Building the AI-Powered Campus

At Ai Everything Abu Dhabi, Ashraf Kotb, Head of Product Management at Ankabut, spoke to CXO DX about the company’s evolution beyond connectivity, its development of an agentic AI-powered EdTech platform, and the role of intelligent personalization, orchestration and governance in transforming the experience of students, faculty and researchers.

What technology solutions is Ankabut building for the education sector, and what challenges are these solutions intended to address?

At Ankabut, we are focused on addressing the challenges educational institutions face as they adopt AI. If you look at most educational institutions today, they have siloed systems, fragmented data and backend processes that are not effectively orchestrated in an agentic way.

What we are building is an EdTech stack that can automate these institutions and provide an agentic framework capable of delivering personalized experiences to different stakeholders.

In traditional systems, the student, faculty member or researcher effectively becomes the integration point. They need to know which system to access and how to interact with it. Am I interacting with the Student Information System (SIS)? Am I interacting with the Learning Management System (LMS)? Why should the user have to know that?

Our approach is different. You should only have to communicate your intent or request. If I want to submit something, I shouldn’t need to know which system it goes to or how it gets processed. The platform should orchestrate that entire journey based on the policies of the institution.

That’s essentially what we are building with EdTech OS. It’s a full stack of different, loosely coupled components and layers, designed with intelligent personalization in mind. The idea is to orchestrate fragmented services and provide a seamless experience to the different stakeholders.

How would this work in practice? Let’s say a university deploys your platform. What would the experience be like for a student?

Let me give you a simple example. Suppose a student misses a class because of an emergency.

In a traditional system, the student needs to understand the university’s policies, submit an absence request, follow up on whether it has been approved, find out what classes or topics were missed, communicate with the professor and understand whether there will be any academic consequences.

So, the student is doing all the integration work between different systems.

With our platform, the student would simply log into the application and say, “I’ve missed a class because of a family emergency.” That’s it. The platform would understand the request and guide the student through the necessary steps. It might ask for supporting information, help submit the absence request and then look at what else has been affected.

For example, did the student miss an examination? Were there important topics covered during that class? Does the student need to contact the professor? Has the student already exceeded the permitted number of absences, putting their academic standing at risk?

These are things the student might otherwise discover only after the fact. With intelligent personalization and real-time context, the platform can bring these things to the student’s attention and guide them on what action to take. Of course, all of this operates within the policies enforced by the institution.

So, beyond informing students about what they have missed, can the platform actually help them catch up?

Exactly. It could advise the student to arrange a meeting with a professor or academic advisor and guide them through booking that meeting. It could also provide tutorials or relevant course materials covering the topics they have missed.

The important thing here is real-time context. In the past, systems were largely driven by predefined rules. But AI can analyze different sources and signals, look at the 360-degree view of a student and advise them on what they need to do at that particular moment.

And that context keeps changing. If I log into my portal today, I might see something completely different tomorrow because my circumstances have changed. Not every student has the same profile, and even the same student’s profile evolves over time.

Let me give you another example. Suppose I have an outstanding tuition balance, and the payment is overdue. The system knows that I have some free time before my next class and that I am close to the student services office. It could alert me and suggest that I use that time to go to the office and resolve the issue.

Or, if I am a student interested in sports, and there is an upcoming sporting event, the platform could recognize my interest and alert me about the opportunity to participate.

There are many such use cases. What we are trying to build at Ankabut is a platform intelligent enough to become your companion, rather than just another system you have to access and navigate yourself. It should be able to initiate interactions and suggest things that are relevant to you at a particular point in time, instead of expecting you to take the driver’s seat and manage everything.

How has Ankabut evolved from its origins as a network connectivity provider for the education sector, and how would you describe its journey so far?

Ankabut started as a National Research and Education Network (NREN), providing networking infrastructure for educational and research institutions. Connectivity is still an important part of what we do. It’s the foundation that enables us to connect different universities. But over the past few years, particularly with the evolution of our leadership and strategy, we have been reviewing Ankabut’s mission and vision.

We are moving towards delivering EdTech technology rather than being seen simply as a connectivity company. That’s an important distinction because people still associate Ankabut mainly with networking and connectivity. But we are doing much more than that.

We are not looking at ourselves only as a provider of individual technology solutions. We want to play a critical role in the wider EdTech ecosystem. If you look at our EdTech AI OS, the stack covers everything from the infrastructure level all the way up to the experience delivered to the student, faculty member or researcher.

Are you now taking these solutions to other universities, and is your focus limited to the UAE?

Yes, we are opening up to other universities and customers. At the moment, our focus is the UAE. Eventually, there may be opportunities to go beyond the UAE, but right now, this is our priority. And our focus is specifically education and EdTech.

We already have different customers using various Ankabut services, including infrastructure services. As we develop the broader platform, we are looking at how we can extend these capabilities across the education sector.

Coming back to EdTech OS, are you essentially developing a single application through which students, faculty and researchers can access different services?

Exactly. We are building an application that covers different personas.

