What makes a company successful at using AI?

Article originally appeared in Harvard Business Review, 28 February 2022

By Vijay D'Silva and Bruce Lawler

What's the article about

Companies in a wide range of industries are trying to integrate analytics and data in forecasting, maintenance optimisation, logistics, transportation and service provision to improve their operations. A McKinsey and MIT study of 100 businesses across a range of sectors explored how they use digital, data analytics and machine intelligence (MI) technologies, what they want to achieve and how they keep track of their progress. The study aimed to identify why some companies do so much better than others in terms of impact and how quickly gains are made when implementing machine intelligence.

Dive into the details

The study looked at 21 performance indicators across nine areas— strategy, opportunity focus, governance, deployment, partnerships, people, data execution, budget and results. Four levels of performance were identified, ranging from leaders, to planners on the cusp of success, to results-oriented executors whose efforts were more disparate, to emerging organisations just starting out in the process. The leaders were differentiated by having an overarching integrated plan for MI and their commitment and level of investment. More specifically, they outperformed the other categories in five areas: governance, deployment, partnerships, people and data availability.

At the governance level, leaders make MI a strategic priority. They are much more likely to have a defined process for the assessment and implementation of digital innovation, including dedicated centres of excellence to support the entire organisation, ensure standards and accelerate deployment. In this fast-moving space, the study pinpointed that leaders continually refine, improve and update their processes, whereas executors and planners often get stuck, limiting their ability to scale successfully.

Deployment is the second area where leaders do better. They apply MI more widely in all aspects of their operations and use more sophisticated approaches. One example of this was a company that used machine vision in their product quality assurance process to cut down the number of false rejects.

Third, the study identified that partnerships with academia, start-ups, existing technology vendors, and external consultants are common in firms seeking to implement MI. Leaders in particular work intensively with a wide range of partners in order to maximise implementation speed and learning. For example, a partnership between a semiconductor firm and MIT resulted in a novel MI quality-control process that meant engineers only had to review five percent of the enormous volume of process data instead of looking at all of it.

Involving more people in using advanced digital approaches is the fourth area that characterises leaders. They provide skills and resources at all levels of the company, including training frontline personnel in MI fundamentals. Leaders view the use of data and analytics as deeply embedded in how they operate, rather than keeping them siloed and restricted to a few employees.

Data availability or the democratisation of data is the fifth area where leaders stand out. Data is made more available to frontline staff, plus a lot of the data collected by the company is fed back to customers and suppliers. Such companies develop a data strategy early in the development process for new MI applications, which is critical to the effective use of analytics.

The takeaways

Successfully implementing MI increases efficiency and reduces costs. In terms of what works, governance, deployment, partnerships, people, and data availability are most effective when integrated into a playbook, often coordinated by a centre of excellence. Each company’s path will be different, but a starting point is to assess themselves across the nine dimensions. They can then develop a transition plan with medium-term targets, the main ones being skills development, investment capacity, and critical infrastructure.

Regarding investment capacity, any company can start with small steps using data and simple tools, but money is another area that sets leaders apart from the rest. The leaders in the study spent 30-60 percent more and expected to increase their budgets10-15 percent, while the others reported little or no planned budget increases. This has implications for widening gaps across companies.