The hardest part of becoming an AI-first company: It’s not the technology

Axtria president and CEO Jaswinder “Jassi” Chadha recently stood in front of more than 50 chief executives at an event hosted by Anthropic. He walked them through how Axtria remade itself as an AI-first company and, during that meeting, he was asked what surprised him most during the whole effort. He said it had nothing to do with models, data pipelines, or tooling.

“Winning the hearts and minds of every person in the organisation, and making them believe in a 10X difference, not incremental, was the toughest challenge. Not technology!”

Right now, life sciences organisations are living with the same difficulty. Pilots have multiplied faster than changed ways of working. Closing that distance is management work, and Chadha revealed how Axtria tackled it in a very specific order.

Experiment before commitment

There was no traditional plan, per se. Instead, Chadha talked to industry leaders, held a few personal convictions, and found one or two curious leaders inside the company who wanted to establish what was possible. The company committed resources before anyone named an outcome. Teams tried the tools on hand, such as Windsurf, Cursor, Co-Pilot, and others, and learned by using them.

Product development productivity moved first. Once Chadha could see those gains, he made a top-down call to become an AI-first company and rolled Claude out to everyone at once, skipping the usual pilot cohort. People inside Axtria have grown steadily more convinced since, as have the clients and partners watching what it does.

Companies more often work the other way round. They write the strategy deck, stand up a steering committee, and then look for use cases to fit a commitment they have already made. But in Axtria’s case, a handful of people produced the proof. Only then did it spend what a company-wide rollout costs in attention and goodwill.

Incremental is the easy thing to believe in

Chadha noted that people believe in incremental gains easily. But a 10X difference asks each of them to imagine their own work done in a fundamentally different way, which takes conviction on top of capability.

Take a commercial analytics function, for example. “Incremental” would mean a quarterly brand performance review done in two weeks instead of three. But a 10X difference would have the review assemble itself from the data each morning, allowing the team to spend its hours on the decision the review was meant to inform. People will sign up to work faster. Accepting that a good part of today’s job has become the machine’s work, and that their own value now sits somewhere else, asks much more of them.

Programs stall there. In Axtria’s case, the sequence helped, because by then people had watched a colleague do something they thought impossible. That buys more credit than a target handed down from above.

4 choices that made it stick

Chadha pointed to culture and to four decisions that held the change in place.

• Leaders went first. Axtria’s leaders shared their own accounts of why AI worked for them personally, so adoption was never something asked only of other people.
• Nobody lost a job, and the bar went up. Chadha was deliberate about that: the change is meant to make each person more capable, and it asks more of all of them.
• Teams experimented in public. Hackathons and shared experiments took the place of quiet pilots, which made the failures visible and instructive.
• Everyone got the same tools. Access went out equally, with no favoured cohort.

Axtria changed its structure to match. It took some of its strongest people out of the business to lead the change and asked a few business leaders to build small teams that could incubate new ways of working. Product R&D went first and showed the way to the rest. It also rewrote the performance assessment criteria so that the way it judges work matches the way work now gets done.

One operating model then went across every function. Engineering moved from scrum teams to AI pods. Finance closes the month faster, with the numbers reaching people at the point of decision. Marketing generates client presentations with AI. The services group sells AI-first offerings, and the product organisation is rebuilding its flagship platform from software into a headless, autonomous system. Each function worked out for itself what the model meant at desk level.

The upskilling ran alongside all of it. Axtria put structured learning programs in front of existing staff instead of hiring the skills in, and reports that 100% of its workforce is now AI-certified. Certification turned the rollout into an expectation: access was universal, so competence became part of everyone’s job.

Governance sits under all of it

Chadha called governance the most urgent pillar of the whole effort. The company reviews every statement of work for AI risk, blocks high-risk tools automatically, and tracks and attributes every AI dollar it spends. He credited that discipline with letting Axtria move quickly without ever putting client trust at risk.

A pharma commercial organisation carries more of this weight. It works under medical, legal and regulatory review, privacy constraints and audit expectations, and it cannot run open experimentation on top of nothing. Building those controls alongside adoption costs very little. Retrofitting them after an incident costs a great deal.

What another company can copy

Axtria reported more than 24,000 hours freed in corporate functions, roughly 7X return on its Claude spend, 246 agents live, and $2.5mn of delivery capacity freed. Product R&D teams ship 35% more story points per release.

The real story is the sequence behind those figures. Evidence before mandate. Measurement rewritten before adoption is expected. Fear addressed out loud. And governance built while moving, not bolted on afterwards.

An organisation still living in pilots can test itself against that order. Who is already experimenting without being asked, and what would it take to fund them before the outcome is clear? Does the performance review reward the way work is done now, or the way it was done three years ago? Has anyone said out loud what will and will not happen to people’s jobs? The answers cost nothing and show where a program will stall. Every item on the list is a management decision, but none of them come with the software.

About Axtria

Axtria helps life sciences companies harness the potential of data science and software to improve patient outcomes by connecting the right therapies to the right patients at the right time. The company is a leading global provider of award-winning cloud software and data analytics to the life sciences industry. We’re proud to deliver proven solutions that help pharmaceutical, medical device, and diagnostics companies complete their journey from data to insights to action, enabling them to earn superior returns on their investments. As a participant in the United Nations Global Compact, Axtria is committed to aligning strategies and operations with universal principles on human rights, labor, environment, and anti-corruption, and taking actions that advance societal goals. For more information, please visit www.axtria.com.