The idea is that students, faculty members and researchers can access the services relevant to them through a common experience, without having to navigate multiple disconnected systems. But that doesn’t mean everyone will see the same thing. The experience will be personalized based on who they are, their role, their context and what they need at that particular time. The application is being released progressively. We are introducing different components and capabilities as we build the overall stack.

What about parents? Would they also have a role in this platform?

That’s a good point. Currently, our focus is higher education, so we are concentrating on the stakeholders directly involved in that environment. When we start looking at K–12 education, parents would naturally become an important part of the journey. We are thinking about that, but it would require us to properly address their role and how they interact with the institution and its services.

Governance is a major concern when institutions start deploying agentic AI, particularly given the amount of student and research data involved. How are you approaching this?

At Ankabut, we understand the risks associated with exposing data. But I think there is a broader question here. Do you want to innovate, or do you want to control? Very often, these are presented as two opposing choices.

Innovation means openness. You allow access to different technologies and platforms. Control, on the other hand, means restricting access and keeping everything within a defined environment.

We don’t see it as a choice between one and the other. It really depends on the use case. Some use cases make sense with open technologies, open platforms or open models. Others require much more control. You have to look at what data is being exposed, what guardrails are in place and how you are going to consume the results generated by AI. And when we talk about governance in education, it’s not only about data security. Academic integrity is also important, particularly on the research side.

If you introduce AI separately into different siloed systems, you still have a governance problem. You might be improving individual systems, but you haven’t addressed the fragmentation. In fact, you could end up with the same silos, only now they are AI-enabled, potentially creating even more problems.

Our approach is to have a service orchestration layer, a service delivery layer and a unified experience through which stakeholders interact with different systems. That gives you a better opportunity to enforce governance consistently.

You don’t want users bypassing the governance layer and interacting directly with different systems that you cannot properly control. So, governance has to be addressed at the architectural level, across the entire stack, rather than trying to solve it independently within every application.

But your governance framework can only control what happens within your environment. What about students or researchers using external AI tools on their own devices? Is shadow AI a concern?

Yes, and that’s where the balance between innovation and control becomes important.

You cannot necessarily control every external tool someone might use. And there may be valid reasons to use external AI tools or open-source technologies for certain use cases. But you need to define the boundaries.

What data are you allowing these tools to access? What information can be shared outside your environment? And how are you going to use the results that come back?

You also have to ask whether those results are trustworthy. I don’t think the answer is simply to block everything outside the institution. Certain use cases may benefit from external tools. But you need to make those decisions on a use-case basis.

As an institution, you have a responsibility to understand which external AI services are being used, what data they consume, whether that data is secure and what information is allowed to move beyond your boundaries. You also need to govern how the outputs are consumed. So, it’s not just about controlling access to AI. It’s about controlling the data you provide, understanding what happens to it and deciding how much trust you can place in the results.

Research is another area where AI is making a significant impact. Could your platform also serve as a gateway for researchers to access different AI tools in a more controlled environment?

Yes. Again, our stack is not built for one specific use case. We see it as a technology enabler within a broader digital AI ecosystem.

It can work with different systems of record. That could be a research system, an SIS, an LMS or another institutional platform. So, when it comes to research, we can provide a controlled, agentic AI framework on top of those existing systems.

We are not trying to change how research is conducted. We are trying to enable research to operate within a more orchestrated and governed environment.

There is also another dimension to this, which is the infrastructure itself. If you look at research today, universities are no longer working in isolation. Research increasingly involves collaboration between institutions, often across different locations and boundaries. That raises questions around network performance, workloads, distributed systems and how those workloads are orchestrated.

How do you distribute workloads? How do you manage interactions between different institutions? How do you maintain control when some of those resources or workloads are outside your own environment?

Orchestration becomes very important in that context. And these challenges are not limited to research. They apply to many other areas as well. That’s why we are building a technology platform that can address these broader requirements rather than creating a separate solution for every individual use case.

Where are you currently in terms of deployment? Is the complete platform already being used within universities?

We are currently in the release phase. We are gradually introducing different components and building on top of them.

We have the service orchestration layer, the experience layer and the service delivery layer. We are also developing our data platform to provide a 360-degree view of student data and other institutional entities. So, there are different layers being developed and released, and we are building the stack as we go.

As AI adoption evolves beyond experimentation, what are Ankabut’s priorities in developing its technology ecosystem and expanding partnerships?

For me, one of the main areas of focus is how we move from experimenting with AI to actually delivering productized services. We have been experimenting with AI for quite some time. Everyone is experimenting. But the question is whether we have really reached the stage where AI is being productized and delivered at scale.

That’s an important point. And I think we also need to change the way we look at AI itself. AI is not just an application or a chatbot. It should be embedded within your architecture. That’s where intelligent personalization and agentic experiences become important. AI should be part of how the entire system operates, rather than something you access separately when you need it.

At Ankabut, we are also looking at potential technology partners. We want to understand what services they are offering, how their AI capabilities are evolving and how we can integrate those capabilities into our ecosystem. That’s really the focus of what we are building and what we are looking for.