In a world increasingly obsessed with AI replacing people, Duraisamy Rajan Palani offers a far more grounded perspective: the future may belong not to larger automated systems, but to smaller teams with sharper judgement. In this conversation, the Founder and CEO of Archimedis Digital reflects on everything from AI hallucinations in life sciences to why his hiring bias leans towards talent from tier-two cities. What emerges is not blind optimism around AI, but a nuanced understanding of where technology genuinely creates value, where it quietly inflates complexity, and why human oversight may become even more important in the age of intelligent machines.
For those who are hearing about Archimedis Digital for the first time, what are you building, and what gap made you start it in the first place?
Sure. Archimedis Digital operates at the intersection of life sciences and technology. Before starting Archimedis Digital, I was part of our family-owned pharmaceutical manufacturing company, where I saw that operations were highly efficient due to the well-oiled processes the industry was always known for, and the availability of strong talent in India. The industry was thriving on its manufacturing capabilities.
But I also noticed that even very advanced pharma manufacturing units were not fully leveraging the advantages of digital technologies. So, we started experimenting with some digital tools within pharma operations. We started with an ERP system for pharma manufacturing and gradually expanded to real-time data capture and related systems.
That gave us the idea that Archimedis Digital could serve the larger life sciences community. By life sciences, we mean pharmaceutical, medical devices, and biotech companies that already possess the scientific capability to advance their businesses but need digital tools and technologies to complement that expertise.
That’s how Archimedis Digital was born. Today, we provide IT services and software products to life sciences companies.
At what point did AI move from being “interesting” to becoming something you had to seriously think about in the context of your team and talent strategy?
I think that since the inception of our company in 2022, we have known AI would disrupt the industry in a colossal way. Ours is a knowledge industry and heavily talent-based, hence, AI was always on our radar.
At the same time, we realised that even if AI takes care of certain tasks, there still needs to be a human in the loop to review and approve the work it generates. That led us to quickly build domain expertise across the life sciences by bringing in people with backgrounds in manufacturing, quality, regulatory affairs, and drug discovery.
We became increasingly earnest about AI in the last two to three years. Internally, we even ran an experiment in which we asked ourselves: if we were to completely replace one part of our operations with AI, what would it be? We chose marketing because it was comparatively less risky for the business.
So we went all in using AI for content creation, image creation, video creation, and more because generative AI was becoming extremely popular at that point. But then we quickly realised that the human touch resonates far better in the market. People can now immediately identify AI-generated content.
That’s when we scaled back and recognised that this was not the right approach for us. What remained from that experiment, however, was the seriousness we brought into AI. We continued experimenting with AI in other functions, customer engagements, and products, but in a far more cautious and balanced manner.
Today, AI is part of our business, product, and talent strategies. But we are careful about where and to what extent we apply it, so that cost and value remain balanced.
Everyone is talking about AI as if not adopting it is a strategic failure. From where you sit, how much of AI in talent strategy is a real need versus the collective anxiety of being left behind?
If I had to put a number to it, I would say AI should probably contribute around 40–50% in a new team setup initially.
Where AI really stands out is in its ability to generate a large volume of output very quickly. But once something is generated, humans still need to validate and review it. That means subject matter experts become even more important.
For example, we experimented with one of our internal application teams that manages our ERP systems. The team had eight people, and we decided that four would continue while the other four roles would be effectively supported by AI tools for code generation.
We discovered that the people behind the AI screens were unable to cope with the sheer volume of generated output. If AI generates 1,000 lines of code, someone still has to validate them all before they can move into production.
That experiment changed our approach to hiring. Earlier, interviews focused on whether someone could write code. Now, we are more interested in whether someone can review and critically evaluate an AI-generated codebase.
Today, we place greater emphasis on critical thinking, design thinking, and the ability to review AI-generated work, rather than on production capability alone.
We also realised that AI contribution cannot be standardised. Some projects may run with 40% AI, some with 60%, and some may not need AI at all. The key is understanding how much of the work involves generation versus review and approval.
Ink Dive: Artificial Intelligence Can Power Healthcare. Human Empathy Will Define It
Can you share a specific instance at Archimedis Digital where you evaluated an AI use case in talent and consciously decided not to pursue it? What made you pull back?
Yes. One example came from a clinical data management project we were executing for a client.
At the beginning of the project, we believed it would require deep domain expertise because reviewing clinical data entails evaluating drug efficacy, adverse events, and patient responses. That requires higher cognitive application and deep domain understanding. However, the client, an early-stage startup, was keen on using AI because the industry was then aggressively moving towards AI adoption.
So, we decided to run both tracks simultaneously. We continued the human-led project while also building an AI initiative in parallel.
