Transforming enterprise architecture with AI and ML

The convergence of Artificial Intelligence (AI) and Machine Learning (ML) significantly transforms and enhances enterprise architecture (EA) capabilities, improving analytics across data, business processes, and agility.

Budiman Bujang, head of Digital Strategy & Innovation at Johor Corporation (JCorp), in his keynote message during the FutureCIO Conference in Malaysia, explained how AI and ML change the current EA practice in their company. He discussed the challenges and opportunities in building ML models to transform data architecture and embedding AI/ML in EA, shedding light on the complexities of this transformation.

Layers of the Enterprise Architecture

According to Bujang, the Enterprise Architecture is structured into several layers, each serving a specific purpose in the EA. These layers include the business strategy or vision, the data layer, the application layer, and the technology layer.

“Each year, we sit down with all stakeholders and the heads of the company to understand their vision and alignment. Once we fully understand the vision, we break it into layers of EA: Business layer or understanding each business process within the organisation, data layer, application layer, and technology layer,” he said.

Embedding AI and ML helps overcome the challenges of understanding all the different processes. Bujang said they map the business layers into three other layers: data, application, and technology; curate the values from the mapping process; and identify which asset or technology returns a significant ROI for the organisation.

“To scale it fast, we must embed AI and ML as part of our integration tool. The area we focused on for the past year is churning out the value of each process, technology, and application. There is no point embedding tons of technology without realising the value it gives to the organisation or stakeholders,” he explained.

He said that the creation of necessary apps and technology come after truly understanding the business process and strategy and mapping out the capabilities of each strategy and process. Only then can discussing integration and coordination begin.

You need to think about your ecosystem as one big map and how these apps can talk to each other without any human intervention, so the only way to scale this is through AI.

Budiman Bujang

“You need to think about your ecosystem as one big map and how these apps can talk to each other without any human intervention, so the only way to scale this is through AI,” he said.

Bujang said organisations can centralise applications for common business functions to achieve increased cost optimisation and efficiency.

From experience at JCorp, Bujang explained that organisations could employ predictive analysis based on the data and the actionable insights created with AI and ML. Afterward, they invest in third-party data to compare their own with other investment holding companies by specific pillars they have. They compare the findings with the best in the market for benchmarking, which helps them know how to position themselves.

“Now we are embedding automation, data security, and infrastructure. We create tons of awareness sessions with users so that they will embed AI as part of their business process. This actually scales them up, or we improve their skill set by adding AI as part of their processes,” he said.

Bujang said that evolving the EA is one of the latest initiatives at Johor. To achieve time-to-value very fast, they also rely on APIs. He also said that they check for integration and collaboration capabilities when choosing technologies and whether they have an API to connect with the rest of the applications within their digital ecosystem.

“Another technique is whether they have AI capabilities. There are tons of solutions out there, such as AI, but you need to ensure that the AI can be integrated within your digital ecosystem,” he said.

Dashboards

Bujang also explained how they transitioned from traditional reports to utilising helpful dashboards during meetings. The dashboards created from datasets significantly reduced meeting times by providing clear visibility and highlighting key interest areas and action items.

“These dashboards have been designed so that we know the focus areas. We put some lights, green, red, and yellow, to know what are the areas of interest and the key actionable items that need to be focused on. Before the meeting, companies will be prepared because they know what they will shut down with the yellow and red in the dashboard. The meeting becomes more toward explaining the items they want to do for the next month,” he explained.

“It is straightforward how we calculate efficiency throughout all the business processes. We continuously perform benchmarks with our current processes. We record it down. We send out a survey to the people and then create the next session items based on that,” Bujang added.

He said that once everything is done, and all the technology is in place, they conduct another benchmarking by sending surveys to respective teams or critical personnel involved in technology adoption and displaying the efficiency gains.

“On top of that, we look into cost optimisation as well. There is no point in having all these technologies without optimising the cost so we measure with our current. We compare with whatever was budgeted previously, and then we look at how we can leverage the current set-up, put in more innovation, and improve the current processes,” Bujang said.

Digital strategy evolution

Bujang shared that Johor's digital strategy has evolved from digital transformation to digital empowerment, a shift that is deeply focused on people, process, and innovation. This transformation, driven by three key initiatives: AI, Automation, and Analytics, is a testament to the power of technology to empower and drive change.

“We are transitioning from a normal technical operation, and we name it AI Ops. We use AI to generate all the cost root cause analysis and good cost analysis that we have. If there are any issues, AI helps us generate those reports. That contributes towards our cost optimisation and return on investment,” he said.

Bujang said that their company practices an innovation framework for rapid innovation and technology adoption to validate their technologies across three areas: customer validation, solution validation, and business validation.

“AI is part of the validation to scale fast so that we can perform the validation. What used to take about 2-3 months to validate the hypothesis against the technology we want to implement now takes less time with AI,” he said.

Bujang added that they leveraged AI to simulate a persona within their organisation and explained that they “develop the new capability design by engaging other personas we manage. The personas can be your CEOs, company staff, or stakeholders. We do a lot of experiments to validate whether this technology solution will meet the needs of the business,” he said.

Once the technologies are validated, the team at Johor commits to continuous improvement, developing a minimum viable product that is presented to the board for budget purposes, ensuring that every aspect of their work is aligned with their objectives.