Key Takeaways
- Focus on Pain Points, Not Hype: Think what AI can do for your business rather than just going with the trend.
- Fix Your Data Before Scaling: Build clean data foundations first, measuring pilots with adoption metrics and financial KPIs.
- Focus on Change Management: Overcome job security fears by framing AI as a tool that automates manual work, freeing up time for strategic focus.
Everyone everywhere wants to have AI.
It is the ultimate corporate buzzword. But beyond the buzz, leadership teams across organizations have to face the truth and answer some hard questions: do we actually need AI? Will we be able to get a real RoI, or are we doing this just because everyone else is?
I was part of CDO Vision 2026 in Tokyo, where I had the opportunity to moderate a panel of tech and business leaders, including Manzur Mahtab (CIO, Carrier HVAC Japan), Moran Gelber (CIO & IT Director, Volkswagen Group Japan), and Ka Ra (Head of Data Governance & Data Science, Siemens Healthineers Japan), all of whom are facing this exact problem. We discussed how organizations can cut through the noise, align on AI strategy, and turn pilot projects into measurable enterprise value.
Here are the takeaways from our discussion.
1. Shift the Focus from “What AI Can Do” to “What Pain Point Needs Solving”
Once a new tool appears in the market, there is a rush where every organization wants to get their hands on it. One thing that clearly resonated across the panel was that before rushing to invest in a new tool just because, the leadership must first analyse if they actually need the tool and if it fits in with their real business requirements.
When finance teams and CFOs demand to know where the highest ROI lies, the simplest answer isn’t a complex financial model. It is identifying your biggest operational hurdle.
- Is it a high-volume, manual process that’s taking up your back office?
- Is it an outdated legacy system forcing employees into endless copy-pasting?
- Is it a breakdown in customer support communications?
When you identify and remove a massive operational bottleneck, the ROI naturally follows, whether through direct cost reduction, time savings, or reallocating resources to important tasks.
2. The Two Pillars of AI Adoption: Systems vs. People
AI implementation is best monitored when your efforts are split into two different cohorts.
- Embedding AI into core systems and products: This includes predictive maintenance, smart diagnostic tools, and automated workflows. Though technically complex, measuring the return here is straightforward, as it either drives new revenue or cuts downtime.
- Embedding AI into daily employee workflows: This is the longer, harder journey. Simply handing out licenses for enterprise AI tools without structure or training will not generate value.
To bridge the gap between human capability and AI potential, organizations are realizing the importance of creating mandatory training programs to ensure staff understand proper usage, guardrails, and data ethics before getting tool access.
3. Build a Solid Data Foundation Before You Scale AI Pilots
The prerequisite for any kind of AI adoption is to ensure that your data is clean and unbroken.
Which is why, before jumping into enterprise-wide AI adoption, you must prioritize data hygiene, governance, and integrity. Investing heavily in an AI pilot without a ready data foundation creates massive business risk and wastes money.
When evaluating proofs of concept (POCs), use a combination of evaluation criteria:
- Leading KPIs: Metrics like engagement rates, user adoption, and tool performance to test initial potential.
- Lagging KPIs: Concrete metrics like direct revenue growth, labor hours saved, and cost reduction to prove true business value.
If an AI pilot meets its target payback window (often targeted within 12 to 24 months), it earns the green light and funding to scale into production.
4. Overcoming the Fear of Enterprise AI on the Ground
The most important thing to keep in mind while embarking on an AI adoption journey is that scaling AI is more a human challenge than a technical one.
When you move from a controlled pilot to enterprise production, you will tend to come across anxious long-tenured employees who worry that AI will take over their job. The best way to overcome this is to go directly to business unit leaders at the top, rather than to enforce the change through HR policies.
Show them how AI agents and automated workflows are designed to free their time and effort from tedious, manual back-office tasks. When employees realize AI is there to help them focus on strategic work, sales engagement, and creative problem-solving, adoption shifts from resistance to enthusiasm.
Looking Ahead
Three decades ago, organizations probably debated whether investing in workplace computing or the internet was worth the cost. Today, nobody asks for an ROI calculation on internet expenses or electricity – it has become the baseline requirement to run a business.
AI is moving in the same direction. In a few years, asking whether an organization “needs” AI will sound absurd. But today, as we navigate this transition, success belongs to the leaders who stay away from the hype, ground their strategy in clean data, and work to solve real human and operational problems.
Catch the full conversation here
