Pilot and delivery
One service first. A number to move. Then the next wave.
Summary: We start with one service and one measure that matters to its owner, prove it on the platform in a fixed-scope pilot, then add services in waves, each accepted in writing, and run them as a managed service from the UAE. You keep your data, rules and configuration, and you can leave with them.
Pilot: one service, one number
Pick a service with a measure the owner already tracks: time to complete, the share of cases handed to staff, applications returned for missing documents, or cases per officer. The pilot builds that service on the platform with its test set in Arabic and English. It runs against your interfaces or, while access is being approved, against simulated ones built from your API documentation, and it ends with the measure before and after. A pilot runs six weeks for a public-sector service and eight for a corporate one.
Go live in waves
Services that share a transaction type share a template, so they go live together. Each wave ends with written acceptance by the entity before the next one starts. A dependency register tracks what each wave needs from the entity: interface access, UAE PASS onboarding, the payment channel and security approval.
Run and improve
After go-live we run the service as a managed service: round-the-clock monitoring by a UAE-based team, incident priorities with response targets set in the contract, and a monthly report on transactions, handovers, overrides, safety blocks and service levels. The test set reruns before each release, and production changes go through a pipeline with human approval.
What you own and what we keep
| Yours | Ours |
|---|---|
| Your data, transcripts and audit records | The platform code and its intellectual property |
| Prompts, knowledge content and service rules | Reusable agent patterns and tooling |
| Workflow configuration | Our demonstration environments |
At exit, everything in the first column comes to you in documented formats, with the data export at no charge.
How we price
Three shapes, chosen per engagement: a fixed-price pilot; a subscription per wave of services, with a one-time activation fee after each wave’s acceptance; or a project with stage payments and a support year. Usage allowances carry alerts at 80% and 100%, and nothing above the allowance is billed without your written approval. We quote after a scoping call.
Who does the work
A UAE-based team leads delivery, security and operations. Our rule: engineers and coding tools work on code and synthetic data, and in production, data, logs and transcripts stay with named UAE-based staff.
Want to see it on your own service? We run the demo live with you, in Arabic or English.
Related: Agentic AI Platform, evaluation, why agents stall between pilot and production, a sovereign AI assistant as a fixed monthly fee.
Frequently asked questions
How soon is the first service live?
The pilot service runs on the platform within its six or eight weeks. Production go-live then depends on interface access, UAE PASS onboarding and your security approval, which we track from the first week.
Can we start before we have API access?
Yes. We build against simulated interfaces made from your API documentation and switch to the real ones when access arrives.
Can we leave?
Yes. Your data, prompts, knowledge, rules and configuration come with you in documented formats, and the data export costs nothing.
Do you work with private companies?
Yes. We scope the same pilot for banks, real estate developers and trading companies.
