well-governed data that decisions and AI can depend on.
Portals, internal platforms and SaaS products that remain easy to change.
Cloud foundations where releasing is routine and costs are visible.
Systems that exchange data dependably, also when a part fails.
Technical direction before major commitments: options, trade-offs and a roadmap you can justify to your board.
03 Solutions
Begin with a problem, not with a model. Every example follows one route: the friction your staff meet now, the system we would deliver, the gain to measure and the decisions that remain with people.
04 Public sector
600+government projects for public authorities in EU Member States, some completed and others still running.
Public authorities need technology that is accessible, secure, transparent and still useful after the contract that bought it ends. These projects for public authorities in EU Member States taught us to work within procurement rules, more than one language and strict data protection.
We apply the same care to AI: decisions on record, oversight by people, and a design that accounts for the EU AI Act from the first workshop.
Dot map of Europe. The 27 EU Member States are highlighted, the other European countries appear in a lighter tone, and the Antwerp and Ghent offices are marked.
05 Enterprise platforms
Your customer, finance, supply chain and operations data is already in Salesforce, SAP or Odoo. We implement, extend and integrate these platforms, including e-Invoice links to MyInvois, and connect AI to them with the right permissions, so automation runs inside the processes your staff already follow.
Summit Partner
Our work
Adding AI
Help with case handling, with drafts your service staff check before sending
Gold Partner
Our work
Adding AI
Document intake into SAP processes, with exceptions passed to people
Gold Partner
Our work
Adding AI
Back-office workflows on Odoo data, automated
Salesforce, SAP and Odoo are trademarks of their respective owners.
06 Engineering depth
The model is one part. Around it sit integrations, permissions, evaluation, approval steps and operations, and they decide whether it still works at nine on a Monday under real load. That is where AI projects succeed or fail without anyone noticing.
AI helps only as far as it can read from and write to your systems. We design the integrations first: mapping, retries, duplicate handling and errors that staff can see.
The system fetches only what the person asking may see. Access rights are enforced in the retrieval layer and never left to the prompt.
Answers and actions are tested against agreed criteria before and after go-live. Any action with real consequences waits until a person approves it.
Infrastructure as code, staged environments and tracking of quality, response time and cost, so the system can be run, audited and improved once it is live.
The drawing shows six layers. From top to bottom: agents and workflows, with approval by a person before any action with consequences; models and evaluation; data and retrieval; the integration layer; and the business systems that hold the records, such as CRM, ERP, case management and document stores. A cloud platform with monitoring sits beneath them all.
Access control crosses every layer vertically, with identity, roles and audit trail at each step. A request enters at the top and travels down through each layer to the business systems.
Four notes: integrations before prompts (integration layer and business systems); retrieval within access rights (data and retrieval, access control); evaluation and approval steps (models and agents); operated after go-live (cloud platform and monitoring).
07 Approach
Four stages, each closing with a deliverable you can examine and retain. You see progress at every step, and what we build belongs to you.
We look at the business problem, the people affected and the systems in place, then agree what to solve first.
DeliverableAgreed problem statement
We select an architecture, agree how success is measured and plan a first release small enough to go live.
DeliverableArchitecture and acceptance measures
We deliver in short, tested increments that your staff can see, try out and question.
DeliverableSoftware that has passed its tests
We hand the system over or operate it, watch it in production and improve it with you.
DeliverableHandover and plan for operation
What operation teaches us shapes the next round of analysis.
08 Trust and governance
Responsible AI is engineering work, not a policy on paper. Four layers of control surround the systems we build, with your data and your decisions in the middle.
We agree the hosting location with you and record it in the contract before we build, including sovereign options where your requirements call for them.
Our information security management holds ISO/IEC 27001 certification. Every delivery includes least-privilege access, threat modelling and secure engineering from the design stage.
Risk class, documentation, audit logs and decision records, designed for the PDPA, the EU General Data Protection Regulation (GDPR) and, where it applies, the EU AI Act, so you can show who did what and why.
Approval steps, escalation and the power to override, wherever an automated decision has real consequences for people.
Your data and your decisions
Credentials
We design the systems we build for the PDPA and the Cyber Security Act 2024 (Act 854), and for the GDPR, the EU AI Act and NIS2 where they reach you.
Data and security09 In Malaysia
Malaysian groups that export to the EU, own companies there or supply EU customers face two sets of rules at once. At home, amendments to the PDPA have taken effect in stages since 1 January 2025, the Cyber Security Act 2024 (Act 854) has bound NCII entities since 26 August 2024, and e-Invoice through MyInvois reached its last phase on 1 January 2026. From Europe, the GDPR, the EU AI Act and DORA or NIS2 contract clauses arrive through exports, subsidiaries and customers. We design a single system that meets both.
Colleagues at our Cyberjaya office take part in delivery and development alongside the European teams; the office also handles administration and is our regional office for Malaysia and Southeast Asia. Our teams in Asia work in English, Malay, Chinese and Thai. We hold our first discussion with you in English, and the language of the project is settled when the engagement begins. For Malaysian groups with companies in Europe, our European teams also work in Dutch, French, German, Spanish and Italian.
600+ government projects for public authorities in EU Member States, some completed and others still running, taught us to build for accountability. For a public-sector reader in Malaysia, that European experience means services that can accept a MyDigital ID login, interfaces that meet WCAG 2.2 level AA, and documentation that holds up in procurement and audit.
A Malaysian manufacturer with sales companies in the EU must issue e-Invoices through MyInvois at home and send invoices to EU buyers over Peppol, from finance processes that differ by company.
A Malaysian software company adds an AI feature to a product used by customers in the EU, and those customers ask how it works, what it was tested on and who answers for it.
An organisation with NCII duties under Act 854 must notify incidents within six hours, while its EU customers expect notice in time for their own NIS2 or DORA deadlines.
Careers
Engineers, architects, consultants and specialists who build AI systems, enterprise software and cloud platforms for public authorities and companies. Solve problems with real users and real consequences, with colleagues who take care to do it well.
Practices
A first meeting covers your business problem, the systems you run today and the most useful next step. You do not need to prepare a presentation.
For groups with companies in Malaysia and the EU, one team answers for the whole programme.
