This article was co-authored with Andy Thurai, vice president and principal analyst with Constellation Research.
Artificial intelligence can be a beautiful thing for business, with a lot of promise. But this promise has yet to deliver tangible results. Many AI projects fail in various stages of experimentation for many reasons. A recent survey of 1,600 executives by Accenture brings out the mediocre results seen from AI so far, with only 12% of firms having advanced their AI maturity enough to achieve superior growth and business transformation. At the same time, the percentage of AI “underachievers” rose to 22% this year from 17% last year. Earlier estimates from Gartner show only 53% of AI projects make it from prototypes to production.
All too often, executives and managers on both the business and technology sides of the house get enamored with the shiny objects – the vast promises made by vendors, the hyperventilation of analysts, and the trumpeting of the trade press on the wonders of AI technology and act on fear of missing out.
What’s behind the lackluster successful results seen so far from AI? What’s the secret sauce? The search for AI’s secret sauce often comes up in conversations with executives who are starting out with their initial AI use cases. They are always curious and ask about the secret sauce of the winning AI organizations, and consider these options:
- Is it data?
- Is it algorithmic choice?
- Is it skillsets and knowledge?
- Is it infrastructure?
- Is it tools?
- Is it executive support?
- Is it collaboration?
As shocking as it is, the secret sauce is something else. That secret sauce to AI success is selecting the right business use case – a robust and expansive business use case. It’s what moves AI initiatives from disjointed sets of projects to masterful performances. It’s the very reason why the business should deploy AI in the first place, the reason for upending its processes, investing in software and services, and investing in skilled developers and data scientists. If not, even if all of the other components are successful, the AI project might fail miserably.
An example of a robust and expansive business use case, for example, is instilling intelligence and predictive capabilities into digital twins — 3D and virtual replications of systems, supply chains, facilities, or entire organizations. With AI, decision-makers can run simulations and model scenarios to understand the longer-term impact of their decisions. For example, airports — notorious for miserable customer experiences — are employing digital twin technology to improve these experiences, by monitoring climate control and predicting passenger traffic flow. At the higher levels of organizations, AI-powered digital twins can assess and predict the impact of decisions on growth, resources, and revenues.
With AI, there are unlimited opportunities for robust and expansive business use cases, only limited by our imaginations. From predicting demand and planning shipping routes for supply chains to working side by side with executives and employees as intelligent assistants.
The key is to visualize, get business buy-in, and prioritize robust and expansive use cases for these possibilities first, then assemble the supporting technology to make it a reality.
Certainly, AI has many moving parts that need to be synchronized, with each part critical to its success. Data needs to be properly collected, vetted, cleansed, and be as accurate as possible. The right infrastructure needs to be in place to provide compute power to process data, create and test models, and to run algorithms efficiently. Models need to be tweaked, or reconstructed altogether, People need to be trained to understand and run the system. Funding needs to flow. A data-driven culture is mandated. Executive level support and the right funding is a must. But a robust and expansive business case is needed to deliver a winning AI effort.
Successful implementation of AI requires a deep understanding and ability to act on a business problem or opportunity. There is a famous saying, ‘“the operation was successful, but the patient died.” The same way, if you just concentrate too much on making the specific project successful and not pay attention to the business value, ROI, and fail to explain properly to the business executives it will result in disaster. In fact, this is the most common case for failure of AI projects.
In our next article, we will discuss how to capitalize on AI’s secret sauce, and develop a robust and expansive AI business case.
Read the full article here
