Center for Continuous Intelligence

How to Speed AI Deployment to Achieve CI Benefits Faster

PinIt

Companies need to build up AI skills and overcome cultural challenges to its use. The key to accomplishing these objectives is to do many projects quickly.

Continuous intelligence (CI) derived from streaming data often requires artificial intelligence (AI) to derive instantaneous insights upon which to act. That’s especially the case in situations such as autonomous systems or real-time fraud detection where the timeframe to act is in the milliseconds to minutes range.

See also: The Case for Continuous Intelligence

Unfortunately, many companies have difficulty embedded AI into their business processes. They incur start-up delays in implementing AI and acquiring the institutional knowledge about AI to make practical use of it. The sooner a company gains confidence in AI and applies it to CI, the faster they will realize CI’s benefits.

How do you accelerate AI adoption and build up the needed expertise to apply AI throughout the organization? Industry experts recently have offered some advice.

Leading AI anywhere proponent Rob Thomas, General Manager, IBM Data and AI, notes that you not only have to overcome a lack of skills, lack of tools, and lack of confidence, etc. But the biggest issue is cultural. He recommends:

“The most important thing is a beginner’s mindset, a willingness to try, and an acceptance of failure. Organizations should seek to do 100 AI experiments a year, knowing that more than 50% will fail. Many company cultures are not suited for that. A more typical approach is to rally around one big AI project, committing a lot of people, time, and money. I do not advise that approach. AI is about mass experimentation, not one big project implementation. This ain’t ERP.”

As companies adopt such strategies, he expects AI anywhere to become the norm. 

Embrace Small Data

A recent Harvard Business Review (HBR) article touts a comparable idea. The article notes the easiest way to get the benefits of AI is to start with smaller projects focused on small data (not big data).

Small data projects succeed by overcoming the cultural issues Thomas cited. They also let companies do many more projects quickly, thus building up the needed expertise to eventually adopt enterprise-wide. According to the article: 

“Unlike big data projects, which often involve dozens of people with disparate agendas, politics, enormous budgets, and high failure rates, the probability of success is high. Thus, small data projects build the organizational data muscle that helps the entire company learn what it takes to succeed with data, gain needed skills, build confidence, and breed the kind of culture that big data demands.”

“Small data projects involve teams of a handful of employees, addressing issues in their local workplaces using small data sets — hundreds of data points, not the millions or more used in big data projects. They are tightly focused and utilize basic analytic methods that are accessible to all. They can be completed in a few months by people working part-time and yield financial benefits of $10,000 to $250,000 annually per project. Companies are loaded with potential small data projects, and it is reasonable to expect that a 40-person department to complete 20 projects a year. The cumulative benefits are enormous.”

Lessons Learned

Essentially Thomas and HBR give the same advice. Companies need to build up AI skills and overcome cultural challenges to its use. The key to accomplishing these objectives is to do many projects quickly.

Certainly, there will be failures along the way. But a failure in a small project does not consume the dollars and resources as would be the case if there was a failure in a single large project.

The best strategy is a bit analogous to the case of new drug development in the pharmaceutical industry. With the cost to bring a new drug to market estimated to be $2.6 billion, eliminating poorer performance candidates in early research and development stages avoids the high costs of clinical trials, FDA approval processes, production, and marketing.

With AI for CI, it is about the same. Fail early and fail often. “Fortune favors the bold. I believe that the trial and error all have gone through – and will continue to go through – is worth the positive outcomes, said Thomas.

Salvatore Salamone

About Salvatore Salamone

Salvatore Salamone is a physicist by training who has been writing about science and information technology for more than 30 years. During that time, he has been a senior or executive editor at many industry-leading publications including High Technology, Network World, Byte Magazine, Data Communications, LAN Times, InternetWeek, Bio-IT World, and Lightwave, The Journal of Fiber Optics. He also is the author of three business technology books.

Leave a Reply