Naveed Afzal, Head of Data Science at Takeda Pharmaceutical Company, speaks with Dominic Sartorio, VP of Product Marketing, Denodo, in a video interview about the processes for aligning the data strategy with AI strategy, the root causes of AI project failures, how to succeed with AI, and advice for data science and AI professionals.

Takeda is a leading values-based, R&D-driven global biopharmaceutical company.

Speaking of processes for aligning data strategy with AI strategy, Afzal states that it is critical to focus on the people and process aspects without overtly focusing on technology. Referring to a recent BCG survey, he mentions that 90% of the challenges are not AI-related; rather, 70% of those are people and processes, and 20% are related to data and technology.

Therefore, to gain value from AI and data initiatives, organizations must have effective change management in place. This involves clear communication and providing employee training and support to help understand AI benefits.

Culture plays a crucial role in change management, says Afzal, and the leadership must take the initiative to foster accountability and set the tone to embrace change and innovation. He also advocates for a culture of continuous learning and adaptability to seamlessly integrate change management into the organization.

Additionally, businesses must balance AI implementation with other strategic goals, prioritize objectives, assess RoI, and establish clear business targets, suggests Afzal. One key challenge here is addressing employees’ concerns about AI impacting their jobs, which leads to resistance and low adoption rates.

To mitigate this, organizations should offer learning and reskilling opportunities while creating an environment that creates a positive mindset for AI adoption. This includes providing retaining programs that support employees in transition.

Afzal further adds that many companies lack the necessary IT infrastructure to handle vast amounts of data. Therefore, investing in IT infrastructure and data governance is critical to enabling AI integration.

“The root causes of AI project failures often stem from a lack of understanding, unclear problem definitions, and misaligned expectations,” says Afzal. For instance, stating a broad goal like “improving customer satisfaction” lacks clarity, whereas a more specific target, such as “reducing customer service response time by 30% within six months,” sets clear expectations and actionable milestones.

Having poor-quality data is another challenge that many organizations struggle with. Afzal says, “If you don’t have good quality data, you don’t expect good quality insight.” He also stresses having a strong data management system to ensure reliable inputs.

Organizational infrastructure should be structured well enough to handle scalable solutions while ensuring that it is realistic, says Afzal. “AI limitations are there, and we have to make sure that our expectations are not unrealistic,” he adds.

To succeed with AI, it is crucial to start with a well-defined problem and engage with the business early on to understand the issue to be addressed before diving into the project. Apart from investing in high-quality data, proper governance, change management, and communication skills, organizations must also have AI and data professionals in executive roles.

“Focus on the problem, not on the technology,” says Afzal. Instead of selecting a technology first to fit it into the problem, he advises identifying the challenge and then choosing the best technology to address it.

Many technical leaders tend to focus too much on technological design rather than deeply engaging with the business side, remarks Afzal. However, it is crucial to understand how a solution will integrate into the company’s workflow.

For instance, a manufacturing company might implement a predictive maintenance system, but if they fail to collaborate with the maintenance and operations teams, the impact will be minimal.

Key lessons for young data and AI leaders

Continuing the conversation on leadership in AI, Afzal shares valuable insights for young professionals entering the field. Drawing from his experiences, he emphasizes the importance of understanding data, developing technical and communication skills, and fostering responsible AI practices.

1. Understanding what data matters: Afzal’s first piece of advice is to identify and prioritize the data that truly matters. “Not all data is equal value, so we have to engage with business and understand what data is crucial.”

He recommends asking key questions to gain a complete perspective:

• How do you use this data?

• When do you look at this data?

• When do you report on it?

• Does it need to be up to date — minutes, hours, daily, weekly?

• What purpose does this data serve?

• Who needs to be notified if data is delayed?

Afzal also recommends the book “The Digital Mindset” as a valuable resource for professionals refining their approach to data.

2. Mastering unstructured data and AI technologies: With an increasing amount of unstructured data from social media, IoT devices, and other sources, Afzal stresses the need for AI leaders to develop strong technical skills.

“Going forward, 90% of the data will be unstructured. Data scientists should develop skills to process and analyze this data using advanced technologies like NLP, and machine learning.”

3. Effective communication and storytelling: Beyond technical expertise, Afzal emphasizes the importance of communication. “Successful data projects require continuous monitoring, optimization, and effective communication. It’s vital to master the art of storytelling to tailor your communication to your audience.”

4. Prioritizing responsible AI: Afzal urges AI professionals to ensure ethical AI practices. “Prioritize ethical consideration in your work, addressing biases, and ensuring transparency. Safeguarding privacy is crucial for building trust, and for data scientists, the most important currency is trust.”

“It takes a lot of time to build that trust with business stakeholders, so make sure that you pay your attention to responsible AI.”

For further learning, he recommends the book “Ethical Machines,” which discusses responsible AI in detail.

5. Collaboration and networking: Afzal highlights the importance of professional networking and staying updated with AI advancements. “Collaborate and network. Build a strong professional network. Attend conferences. Stay up to date with the latest trends in technology, AI, and data science.”

6. Focusing on impact and business value: Afzal stresses the importance of aligning AI projects with real business outcomes. “Be very clear about the impact of a project. Sometimes projects continue while the data has changed. There’s no real value for that project to be deployed in production. So be more agile and have an end goal in mind.”

7. Lifelong learning and achieving ROI: Afzal encourages continuous learning in the evolving field of AI. “Embrace lifelong learning. The field of data science is constantly evolving. So stay curious and committed to continuous learning.”

He also shares his approach to achieving ROI. “Invest in scalable core business use cases that will add value to the business. Prioritize data hygiene and governance. Foster a culture of testing, learning, and collaborating with your colleagues and working closely with the business.”

8. Balancing leadership styles: From the leadership perspective, Afzal says, “It’s crucial to find a balance between being too permissive or being too authoritarian. If you are not demanding at all as a leader, you risk becoming permissive. But if you are too demanding, you risk becoming authoritarian. You have to find the right balance between too demanding and too permissive.”

CDO Magazine appreciates Naveed Afzal for sharing his insights with our global community.