AI became the number one topic in boardrooms, but very few companies know how to move from talk to action
Digital transformation has become a mandatory agenda item in strategy meetings at major companies. It doesn’t matter the industry, the size of the operation, or the geography — everyone wants to understand how artificial intelligence can accelerate results, cut costs, and open up new growth opportunities. But between the desire to adopt AI and the reality of making this technology work day to day, there’s a gap that most people still don’t know how to bridge.
Irfan Khan, President and Chief Product Officer at SAP Data and Analytics, noticed something interesting inside some of the world’s largest organizations: there’s far more anxiety than action when it comes to making artificial intelligence actually work in everyday business operations. According to him, this holds true for large, mid-size, and small companies alike. The picture looks pretty much the same across the board.
And the villain in this story is almost always the same: data scattered across different systems, lacking structure, missing context, and hard to access. Critical business information gets trapped in organizational silos, duplicated in spreadsheets nobody updates, and stored on legacy platforms that don’t talk to each other. This scenario turns any attempt to implement artificial intelligence into an exercise in frustration.
The situation is serious. A recent SAP survey found that only 33% of business leaders trust their own data. That lack of trust creates a massive gap between intention and execution. On top of that, organizations that have already matured their data infrastructure are three times ahead in terms of AI readiness, according to the same study.
And as 2026 moves forward, this gap is going to split companies into two very distinct groups:
- Those that built a solid foundation to operate, compete, and grow with AI
- Those that just kept talking and will face real consequences for it
The good news is that it’s possible to change this picture right now. Khan outlined four practical moves that any company can start applying today to get ready for the age of artificial intelligence. 👇
Why most companies still stall when it comes to using AI
Before getting into the steps themselves, it’s worth understanding why so many companies are still spinning their wheels on this transition. The answer isn’t a lack of available technology or a lack of willingness from leadership. What holds most organizations back is something far more basic and, at the same time, far harder to fix: the quality and accessibility of the data they already have.
Legacy systems that don’t talk to each other, duplicate information scattered across spreadsheets in different departments, and processes that were never designed to feed an artificial intelligence layer — that combination paralyzes any initiative before it even gets off the ground. The data exists, and in massive volumes, but it’s fragmented in a way that makes it practically useless for AI models that need consistent, contextualized, and real-time accessible information.
Khan likes to say that AI is only as good as the data that feeds it, and that line sums up the core problem companies face today. There’s no point in deploying the best language models, investing in cutting-edge cloud infrastructure, or assembling a team of data scientists if the raw material — the company’s own data — arrives dirty, incomplete, or contradictory. The result is predictable: unreliable outputs, decisions based on noise, and growing distrust in the results the technology delivers.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
This cycle of distrust is especially dangerous because it feeds itself. When leaders don’t trust the data, they stop investing in improving its quality. When quality doesn’t improve, AI keeps delivering questionable results. And when results are questionable, the perception that the technology just doesn’t work for that particular business takes hold — even though the real problem was never the technology itself. Breaking this cycle requires a structured approach, and that’s exactly what the following four steps lay out.
The 4 steps to make your company AI-ready right now
Embrace hybrid cloud without leaving your data behind
Despite years of cloud migration, many companies still aren’t fully cloud-native — and probably won’t be anytime soon. Between sunk costs in legacy systems and regulatory constraints, hybrid environments are an unavoidable reality. And that’s perfectly fine, as long as the company knows how to handle that complexity the right way.
The key insight, according to Khan, is that AI doesn’t necessarily require a complete break from on-premises systems. What truly matters is whether the company can access its data consistently and securely, regardless of where it’s stored or where the processing happens. In his words, you need to be able to access data independently of where it or the runtime environment operates.
That’s why SAP has been investing heavily in data fabric — a critical layer that connects raw data sources to artificial intelligence applications. This approach helps provide the business context that AI agents need to actually work, without requiring the company to abandon its existing infrastructure overnight.
A practical example comes from Yamaha, a corporation that operates across segments as diverse as musical instruments and motorsports. That product variety creates an enormously complex data ecosystem. With SAP Business Data Cloud, Yamaha was able to connect product information, supply chain operations, financials, and customer data into a unified layer, gaining the agility to move fast in the age of artificial intelligence.
Build a database designed specifically for AI
Traditional databases simply weren’t built for the kind of reasoning and inference that AI demands. That’s why companies need to take action and implement a data foundation that’s genuinely prepared for artificial intelligence. In Khan’s view, an AI-ready database needs to support multiple data models simultaneously — whether that’s vectorization, knowledge graphs, real-time processing, or any combination of them.
That’s the reason SAP adopted a multimodal design approach. With it, companies don’t need to worry about the volume or temperature of their data. Instead, they’re free to use AI systems that query data directly, without the need to copy or duplicate it. Khan emphasizes that data federation is an intrinsic and fundamental capability of any AI-based data offering.
The Chilean company Martinez and Valdivieso, which supplies agricultural inputs, is a good example of how this works in practice. By implementing SAP HANA Cloud — a multimodal database that’s part of SAP Business Data Cloud — the company equipped its field sales reps with an AI-powered solution that lets them customize quotes in real time. The solution triangulates information from multiple on-premises and cloud sources, enabling sales reps to offer smarter quotes from anywhere, improving both customer experience and profitability at the same time.
