Three ways organisations can use data to power AI initiatives that drive value

Authored by:

• Yogesh Dhimate | Senior Partner Solutions Architect, AWS
• Martin Ristov | Senior Partner AI Technologist, AWS

Through consumer-facing applications, we’ve all witnessed how powerful the latest machine learning models have become. What typically takes most technologies two to four years to achieve, generative artificial intelligence (AI) has accomplished in less than 12 months. And with the uprising of big data, the game-changing access we have to analytics has never been greater.

However, AI models are just one small part of the big picture; the end goal is not as straightforward as just having visibility over your data. It’s how you use that data in a way that drives meaningful business value.

Crucially, you may be implementing AI initiatives to supercharge value, but those initiatives need your best data to deliver tangible results. Let’s navigate that.

Garbage in? Garbage out.

These breakthroughs and their widespread adoption underline the fact that organisations that will be successful in building AI applications with real business value are those that will do so using their data.

While these solutions can help us advance how we work, they are very much an equitable toolset. Organisations everywhere have the same access to the same machine learning models.

The models are an efficiency enabler, but they are just one small piece of the AI puzzle. To drive a better customer and user experience, you need good customer data, whereby analysts can derive insights from massive datasets with efficiency and accuracy.

As goes the computing adage “garbage in, garbage out,” a poor or incorrect input is only going to produce a faulty output. The same goes for your customer data. A lack of “clean” or “good” data can block organisations from delivering meaningful customer experiences—and, as a result, leave them unable to drive value. This leads to a value void—where organisations embracing new technologies struggle to maximise the full impact they can have on their core operations strategies.

The culture crux

To put it simply, unclean data is the result of misaligned systems, people, and processes. It’s also when data is in different languages, formats, and frequencies, each set with its own privacies and regulations—such as the Health Insurance Portability and Accountability Act (HIPAA), the General Data Protection Regulation (GDPR), or the California Consumer Privacy act—about when and where it can be used or accessed. The consequence of unclean data can be significant. For example, it can lead to hallucinations that reflect bias or inaccuracy.

An example of this might be that when a business is acquired, and people and cultures merge, barriers are created, and inconsistencies appear. Even though newly formed workforces and onboarded customer bases bring rich ideas and cultural nuances, they can also generate misalignment.

In fact, beyond the avalanche of data this creates, the myriad of protocols, mechanisms, and data residencies also create data silos. It’s almost like trying to cook with ingredients stocked in completely different kitchens. When data isn’t centralised, how can your new AI-powered tool churn out the best quality outputs? To be completely frank—it can’t.

So, how do you strike the right balance between technology and cultural transformation that sanctions not only the implementation of AI but also gives your AI initiatives fuel to impart lasting change?

Three ways your business can power AI initiatives with good data

1. Cross-functional collaboration

In 2025, proof packs more punch than a hunch. With access to smart technology today, evidence-based action has superseded intuition-based action across all organisational levels. As highlighted in the “How Possible Happens” report by Infor™, the most productive organisations are now using data to create a culture of experimentation and as a common language to collaborate for improved customer experiences.

Additionally, Deloitte’s recent “Is your data AI-ready?” report outlined how successful AI starts with connecting key stakeholders to the data products your organisation deploys. This means policymakers for building components, data engineers for collecting and transforming data, and data scientists for analysing the results. For this kind of collaboration to work, increased communication and tracking across all organisation levels is crucial—not to mention standardised tooling and processes to ensure data hygiene and handling.

2. Complementary AI

On their own, both generative AI and predictive AI offer strong advantages. While the former generates content in mere seconds, the latter uses algorithms to predict and forecast future outcomes. Together with solid data foundations, the two can be used to drive value.

For example, organisations could use traditional machine learning models to predict what a customer may want to buy next—based on previous purchases—before using generative AI to present that prediction in a way that’s personalised to that user, creating a more meaningful experience that encourages customer engagement.

3. Culture of data

In the quest to drive value, building a robust culture around data within your organisation can’t be understated. After all, you don’t just want to make tech successful—you want it to deliver tangible business value. Encouraging widespread data literacy across the workforce encourages a quality benchmark for how data is stored, shared, and kept clean.

When data is made more accessible, and its impact in driving value is recognised more openly, it is more easily seen as a strategic asset. A strong data culture also includes the building of a solid data governance framework to ensure data quality, security, and compliance.

Implement, monitor, iterate

While data is important to power AI, there are common pitfalls of overthinking it, which can lead to a sort of analysis paralysis stage. The rate at which new and shiny tools, frameworks, and models are released is unlike anything the tech industry has seen before. It’s critical for organisations to not only embrace new solutions but ensure they deliver tangible organisational value. And there are ways to make sure that happens.

It all starts with a bit of experimentation. To get the most from this process, organisations should foster a culture where failure is acceptable. Failing in an enterprise setting can be tricky, but doing so in a low-expectations, experimental environment can do wonders for the culture, encouraging teams to push the boundaries of what’s possible. As leadership expert John Maxwell said: “Fail early, fail often, but always fail forward.”

Once you begin establishing some quick and easy wins, you’ll earn the trust of stakeholders to invest more in initiatives. Showing that value early on will really drive home that successes are there and ready for the taking—and celebrating those wins paves the way for higher effort but much higher value use cases.

Lastly, it’s important to take a step back. Look at where we are now on a much bigger timeline. From the very early days of AI development to where we might be in ten years, we’re currently at a crucial stage for defining industry standards around data. Beyond the value we can all unlock, clean data is paramount for minimising unintended bias, ensuring transparency, and protecting privacy—something we should all be striving for.

Uncover the full story and get practical advice on filling the value void in Infor’s Report:  “How Possible Happens.”