Investments in AI are wasted without a solid data foundation

Artificial intelligence and it’s offshoot, machine learning are promising to be nothing short of a revolution for business. These new technologies allow organisations to gain competitive edge and deeper insights, in turn contributing to consistent increases in performance and productivity.

As it’s reached a new level of mainstream acceptance in recent years, AI should be included in one or more digital strategies for every enterprise. At the same time, not every investment into AI and machine learning applications will yield the same results. The variance we see is often driven by the organisation’s capabilities for collecting, processing and analysing data before AI is actually introduced.

For the full transformative capabilities of AI to be realised within an enterprise, data scientists spend up to 80% of their time behind the scenes wrangling with data to create the necessary data foundation. To modernise your data strategy to the point where you can expect the best returns on AI and machine learning investments, you need to start by addressing the following key points:

• Are you capturing all of the raw data available from your most vital business processes?
• What level of unstructured data lies outside of your current databases in terms of documents, images, video, audio and system logs?
• Can you bring together the disparate data silos across your organisation, as well as in off-premise SaaS and cloud environments, to create an up-to-date and integrated view?
• Do you have the capability to query your raw data to answer vital new questions as they arise?
• As opposed to analysing historical data, can your processes gain access to real-time streaming of data?
• Can you democratise your data across the business so that everyone has the ability to query data, not just data scientists?
• How flexible is your data architecture? Can your strategy innovate and respond quickly to changing market conditions?

Once the above questions have been addressed, enterprises need to begin experimenting with smaller AI use cases now, so they can better understand the necessary steps to prepare for larger projects in the future. The organisations to achieve the greatest level of success will be those that experiment with versatile machine learning applications that can lead to process improvements and enhance productivity in the business.

The investment focus should be on initiatives that takes market share away from competitors; improving customer experience while unlocking and engaging new segments of customers. But deploying these types of initiatives takes considerable time and effort, so it’s advised to begin with smaller projects that provide measurable benefits and ROI.

The largest area of potential we see in using machine learning is to simply “follow the money” meaning to focus on the areas where it returns the most value e.g. in sales and marketing for retailers and predictive maintenance in manufacturing or better forecasting in supply chain.

Ultimately, AI and machine learning applications require access to large amounts of easily accessible and well-integrated data to automatically learn and improve from experience, and make decisions with minimal human intervention.

About the author:

As the General Manager of InfoCentric, my team and I provide specialised expertise, advisory and implementation services in digital and data analytics to help our clients convert data assets into business insights for competitive advantage with faster, smarter and more accurate business decisions. If you’d like to discuss how your organisation can leverage your data assets to drive operational efficiencies, attract new business, and minimise risk, please feel free to get in touch with me at [email protected].