Frequently asked questions

We help organizations put AI to work. Our four practices cover AI strategy and operating model, agentic AI engineering, AI governance and risk, and adoption, change and training. Clients can engage one practice or several. Most use more than one, because a system only pays off once people trust it and use it.

Our discovery sprint runs for about two weeks. A working pilot on your own data usually follows within four to eight weeks, depending mostly on how quickly we can get access to the relevant systems.

Government entities, enterprises and growing businesses across the UAE. The work ranges from a two-week sprint with a single team to programs that run over several quarters and departments.

No. You own everything we build for you, including source code, prompts, test sets, connectors and documentation. Training is part of every delivery, so your own team or another partner can take it on.

No. We are independent of model providers and platform vendors, and we recommend whatever best fits your needs on quality, cost, data residency and support. Sometimes that is an off-the-shelf product, and when it is, we will say so.

A chatbot answers questions. An agent can also act. It reads a request, works out the steps, uses tools such as your CRM or case system, and either completes the task or passes it to a person for approval. Because agents can take action, they need tighter controls than a chatbot does.

The harness is the software around the model. It holds the instructions, the tools the agent may call, its memory and permissions, the points where a person must approve, and the logging and testing. General-purpose harnesses are designed for everyone. A custom one reflects your workflows, approval rules and systems, and that is usually where quality, safety and cost are decided.

Yes. We build connectors to the systems your teams already use, such as email, document management, CRM, ERP, case management and internal APIs. Where it makes sense we use open standards such as the Model Context Protocol (MCP). Every connector has scoped permissions and an audit trail.

Single sign-on and role-based access. Each tool an agent uses gets only the permissions it needs, and consequential actions wait for human approval. Everything is logged, every release is tested against your own cases, and quality, cost and safety are monitored in production. In short, the system is designed to pass your security review.

No system is error-free, so we plan for errors. We measure quality against test sets drawn from your real cases, keep a person in the loop wherever a mistake would be costly, limit what each agent is allowed to do, and monitor live behavior so problems surface early.

Yes. We design and test everything with Arabic and English users, and our test sets cover both languages.

Wherever your policies require. During discovery we agree hosting, model providers and data flows with you, including in-country options, and we document them before any build starts.

It sets out who can approve an AI use case, how its risk is assessed, which controls apply at each level of risk, how it is monitored once live, and who is accountable if something goes wrong. A good one speeds up approvals, because teams know the rules in advance.

UAE federal and emirate-level AI guidance comes first. Where useful, we also draw on international references such as ISO/IEC 42001 and the NIST AI Risk Management Framework. The aim is to fit governance into the risk, security and data policies you already have.

AI programs usually stall for people reasons rather than technical ones. Roles shift, workflows change and trust takes time to build. We plan for this from the outset, with stakeholder mapping, clear communication, champions in each team, and adoption measures we keep tracking after launch.

Programs shaped to each audience. They include AI literacy for all staff, practical sessions for teams adopting a new agent, executive briefings on strategy and risk, and hands-on workshops where engineers learn to build agents, harnesses, connectors and evaluations.

It describes how your organization runs AI as an ongoing capability. That covers where the AI team sits, how use cases are funded and prioritized, which platforms are shared, which skills you keep in-house, and how value is measured. Without one, each new project tends to start from scratch.

Yes. Our fixed-scope starting points suit smaller teams that want results without a large program or platform licence.