By Mohsin Pervez, Associate Vice President, Technical Account Management (EMEA), PubMatic
Around 91% of businesses now say they use AI in some capacity, yet the same McKinsey research suggests only 1% consider themselves truly mature in deploying it. I’ve been at PubMatic for more than 15 years and have seen plenty of technology shifts along the way. I’d argue this gap is the real story of AI in the workplace. Technology is everywhere. Genuine readiness is not.
Closing that gap requires a deliberate, structured programme to make sure every part of the organisation can use it meaningfully, not just simply talk about it.
If I had to summarise what AI-readiness actually requires, it comes down to three things: bring the whole business with you, make AI genuinely useful in daily work, and ensure leaders actively champion adoption.
Technology matters, but the way a business approaches AI matters more. Success depends on clear objectives, cross-functional collaboration, and a culture where experimentation and feedback are encouraged. Here’s what worked for us.
Get everyone involved
Around 10 months ago, PubMatic formed a Global AI Council to ensure adoption was not led solely by the engineering team. The council includes representatives across EMEA, the US and APAC, spanning various teams across functions: customer success, demand, marketing and technical operations, engineering and leadership.
Balanced representation, including gender diversity, was a priority from the outset, and that mix of perspectives has been one of the biggest drivers of progress.
For example, a marketing colleague spots different friction points than those I see, while a customer success expert better understands what their teams need day-to-day. Different regions also bring different operational realities. Without that range of input, it would be far harder to create tools and training programmes that deliver benefits across the business.
Diversity not only relates to areas of expertise but to seniority as well. This matters because of what BCG calls the “silicon ceiling”: research shows around 75% of leaders use generative AI several times a week, but only 51% of frontline employees do. Adoption can’t progress if it stops at the top.
Make AI useful from day one
One of the earliest priorities was integrating AI into our internal workflows.
PubMatic custom-trained AI models based on company documentation, allowing employees to query internal knowledge securely rather than solely relying on public tools. For instance, I can ask about a technical protocol and receive answers based on our own expertise built over the years, instead of generic information.
We also introduced an AI layer across our internal systems, including third-party CRM tools. That means someone in marketing can understand what the customer success team has been working on, unlocking levels of cross-functional visibility that once required manual requests or waiting for reports. Now it takes seconds.
Across marketing, AI is used throughout the workflow: refining communications for different audiences, creating assets, activating and optimising campaigns, and measuring outcomes. AI planning agents are being developed to support prioritisation across go-to-market teams. What were once disconnected stages now feel far more joined up.
Give everyone a voice
One of our best decisions was opening the door for everyone to suggest opportunities for improvement. We now have hundreds of ideas logged from across the business. Each is assessed against multiple criteria, for example, efficiency gains or cost savings. If it meets the majority of the target criteria, stakeholders progress it forward.
The value of this is that innovation is shared across the organisation, not confined to one team. For example, some of the most useful suggestions have come from colleagues in HR and finance, people who are not in the weeds of the technical side of things day-to-day.
Leadership has to lead
AI adoption cannot be left to individuals to figure out alone. A significant part of our progress has come from leadership making AI upskilling a priority and embedding it into team objectives. That top-down commitment is what allows a pilot programme to become a standard way of working.
This is the biggest lesson for any organisation looking to accelerate AI adoption: leaders have to drive it, and they need to be using the tools themselves. Industry estimates suggest that 70%-80% of AI initiatives fail, and the cause is rarely the technology. Success is almost always a matter of change management.
Tackle concerns head-on
One common reaction we’ve seen across departments and levels is the belief that using AI is somehow cheating. If AI helps structure a document or draft communication, then the concern is that the work is no longer yours. We decided to address that directly.
Using AI to write an email is fine, provided the facts are accurate, and you have checked them before sending. Using AI to analyse account signals is no problem, provided you are still making the judgment call on what happens next.
AI simply increases the scale and speed at which human judgment can be applied; it does not devalue or downplay human involvement.
We’ve been clear that our AI initiatives are designed to remove repetitive tasks so people can focus on higher-value or more meaningful work, which requires their specialist skillset. Additionally, our legal team has been involved from day one, ensuring that governance and policy are not only defined and shared, but cruciually, are understood across the organisation.
What comes next
Daily use of custom-built, AI-enhanced workflows and processes will be standard practice across our global workforce by year-end, and EMEA is tracking nicely against the ambition.
But the bigger ambition is to create an organisation where using AI becomes as instinctive as using any other workplace tool, and experimentation is normalised.
The training we ran nine months ago has already evolved, and our tools continue to improve. Training is, de facto, ongoing and it is coupled with sharing best practices and new approaches, as well as failures. That progress did not start with technology. It started with people, culture, and readiness.
