Incorta serves its customers in Saudi Arabia from inside the Kingdom

Incorta processes billions of tokens on aixplain, across two dedicated deployments of the aixplain OS: one in a cloud region inside Saudi Arabia, one on Incorta's own servers in the US. Incorta gives business users live operational data from the systems a company already runs, and since 2024 the models behind its GenAI features have run on aixplain. For its customers in the Kingdom, the data and the models stay where the customer's rules say.

“The main reason we are partnering with aixplain is to… automate this whole process, from utilizing the right LLM, serving the right LLM, finding the right LLM.”

The challenge: one product, local rules

Incorta's customers wanted to ask their operational data questions in plain language. Incorta wanted to give them that without becoming a model infrastructure company. For GenAI, Elkady said on the aixplain'd podcast, “we don't have that team, we don't have that expertise.”

Its customers in Saudi Arabia include government and sovereign organizations. Their data has to stay in the Kingdom, and so do the models that read it.

And the product had to stay a single product: the same features and the same API, wherever a customer sits.

Why aixplain: the same OS, wherever the rules require

• Your infrastructure. The aixplain OS runs in aixplain cloud, as a dedicated deployment in a region, or on a customer's own servers. It is the same OS in each place.
• Your models. Every model sits behind the same interface, so changing one is a change of configuration. Incorta builds a feature once and keeps it when the model changes.
• Your rules. Access and limits are set per key and per model, and they hold in every region.

The solution: a dedicated aixplain OS in the Kingdom

In-region. aixplain runs a dedicated deployment of the aixplain OS for Incorta in a cloud region in Saudi Arabia. It serves the large language and embedding models behind Incorta's GenAI features to Incorta's customers in the Kingdom, and their requests are processed there.

The same API on both sides. In the US, the same OS runs on Incorta's own GPU servers. Incorta's application calls both through the same endpoints, so a feature built once works in both regions.

Governed from the start. Incorta's admins set rate limits per API key and per model, and decide which keys can reach which models. Incorta built its own monitoring dashboard on aixplain's usage endpoints.

“aixplain making GenAI is like drinking a cup of water. You don't need to think about it.”

What's next: from models to agents

On the same OS, aixplain and Incorta have built agent prototypes for Incorta's customers: a BI analyst agent that answers demand and shortage questions across operational data, and an invoice-processing agent that checks each invoice against vendor and purchase-order rules. “For the pipelining we use aixplain,” Incorta Principal Machine Learning Engineer Dr. Ebrahim Alareqi told an AICamp meetup in San Francisco in January 2025.

When those agents reach Incorta's Saudi customers, they run in the same in-region deployment, under the same rules. Every agent answers to an owner, inspector agents check every run, and every run is recorded with what it did and what it cost.

Use the infrastructure that makes sense today. Keep the freedom to change tomorrow, without losing the agents, models and controls already built.

“aixplain, which is one of our very important partners.”

See aixplain at work

aixplain is the operating system for accountable AI agents you own. Organizations run more of their operations and services on agents, on any model and any server. Every run is checked against policy, and every agent answers to an owner.

Book a demo to see how your agents and models can run in the region your customers require.