Focus your efforts where they actually move the needle
With the pressure to implement AI across every front, many companies end up spreading their efforts too thin. Khan warns that this is a mistake. Instead of trying to solve everything at once, organizations should start at the use-case level — identifying the areas of the business where uncertainty and risk are highest and, from there, deploying artificial intelligence in a meaningful and targeted way.
Supply chains are a classic example of an area with enormous potential for measurable results. Most companies struggle to forecast supply and demand in order to get ahead of spikes or disruptions. This is a problem tailor-made for AI. As Khan explains, it’s no longer about waiting to predict that something is happening — it’s about genuinely sensing when something is going to happen, before it materializes.
That same logic applies to finance, workforce planning, operations, and other critical areas of the business. By combining predictive models with this real-time sensing capability and then automating the response, companies can shift from a reactive posture to a proactive one. Instead of reacting to problems after they happen, they start anticipating scenarios and acting preventively.
Rethink teams and operating models for an AI-native world
Artificial intelligence is rapidly changing the way work gets done, and Khan says companies need to move fast to align their operating models with this new reality. In the short term, he sees a kind of buddy system where humans and AI agents work side by side. The agent handles repetitive tasks like testing or quality assurance, while the human focuses on creativity and oversight.
In the long run, the shift toward AI could put even more pressure on existing operating models. Software development teams that traditionally run on waterfall or even agile methodologies, for example, will need to adapt to deliver results dynamically. This need for speed and adaptability brings Khan back to his original point: companies need to cut their losses with data management systems that no longer serve them and implement a cloud-based or hybrid infrastructure that enables this agility.
Khan’s closing message is direct and powerful: if you’re still looking for a business case to justify AI adoption, you’re going to get buried under business-case logic while the world moves on at a much faster pace. Companies need to embrace the mindset of experimenting fast, failing fast, learning, and iterating — treating AI as something you build in practice, not just in presentation decks.
Culture matters just as much as technology
Technology without culture goes nowhere. It doesn’t matter if you have the best cloud platform, the most well-organized data, and the most sophisticated AI models on the market — if the people inside the organization don’t trust what the systems show them and keep making decisions based on gut feeling or side-channel information floating around in emails. Building a culture of data trust is a process that starts with transparency: teams need to understand where the numbers come from, how they’re calculated, and why they’re reliable.
An essential element in this effort is data governance. This involves defining who’s responsible for each dataset within the company, what quality standards need to be met, how errors are identified and corrected, and how updates are communicated to the people who depend on that data to do their jobs. When there’s a clear governance process in place, people start seeing data as something trustworthy and managed — not something that could be outdated or wrong without anyone knowing.
The cultural transformation also involves training and education. People who understand how AI works, what its limitations are, and how to interpret the results it delivers are far more likely to trust the technology and use it effectively than those who see artificial intelligence as some kind of magical black box. Investing in data literacy and hands-on training for teams isn’t an extra cost on the digital transformation journey — it’s a fundamental part of the path.
Measuring results and iterating consistently closes the loop
Launching an artificial intelligence solution and expecting it to keep running smoothly forever without adjustments is wishful thinking. Both the business environment and the data feeding the models change constantly. Companies that understand this create regular cycles of evaluating results, identifying deviations, and updating models — treating AI as a dynamic asset that needs ongoing maintenance and evolution.
Defining success metrics before launching any AI initiative is essential for making that evaluation possible. These metrics need to be tied to the business objectives the solution is meant to address: reducing time in a specific process, increasing the accuracy of demand forecasts, decreasing errors in particular operations, improving customer satisfaction. Concrete, measurable numbers linked to real business impact are what separate a purposeful AI initiative from one that was launched just to check a box on a report.
Cloud infrastructure plays an important role at this stage because it lets companies scale resources up or down based on what the results show, without requiring large fixed investments in hardware. This flexibility is especially valuable in the iteration process, because it allows teams to test hypotheses, adjust approaches, and experiment with new configurations without the financial burden that kind of experimentation would carry in traditional environments. Companies that combine rigorous measurement with flexible infrastructure have a much better chance of turning their AI initiatives into sustainable competitive advantages.
What separates the companies that get there from those that fall behind
At the end of the day, the difference between organizations that manage to make AI work in their favor and those that stay stuck in the planning phase isn’t about available budget, the size of the tech team, or access to the most advanced tools on the market. It’s about the willingness to tackle the problem at the root — which always starts with data quality and organization — and the ability to connect every technology decision to a concrete business outcome.
The cloud has democratized access to tools that used to be the exclusive domain of the world’s largest corporations. Today, companies of all sizes can tap into scalable infrastructure, advanced analytics platforms, and state-of-the-art AI models without having to build anything from scratch. What’s still scarce — and has therefore become the real competitive differentiator — is the ability to use all of it strategically, in a connected way, and driven by trustworthy data.
The four steps Khan described — embracing hybrid cloud, building databases designed for AI, focusing on high-impact use cases, and rethinking operating models — form a roadmap tested at some of the world’s largest organizations. They’re not a magic formula, but they are the most consistent path for breaking out of the anxiety cycle and entering the action cycle for good.
2026 is shaping up to be the year when the stakes get higher and the differences between prepared companies and those that fell behind start showing up in financial results with much greater clarity. The time to act is now. 🚀
