Applying AI in the real world

How to actually apply cognitive technologies (AI) in the real world? How to make projects successful? And how to scale them across the company? The answer appeared to be a lot closer to my daily practice than expected.

In a previous blog, we explored the three types of AI. Process Automation, Cognitive Insight and Cognitive Engagement. In their research, Davenport & Ronanki also examined patterns of (un)successful AI implementations. They see an increased use of cognitive technologies to solve business problems. But the most AI projects still fail. Davenport & Ronanki advise organizations should take an incremental approach and focus on augmenting rather than replacing human capabilities.

Understand the technologies

Before launching an AI initiative, organizations must understand which technologies perform what types of tasks. And recognize the strengths and limitations of each technology. Otherwise AI initiatives could result in wasted time and money pursuing the wrong technology for the job at hand.

Having the right capabilities is also essential to progress. In particular, organisations will need to leverage capabilities of key employees, like data scientists. A main success factor is their willingness to learn. If you don’t have the data science or analytics capabilities in-house, you’ll probably have to build an ecosystem of external service providers in the near future.

Creating a portfolio of projects

Systematically evaluate needs and capabilities and then develop a prioritized portfolio of projects. Determine which areas of business could benefit most from cognitive applications. Ask yourself which cognitive applications would generate substantial value and contribute to business success. How critical is it for your overall business strategy? Prioritize accordingly. Select the right technology for now. Davenport & Ronanki also encourage to take incremental steps with the currently available technology while planning for transformational change in the not-too-distant future.

Launching pilots

Because the gap between current and desired AI capabilities is not always obvious, companies should create pilots for cognitive applications before rolling them out across the entire organization. If your firm plans to launch several pilots, consider creating a cognitive center of excellence to manage them.

Think of how workflows might be redesigned, focusing specifically on the division of labor between humans and the AI. Systematic redesign of workflows is necessary to ensure that humans and machines boost each other’s strengths and compensate for weaknesses. Most cognitive projects are also suited to iterative, agile approaches to development.

Scaling up

Many organizations have successfully launched cognitive pilots, but they haven’t had as much success rolling them out in the entire organization. To achieve their goals, organizations need detailed plans for scaling up, which requires collaboration between technology experts and owners of the business process being automated. Make sure your business process owners discuss scaling organisations with the IT organization before or during the pilot phase.

Through the application of AI, information-intensive domains could become more valuable and less expensive to society at the same time. The great fear about cognitive technologies is that they will put masses of people out of work. However, cognitive systems perform tasks, not entire jobs. Repetitive, administrative work could be allocated to machines, freeing up human workers to be more productive and creative.

Read the full article of Davenport & Ronanki in Harvard Business Review.