Initially, the AI system moved very quickly. It was able to identify patterns and highlight areas where data needed closer review. Those insights were useful because the human teams could leverage them effectively. But midway through the project, we realised that the amount being spent on AI infrastructure and token consumption had become extremely high — far higher than what we were spending on the human teams.
At the same time, the actual incremental value from AI remained limited because, eventually, the human teams would likely have identified those issues anyway.
So, we advised the client that while AI had delivered value during the early stages of pattern identification, continuing to use AI throughout the project was not generating proportional value. We recommended continuing primarily with the human-led approach.
That project reinforced for us that the important question is not whether AI should be used everywhere, but exactly where within a project it creates meaningful value.
There is a narrative that AI reduces bias. There is an equally strong argument that it scales. Where do you stand, especially as someone building teams in a growing organisation?
I definitely think AI creates bias because, at least today, AI has no independent thinking of its own. It relies on the knowledge and data it has been trained on.
To me, bias itself is an outcome of knowledge and experience. Whenever humans make choices, those decisions are shaped by collective experiences and accumulated knowledge. AI works in a similar way because it learns from existing datasets.
So naturally, the knowledge we feed into AI shapes its behaviour and, consequently, its biases.
In life sciences, especially, this becomes extremely important because we are dealing with life-saving products and decisions. An AI hallucination should never be allowed to affect patient safety. That is why governance frameworks and guardrails are critical in our industry. In certain areas, we even restrict AI’s ability to operate beyond predefined parameters.
Another thing I have noticed is that AI often tries to please the user by generating answers that sound convincing or agreeable. But in industries like ours, that approach can become dangerous if it is not properly governed.
So yes, AI definitely carries bias, which makes human oversight extremely important.
Do you think we are moving towards smaller, more strategic talent teams augmented by AI, or simply more overloaded teams with better tools?
I’d like to believe the movement is towards smaller, AI-enabled teams.
As I mentioned earlier, AI creates tremendous value in generating solutions and accelerating work, especially during the early stages of problem-solving. But the real value emerges when subject matter experts are empowered to review and validate AI-generated work efficiently. Whether it is coding, clinical operations, manufacturing, or any other function, domain expertise still matters deeply.
I still do not believe AI has surpassed human cognitive capability. So, in my opinion, the future lies in smaller teams with strong subject matter expertise that are augmented by AI rather than large, traditional teams.
If you had to be brutally honest, where do you think companies are currently over-investing in AI within talent, and getting very little in return?
I think analytics is one major area where companies are over-investing in AI while getting limited returns.
Organisations have always had access to analytics tools, dashboards, reports, and data. The real challenge was never the lack of information; the challenge was converting that information into meaningful decisions. Today, many companies are using AI simply to accelerate data analysis and reporting, which traditional tools could already handle effectively.
The true value of AI should come from helping humans model different scenarios, generate possibilities where data may not yet exist, and assist in decision-making. Instead, companies are often investing heavily in accelerating analytics without improving the underlying data quality or the actual decision-making capability.
That, in my opinion, is where the focus needs to shift.
Every founder has a certain bias when it comes to hiring, whether they admit it or not. What is yours, and how has it shaped the kind of team you are building?
My bias is towards tier-two cities.
A lot of founders believe knowledge and talent are concentrated in tier-one cities because of the opportunities available there, and that is true to some extent. But I believe the future, especially in the AI era, depends far more on critical thinking and adaptability than purely technical skills.
People in tier-two cities often navigate relatively limited resources and opportunities, which means they tend to apply stronger critical thinking and problem-solving abilities.
Also, AI is such a new technology that people need to keep unlearning and relearning. I feel professionals from tier-two cities often demonstrate stronger learning intent and adaptability.
So, our hiring strategy gradually evolved from purely technical assessments to identifying people with strong critical-thinking abilities and then empowering them with AI skills.
This bias has actually become part of our execution strategy. We have grown from around 20 people to nearly 200 people in just three to four years, and a significant part of that growth has come from talent across tier-two cities.
Duraisamy Rajan Palani is the Founder and CEO of Archimedis Digital, bringing more than 25 years of experience across technology, healthcare, life sciences, and digital services. His role covers business strategy, revenue planning, customer relationships, and daily operations. He is also deeply involved in building leadership within the organisation and ensuring that the company operates with financial discipline and strong governance. His approach is hands-on and rooted in clear execution. Rajan has a strong academic background in both technology and business. He has completed executive programmes at Harvard Business School Online (with a focus on finance) and at London Business School under The Entrepreneurial Edge Programme. He holds a Master of Science in Software Systems from BITS Pilani and a Bachelor of Technology from Anna University, Chennai. This blend of education has helped him work across complex and regulated industries.
