From AI Literacy to AI Product Judgment: Lessons from “Pharma Tech Essentials: A Modern Guide for Safety Professionals”

Reflections on one of my favourite books

The greatest obstacle to discovery is not ignorance — it is the illusion of knowledge. — Daniel J. Boorstin

I have been thinking a lot about what separates AI literacy from real AI product judgment.

AI literacy helps you understand the vocabulary: models, data, automation, agents, copilots, prompts, evals, guardrails. Product judgment is harder. It is about knowing which problem is worth solving, which data can be trusted, what the model can and cannot do, how the system should be evaluated, where a human must remain in the loop and which constraints should shape the product from day one.

That is where Pharma-Tech Essentials by Gaurav Goel felt especially useful to me.

I have known Gaurav for more than a decade, from Vitrana’s early days. I was in Product Management and he was leading the AI and Analytics function. He was also one of the first people who guided me seriously into machine learning and AI - not through theory alone, but through patient conversations, sharp questions and a practitioner’s instinct for what actually works.

Gaurav is a great simplifier, which is his superpower. And it shows in this book!

We worked in the same team then. Years later, we still speak almost every week, usually exchanging notes on AI, product thinking and where the industry is heading. So when he wrote this book, I read it not just as a reader, but as someone who knows the practitioner behind the writing.

I have now read the book twice end to end. :-)

The first reading helped me understand the structure and flow of the book. The second reading was more useful. I found myself going back to specific sections as a product person: data, systems, language models, evaluation, privacy, regulatory thinking and the changing nature of work inside pharmacovigilance.

That is when this book started feeling less like something you finish and more like something that should sit on your desk.

What I liked most is that it does not treat AI as a thin layer added on top of business processes. It starts with the basics. How do you think about data? What is the difference between transactional and aggregated data? What changes when the data is structured versus unstructured? How does daily operational work become actionable intelligence? How to understand regulatory point-of-view and their perspective on AI?

These may sound like simple questions, but in real product decisions, this is exactly where teams often move too quickly.

The book also explains the pharmacovigilance landscape in a way that is practical and current. PV, at its core, is about patient safety. But the work around it is changing. As AI, automation and agentic systems enter the industry, the role of the domain expert is expanding. It is no longer enough to understand the process in isolation. The people shaping these systems increasingly need to understand data flows, IT systems, product trade-offs, model behaviour, validation, privacy and regulatory expectations.

That is a meaningful shift.

The part I found particularly useful as a product leader was the first-principles explanation of language models, machine learning concepts, the modelling stack and evaluation metrics. It is written with enough depth to be useful, but without turning into an academic detour. More importantly, it helps a product person ask better questions.

And that is where my own thinking has sharpened.

For AI product managers, knowing what AI can do is not enough. The harder question is: how do we know whether it is working well enough to trust?

As product managers, we are trained to start with customer needs and business needs. That remains non-negotiable. But in an AI and agentic world, another question has to sit next to it from the beginning: how will we evaluate this?

What are we evaluating? Who evaluates it? What does good look like? What are the acceptable failure modes? Where does a human need to stay in the loop? What evidence gives us confidence to ship, scale, or stop?

In traditional software, correctness is often easier to reason about. In AI systems, that confidence has to be built differently. You need an evals-first approach. You need to design for testing, validation, monitoring, explainability and traceability from the start.

This is especially important in pharmacovigilance, where patient safety, auditability and regulatory trust are not optional. You cannot bolt them on later. Gaurav makes the regulatory perspective a central part of the book and that is one of its biggest strengths.

It pushes you to think backwards from what matters: patient safety, traceability, explainability, privacy, compliance and responsible use. That way of thinking is deeply relevant to PV, but it is not limited to PV.

Any product manager building AI in a regulated or high-stakes domain needs this discipline. Healthcare, life sciences, finance, insurance, public sector, enterprise automation — the principle is the same. Build backwards from the constraints that do not move. Be honest about what AI can do today, what it cannot do yet and where the real trade-offs sit.

That honesty matters.

We are in a phase where it is easy to get carried away by what the next model can do. But product building is not just about chasing capability. It is about understanding the system around the capability: the user, the workflow, the data, the risks, the operating model, the regulatory boundary and the evidence required to trust the output.

That is why I would recommend Pharma-Tech Essentials not only to pharmacovigilance professionals, but also to product managers building AI products.

PV professionals will recognise their world in it. Product managers will find a grounded way to think about AI in domains where mistakes have consequences.

The credit for this goes to Gaurav , not only for knowing the subject, but for distilling years of hands-on practitioner experience into a guide that is clear, practical and usable.

Highly recommended!

It is already on my desk.

Pharma-Tech Essentials by Gaurav Goel: https://www.amazon.in/Pharma-Tech-Essentials-Modern-Professionals-ebook/dp/B0FM4GTCW5