Most enterprise leaders ask the same question when AI becomes a board-level priority.
Do we modernize legacy systems first and then build AI on top of them? Or do we push AI forward now and somehow keep the old world running at the same time?
In this episode of CXO Spotlight, Chirag Khanijau sits down with Munir Hafez, CTO at Independence Pet Holdings, to unpack what it actually looks like to make that decision inside a live enterprise. Not in theory. Not in a clean-room strategy deck. In the middle of real operational complexity, M&A sprawl, and hard business tradeoffs.
Munir is leading one of the more difficult transformations in enterprise tech right now. Independence Pet Holdings has grown through acquisitions, which means inherited complexity at scale: multiple acquired businesses, multiple policy administration systems, deeply embedded business logic, and large operational dependence on legacy platforms that cannot simply be switched off. At the same time, the company is building toward an AI-native future.
That tension is what makes this conversation valuable.
A central idea from the episode is that “stabilize first, then innovate” sounds sensible but often kills momentum. If the same people, same priorities, and same incentives are asked to both protect the current revenue engine and build the future, the future usually loses. Munir’s answer is bifurcation: create two parallel operating motions, one focused on protecting business-as-usual and another focused on building the new architecture.
It is not the cheapest short-term move. But it may be the fastest path to real transformation.
That decision becomes even more interesting when he explains how he thinks about people. Instead of treating legacy talent as an obstacle, he frames tribal knowledge as an advantage. The teams who understand the old systems best are often the ones holding the business together. Rebadging, redeploying, and restructuring roles can preserve institutional knowledge, protect morale, and create space to fund the next stack without defaulting to blunt-force layoffs.
Another sharp insight in the episode is Munir’s distinction between “blue money” and “green money.” Blue money is the soft value most AI projects claim: better productivity, better experience, faster decision-making, more insight. Green money is what actually hits the P&L. His argument is direct: most AI projects never make the jump from soft benefit to measurable dollars. That is where many enterprise AI strategies fall apart.
The implication for CIOs and CTOs is clear. If a project cannot be tied to cost removal, revenue growth, risk reduction, or some other tangible business lever, it may be interesting, but it is not transformation.
He also makes a point that deserves more attention in enterprise tech: copying bad processes into modern tools does not create innovation. It just creates expensive complexity. That is true for AI. It is true for ERP. It is true for every transformation program that mistakes tooling for operating change.
One of the strongest sections in the conversation is around strategy communication. Munir argues for a one-page strategy instead of twenty-page presentation decks. Why? Because transformation only works when people understand the direction well enough to act on it. The goal is not documentation. The goal is alignment.
That becomes especially important in organizations where technology leaders are trying to bring boards, executive peers, and delivery teams onto the same page while moving at speed.
The episode also goes beyond systems and operating models into talent. Munir offers a practical view of how developers, data scientists, and IT professionals should think about staying relevant as AI commoditizes more of the execution layer. The takeaway is not panic. It is repositioning. Technical professionals who understand systems, business context, and process redesign will stay valuable longer than those who only optimize for narrow tool familiarity.
This is a serious operator’s conversation. It is about modernization, yes, but also about sequencing, incentives, organizational design, and the uncomfortable choices that real transformation requires.
Why you should watch: A lot of enterprise AI content still assumes the company is starting clean.
Munir Hafez is operating from the opposite reality: acquired systems, fragmented platforms, embedded complexity, and a business that cannot stop while transformation happens. That is the world most enterprise leaders actually live in.
This episode gives a more honest playbook for how transformation works when the stakes are real.
Munir explains:
- Why running legacy and AI-native organizations in parallel can be faster than sequencing them
- What most leaders get wrong about “stabilize first, then innovate”
- Why tribal knowledge is an asset, not just a transition risk
- How rebadging talent can preserve morale and protect transformation speed
- What “blue money” versus “green money” means in AI programs
- Why most AI projects never reach true P&L impact
- The three common failure patterns in enterprise transformation
- Why tools fail when broken processes are left untouched
- How a one-page strategy creates alignment faster than long presentations
- What developers and data scientists should focus on as AI changes the value stack
🎧 Listen on Spotify · Apple Podcasts
🔗 Follow
Chirag Khanijau
