How far are you on the path to becoming a data-driven company?

• Many companies aspire to continuously generate value and competitive advantage through data and artificial intelligence (AI) technology.

• What does this journey towards being a data-driven company look like? And what barriers do companies face along the way?

The potential of data and artificial intelligence is vast and beyond dispute. By 2030, AI technologies is expected to contribute around $15.7 trillion to the global economy. Most companies invest in data and AI projects, but many report that the potential of these projects is only partially exploited. 
 
We surveyed over 110 international businesses in a study to understand how decision-makers view the potential of data and AI, how far companies have advanced towards becoming a data-driven company, and what obstacles they are encountering. 

5 typical barriers companies face in becoming data-driven

1. An inactive data innovation pipeline  

A key element for a data-driven company is an active data innovation pipeline. This refers to the continuous planning and implementation of data and AI projects that generate value for the organisation. However, this pipeline is not active in many companies due to the misalignment between business goals and technology or the lack of a defined strategy.  
 
43.6% of respondent companies thought that AI initiatives failed because there was too little consideration of business success or too much emphasis on technology. The second-most common driver was the lack of a company-wide data and AI strategy (33.6%).  

2. Proof-of-concepts falling by the wayside 

Proof-of-concept (PoC) studies act as a checkpoint where the feasibility of projects is reviewed. Many projects remain stuck at this phase due to a narrow PoC or insufficient expertise for operationalising use cases. 
 
32.7% of respondent companies indicated that their AI initiatives faced problems associated with PoCs. The most frequently cited reason (40.9%) was the poor quality of data and models.  
 
In an example at a plant construction company, we found a lack of infrastructure for operationalisation. Although the project team developed good models, it was unable to progress beyond the PoC as there was no direct access to the customer’s systems. Creating access would mean integrating multiple systems across the globe, costing millions. 

3. Technically perfect solutions are not used as planned 

A user group may fail to use an application as planned. This may be caused by a solution that is insufficiently tailored to real-life practice, prejudice against AI solutions or user experience difficulties.  
 
Failing to integrate AI solutions was the most common reason for the users failing to adopt the solution (52.7%). Further reasons cited were insufficient training of end users and distrust towards AI solutions. 
 
In one project with a vehicle manufacturer, a pilot MVP AI solution was developed to forecast demand for replacement parts. However, it emerged that the predicted parts were either so small that inventory was always available on hand anyway or so large that they had to be ordered. This meant that the solution delivered no added value for its users. 

4. Inadequate technical competencies 

Technical competencies represent a major barrier for successful implementation of data and AI projects. Some companies are unclear about what skills are required.  Others may have the right job profiles, but their employees are unable to apply their skills because of sub-optimal structures.  
 
55.5% mentioned that the lack of collaboration was a major issue. This shows the importance of deploying in-house data consultants who have experience in implementation projects and can build bridges between teams.  

How far have companies progressed on their journeys?

We identified 3 levels of maturity during a cluster analysis to understand how these companies progressed to becoming a data-driven company. 
 
1. "Innovation" – They face challenges across all five barriers  

These companies recognise the potential and significance of data and AI. However, there was a lack of holistic implementation concepts.  
 
2. "Foundation" – They experience challenges in the last two barriers 
 
These companies can be found further along the journey, but they face challenges in establishing foundations commonly associated with data-driven companies. 
 
3. "Data-driven" – They can already be described as data-driven  
 
These companies are the most advanced along the journey to a data-driven organisation but continue to anticipate an uptrend in data analytics and AI use in the near future. 
 

Becoming a data-driven company

From the barriers identified, we derive the following success principles for data and AI projects: 

• Company-wide planning and orchestration 
• Business orientation 
• Early user involvement of AI solutions 
• Fast and agile execution, plus a willingness to learn as an organisation 

Organisations can incorporate these principles within a data-driven company framework in 3 main steps:  

1. Determine the vision at the C-Level  

Set up a core team at the management level to align decision-making, develop a strong vision and formulate outcomes that the company aims to achieve. This ensures that the whole company can be aligned on this journey. 

2. Define the data strategy and establish AI portfolio management  

Develop a data strategy using the company strategy as a starting point, then operationalise it through a portfolio of concrete projects and initiatives. Furthermore, a portfolio process in place can function as an impulse generator for the project pipeline. 

3. Create the foundations on an ongoing, incremental basis while implementing value-adding solutions  

Establish the foundations step by step while projects are implemented in parallel. These foundations include data platforms, structures and processes that are based on concrete use cases and fit for real-world use. For initial implementation, choose use cases with sound prospects for success. These projects will serve as a beacon and strengthen internal acceptance of the company’s transformation. 

“The era of experimentation in AI is over. You expect to see tangible results from AI investments. And we believe that an integrated and holistic approach is the way to achieve this.” 
 
By laying the right foundations and fostering an open culture, companies can progress on their journey to be more data-driven, paving the way for organisations to operate efficiently, compete effectively and navigate the digital future.  
 

This article was originally published in Issue 90 of the Orient Magazine, Official Magazine of the British Chamber of Commerce in Singapore.