How we work
One workflow at a time.
Every engagement follows the same arc: a scoping conversation, a proof on real data, production inside your environment, then measurement before anything widens.
01The arc of an engagement
-
01
Scoping conversation
We talk through one workflow: who runs it, where it waits, which systems it touches, and whether AI is the right tool at all.
- You get
- A plain answer on fit, and a proposed proof.
- We need
- A conversation with the person who owns the workflow.
-
02
Proof on one workflow
We build a focused proof on real data, with a baseline taken first, so the result is a comparison rather than a demo.
- You get
- A working proof, a baseline and a measured result.
- We need
- A representative data sample, and time with the people who do the work.
-
03
Production in your tenant
If the proof holds, we take it into production inside your environment, with permissions, human checkpoints, audit logging and an evaluation harness.
- You get
- A running system, runbooks and a named internal owner.
- We need
- Environment access, integration points, and an owner who will run it.
-
04
Measure, then widen
We measure after go-live against the baseline. The next workflow starts only when the first one holds.
- You get
- Release reports, and a clear decision on what comes next.
- We need
- Agreement on the thresholds that decide a release.
The scoping conversation carries no fee.
02Three ways to engage
-
01
Co-build
We build with your team and hand over the capability.
- You
- Provide the people who will own and run the system.
- We
- Engineers work alongside your team until the system and the know-how are yours.
- Fits when
- You want to own and run the system yourselves.
- Your stack, your tenant, your data
- Capability transfer, not lock-in
- Runbooks and a named internal owner at handover
-
02
Managed service
We run, monitor and improve the system; you receive the outcome.
- You
- Own the workflow and the decisions; agree the thresholds that govern releases.
- We
- Operate the system in your environment, with evaluation and monitoring built in.
- Fits when
- You want the outcome without building an AI team.
- Evaluation and monitoring built in
- Citations and audit trail by default
- Release reports to your owner
-
03
Assessment
We tell you what to build first, and what not to build.
- You
- Open access to the people, systems and data behind your candidate use cases.
- We
- Assess readiness and return a sequenced roadmap you own.
- Fits when
- You have more ideas than certainty.
- Readiness across data, systems, governance and people
- A sequenced roadmap with prerequisites
- Explicit decline recommendations
03What stays constant
- Arabic and English from the first commit.
- An evaluation harness on every system we ship.
- Customer data stays in the customer's tenant.
- People approve the actions that matter.
04Claims we will not make
Some claims win meetings and lose trust. We leave these out of every proposal, every page and every release note.
Zero hallucination.Guaranteed accuracy, or guaranteed savings.Certifications we do not hold.Results we cannot reproduce.
Where we describe alignment with PDPL, NCA controls or SDAIA principles, we mean the system is designed against them, not that a third party has certified it.
05Teaching your team
Itqan is Siyada Tech's education initiative: practical AI programmes in which your team works on its own tasks and leaves with an output it can use.
?Asked often
Questions
Is the first conversation paid?
No. The scoping conversation carries no fee.
What if AI is not the answer?
We say so. Sometimes a process change, a report or a better form does more than an agent would.
Do you work on site?
In Jeddah, across the Kingdom by arrangement, and remotely.
Who owns what we build together?
Ownership is set in the contract for each engagement. In a co-build, the aim is that you own and run what we build together.
Can you meet our IT and security requirements?
Deployments are built to sit within your existing cybersecurity controls, identity provider and change process.
Start with one workflow
Tell us where the work waits. The first conversation is about fit, not a pitch.