AW Rostamani Group scales enterprise AI through agentic intelligence and governance

For more than 70 years, AW Rostamani Group (AWR Group) has grown alongside the UAE, evolving from a single Dubai bookshop into a diversified, multi-sector enterprise. Today, its automotive business brings together 14 brands, 26 showrooms and 27 service centres across vehicle sales, mobility solutions and aftermarket services. 

As part of its wider innovation agenda, the Group is building the internal expertise, governance frameworks and technology partnerships required to apply AI responsibly to real business priorities.

One example is an agentic AI solution developed by AWR Group’s Data, AI, and Analytics team with collaboration with the  automotive Spare Parts business to investigate lost-sales opportunities more quickly and consistently.

Revealing the full story behind every lost sale

When a spare parts enquiry does not convert into a sale, the underlying cause may be inventory availability, competitor pricing, an unusual dealer request or a lead that was never a genuine opportunity.

Finding the answer traditionally required teams to examine order history, inventory, pricing, dealer activity and market data across multiple systems. This made it difficult to identify the cause and respond quickly.

To address this, AWR Group developed the Spare Parts Loss of Sales Agent. Built using the Dataiku LLM Mesh, it draws on the Group’s data lake, business systems and relevant external market sources.

Five specialised subagents examine customer purchasing patterns, part performance, inventory, competitor pricing, dealer behaviour, parallel imports and foreign exchange conditions. More than 20 governed tools connect the agents with the data required for each investigation

Dataiku orchestrates the agents, compares their findings and consolidates the analysis into a single response. What initially appears to be an inventory problem may have a different explanation once pricing, dealer behaviour and wider market conditions are considered.

This gives business teams a more holistic explanation without requiring them to trace each possibility across separate systems. Instead of waiting for monthly analysis, they can identify falling conversion rates and emerging inventory or pricing issues sooner.

Each investigation is estimated to save ten hours of analysis on average, freeing teams to focus on customer engagement, high-value negotiations and action on missed opportunities.

Enabling responsible AI through Dataiku

Governance is central to AWR Group’s approach to AI. The Group requires AI projects to pass through its governance process before progressing into production, making effective oversight essential to moving beyond experimentation.

Dataiku provides a unified environment for developing, testing, evaluating and governing the agent while connecting it with AW Rostamani Group’s existing data platforms, tools and models.

This enables the Group’s Data, AI and Analytics team to control access, apply guardrails and evaluate performance throughout development, providing a strong foundation from which to scale AI responsibly.

Business expertise is equally important. The Data, AI and Analytics team develops the tools and integrations, while subject matter experts test responses, refine semantic definitions and add golden queries based on their operational knowledge.

Their participation improves how the agent interprets information and strengthens the relevance and reliability of its responses, without requiring every adjustment to be routed through the central data and AI team.

Creating a governed model for broader AI adoption

The Spare Parts Loss of Sales Agent was launched as a pilot, with business users actively testing and refining it. The agent surfaces findings and alerts relevant teams, while the people closest to the issue interpret the analysis and determine how to respond.

This reflects AWR Group’s wider approach to AI: using technology to make important information available sooner and strengthen decision-making, while retaining appropriate human oversight and accountability.

The Group now intends to build additional use cases on the foundation established through the pilot. The same inventory, pricing, dealer and market-analysis capabilities could support new agents, including one for dynamic-pricing recommendations.

By combining its internal Data and AI expertise with Dataiku’s technology and governance capabilities, AWR Group is establishing a scalable model for applying AI to practical business priorities and helping teams respond to important signals with greater speed and confidence.