Tarun Sukhani, Founder and CTO of Abundent, brings a pragmatic, deeply technical perspective to the current AI hype. He discusses why he sees Gen AI as an incremental shift, not a revolution, and why the biggest cause of IT project failure is choosing the wrong tech stack instead of the most appropriate one.
Follow The Ortus Club on LinkedIn to keep up-to-date on our conversations with today’s top IT leaders.
Here’s a glimpse of what you’ll learn:
- Gen AI is an Incremental Shift, Not a Revolution. It’s a significant advancement of 50-year-old technology, now supercharged by data and compute power. The focus should be on practical integration, not hype.
- The Most Important Question is “What is the Appropriate Stack?” Project failures often come from following the bandwagon instead of choosing technology that aligns with the company’s unique culture, competencies, and goals.
- The Real AI Challenge is Governance and Value Delivery. Beyond the technology, project failures are most often rooted in poor governance and a misalignment between the AI’s output and core business objectives.
Please tell us about yourself and your role.
My name is Tarun Sukhani, and I’m the founder and CTO of Abundent. We are an ed-tech solutions provider and a consulting firm based in Southeast Asia, focused on intelligent automation projects that mix AI and automation.
I’ve been a senior-level IT executive for about 15 years, but the vast majority of my career has been spent in deeply technical roles like programming, architecture, and data science.
The IT landscape evolves rapidly. What challenges are you currently facing?
The latest AI has been around for 50 years; what we’re seeing with Gen AI is an incremental set of improvements, though this one is more monumental because it involves human language. We don’t look at it as a completely new emergence, but rather as a new layer of complexity.
Our focus is on how we can integrate these new capabilities into our existing solutions to provide a much better way for humans to get work done. This could be through conversational interfaces, more automated data analysis, or customer support. Fundamentally, the algorithms have existed for decades; we just now have the data and the compute power to train them at a massive scale.
What trends or technologies do you believe will have the biggest impact on the IT industry?
If you look at the surveys, more than 70% of the expected changes are because of Gen AI. The focus is shifting to enterprise AI and how to effectively integrate it into existing systems. We will see more workflow automation and more interaction with AI bots and avatars to get basic stuff done. This opens up the plane for humans to do more value-added activities and have more meaningful interactions.
How is the regulatory environment changing, and how are IT leaders navigating these shifts?
Everyone is familiar with the EU AI Act, GDPR, and local Personal Data Protection Acts (PDPA). In Europe, the focus is on data security, privacy, and the proper use of AI in terms of explainability, bias, and ethics. The frameworks and guidelines for how AI should behave do exist.
The real question is whether companies are willing and able to implement these measures. From our front, we look at these regulations and try to conform as best we can, especially within the realm of ethical AI and minimising bias in decision-making.
How do you balance the need for innovation with the demand for stability and security?
This is the age-old question of trade-offs. Our approach is to implement something, go full tilt, and then start to measure. We look at the decisions being made, where biases are being produced, and why. Then we have to go back to the training data and our fine-tuning methods to understand what is causing the biased behaviour.
There is an established field of AI explainability and alignment research. We try our best to achieve minimal-bias systems, but we know it’s inevitable that some will be introduced, so we have to work with our customers to ensure such biases are minimised.
Reflecting on your experiences, what’s been a memorable insight from knowledge-sharing discussions?
A lot of the discussion in the AI community nowadays is mostly along the lines of AI project failures. Studies have pointed out that many issues revolve around a misalignment between value delivery and business objectives, and a lack of proper governance.
So, the key conversations we’re having are about what better frameworks we can use for defining business objectives and for data and decisioning governance. Who is responsible for a bad decision made by an AI? These discussions have far-reaching consequences.
What’s one question all CTOs should be asking themselves today?
I think all CTOs should be asking themselves, “What is the most appropriate stack for my company and for my personnel?”
A lot of bad decisions are made, and a lot of money is misspent because companies tend to follow the bandwagon rather than looking at their own internal capabilities and objectives. This misalignment is why we see so many project failures. It comes down to choosing a stack that fits with your company culture, your competencies, and your business objectives.
How do you approach mentoring and developing the next generation of leaders?
As an ed-tech company, we are intimately involved with the educational community. We focus a lot on upskilling, reskilling, personal goal setting, and apprenticeships. We cover the whole gamut of learning, from kids’ camps to professional lifelong learning. We try to bring that external educational focus in-house to our own staff, figuring out what incentive structures will encourage them to improve their competencies.
More Interviews
