Healthcare AI Expert in Abu Dhabi

Healthcare AI expert in Abu Dhabi: Karthik Telidevara, Manager – AI Strategy & Transformation at Burjeel Holdings, on governed, measurable clinical AI.

Karthik Telidevara is a healthcare AI strategist based in Abu Dhabi, working as Manager – AI Strategy & Transformation at Burjeel Holdings, where he helps take clinical and operational AI from idea to governed production.

A Healthcare AI Strategist Working Inside an Abu Dhabi Health System

Most writing about healthcare AI in Abu Dhabi comes from outside the hospital. My vantage point is different. I lead AI strategy and transformation at Burjeel Holdings, and I am based in Abu Dhabi, so the questions I work on every week are the practical ones: which use case deserves funding this quarter, how a model is validated before a clinician relies on it, who signs off, and how we will know in six months whether it actually helped.

That work sits on top of more than 13 years spent across pharmacy and drug analysis, healthcare business strategy, hospital operations, digital transformation and, more recently, production AI. The mix matters. A pharmacist sees medication risk differently from an engineer, an operations lead sees bed flow differently from a data scientist, and good healthcare AI strategy needs all of those views at the same table.

Burjeel's AI initiative, AiMed, is where much of this comes together. It gives a structure for experimenting responsibly, documenting what works, and scaling only the tools that earn their place. The rest of this page explains how that approach plays out in the Abu Dhabi context specifically.

Governance Before Code: How Projects Start

Every AI project I am involved in begins with governance rather than a demo. A cross-functional AI governance committee, with clinical, operational, IT, legal and quality voices, reviews the use case before any build starts. The committee asks whether the problem is real, whether AI is the right tool, what the patient-safety exposure is, and which data the system will touch.

The second step is baselining. In my interview with The AI Times in October 2026 I put it this way: "We need to understand from day one what the current baseline is, where we want to go with AI, and whether it is a real win." If we cannot measure the current state of documentation time, call-back rates or turnaround, we cannot honestly claim improvement later, so KPIs are agreed and measured before work begins.

Only after that does a controlled pilot start, with a defined group of users, a fixed time window and clear exit criteria. Clinical validation runs alongside the pilot, not after it. Scaling is a separate decision, made on evidence. You can read the full feature and the other coverage on the press page.

Working Within the Abu Dhabi Regulatory Environment

Healthcare in the emirate is regulated by the Department of Health – Abu Dhabi (DoH), which sets expectations on data protection, clinical accountability and the use of health information. Any AI system that reads or writes patient data has to respect those rules, and has to fit within the emirate's health information exchange, Malaffi, rather than work around it.

In practice this shapes architecture decisions early. Data residency determines which cloud regions and which model endpoints are acceptable. Access control has to map onto existing hospital identity systems. Audit trails need to show who saw what, which model version produced an output, and which clinician accepted or rejected it. These are not features added at the end; they are the frame the project is built inside.

Where international frameworks such as HIPAA or GDPR are useful as reference points, I treat them as helpful benchmarks rather than substitutes for local requirements. For a wider view of UAE regulators, including DHA and MOH, see the overview of healthcare AI regulation and adoption in the UAE.

Use Cases Taking Shape in Abu Dhabi Hospitals

Two families of use case have moved furthest in my own work. The first is ambient clinical AI: tools that listen to a consultation, with consent, and draft the clinical note for the doctor to review and sign inside Oracle Health. The clinician remains the author. The goal is to give time back to patient conversation and reduce after-hours documentation, not to automate clinical judgement. I have written more about choosing these tools in the guide to ambient clinical voice technology.

The second is voice AI agents built in-house for patient communication: triage support, post-surgical follow-up calls, and reaching patients who missed an appointment so they can be rebooked directly into the scheduling system. That last workflow is described in a Telnyx customer story. Voice agents are attractive because a large share of hospital friction happens on the phone, and because Arabic and English both need to work naturally. The AI voice agents in healthcare guide covers the design choices in depth.

Behind these sit less visible but equally important projects: making operational data usable, standardising how teams request AI help, and giving staff confidence with everyday assistants such as Microsoft Copilot.

The Technology Choices: Microsoft and Google

My platform work is centred on the Microsoft and Google AI ecosystems. On the Microsoft side that means Azure AI Foundry and Azure OpenAI for model access and orchestration, and Microsoft Fabric for bringing clinical and operational data into one governed analytics layer. On the Google side it means Vertex AI and Gemini, with MedLM where a healthcare-tuned model fits the task.

Choosing between them is a workload decision, not a loyalty decision. Questions that drive it include where the data already lives, which regions satisfy residency requirements, how strong the evaluation and monitoring tooling is for the use case, and what the hospital's existing identity and security estate looks like. Keeping to two well-supported ecosystems also reduces the integration and security surface the team has to maintain.

Whatever the platform, every production system I support carries the same basics: an evaluation set agreed with clinicians, monitoring for drift and failure modes, a named human checkpoint for clinical outputs, and a plan for turning the system off safely if it stops performing.

What Abu Dhabi Organisations Can Take From This

For hospital leaders in the emirate, the most transferable lesson is that the hard part of healthcare AI is rarely the model. It is agreeing the baseline, earning clinical trust, fitting into regulated data flows, and being honest when a pilot does not deliver. Organisations that build those habits first tend to move faster later, because each new use case reuses the same governance and measurement scaffolding.

The second lesson is about people. Tools fail when staff do not trust them or do not know how to use them well. Investing in training, champions and clear guidance on what AI should and should not be used for is as important as the technology itself.

If you want the national picture, read about my work as a Healthcare AI Expert in the UAE. For the method behind these projects, see the healthcare AI strategy framework, and for risk and controls, the page on healthcare AI security and governance.

Frequently Asked Questions

Who is Karthik Telidevara?

Karthik Telidevara is a healthcare AI expert in Abu Dhabi. He is Manager – AI Strategy & Transformation at Burjeel Holdings and has more than 13 years of experience across pharmacy and drug analysis, healthcare strategy, hospital operations, digital transformation and production AI.

How does a healthcare AI project start in an Abu Dhabi hospital?

In the approach Karthik follows, a cross-functional governance committee reviews the use case first, KPIs are baselined before any build, and a controlled pilot with clinical validation runs before any decision to scale.

Which regulator governs healthcare AI in Abu Dhabi?

Healthcare in the emirate is regulated by the Department of Health – Abu Dhabi (DoH). AI systems that use patient data must respect its data protection and accountability requirements and fit within the Malaffi health information exchange.

What kinds of AI are being used in Abu Dhabi healthcare?

Examples from Karthik's work include ambient clinical AI that drafts notes for clinicians to review in Oracle Health, and in-house voice AI agents for triage, post-surgical follow-up and rebooking patients who missed appointments.

Which AI platforms does Karthik work with?

He works within the Microsoft and Google ecosystems: Azure AI Foundry, Azure OpenAI and Microsoft Fabric, and Vertex AI, Gemini and MedLM.