Electronic Products & Technology

Outdated innovation processes threaten to derail AI

Stephen Law   

Automation / Robotics Electronics AI artificial intelligence

Lux Research report proposes overhauling Stage-Gate processes to manage AI innovations

While manufacturing companies have been investing in AI heavily, the results have been mixed. Companies have been hampered because they have been using traditional processes to manage AI innovations, according to a recent report from Lux Research. The study outlines the challenges companies face when it comes to AI innovations and how they can change their innovation processes to improve their success rates with AI.

Source: Getty Images

“It’s imperative to not apply old thinking to new problems, especially when it comes to AI,” said Lux research director Dr. Shriram Ramanathan. “The fundamental challenge lies in that the underlying logic in an AI solution is intricately tied to the raw data that it provides insights on. While this allows AI solutions to adapt easily to continuously changing environments, it is also AI’s Achilles’ heel. AI solutions quickly start deviating from their original purpose the minute they are deployed in the real world. Companies need to implement a continuous and ongoing developmental effort to keep their AI models current.”

Lack of a clear product definition

While companies have been instituting Stage-Gate processes to manage innovations for decades, they are not well-suited to AI applications for several reasons. For one, it’s difficult to lock down a definition for the final product, as the adaptability of AI by its nature prevents a clear definition for a finished product. The lack of a clear product definition leads to uncertainty in the business case. This also makes it difficult to define key performance indicators that can measure the effectiveness of an AI application. Lastly, there are challenges in keeping AI products current due to its dependence on external data.

Lux has found that the Stage-Gate model can be applied effectively to AI with some minor modifications. To begin, define the product use case as narrowly as possible while planning for a broad range of scenarios and outcomes. Define suitable KPIs to measure the true ROI of an AI solution while keeping in mind that these solutions are not traditional software plays but rather bridge the physical and digital worlds. Lux Research says that companies should plan for real-world deployment early in the development phase to get a clearly defined product. They can do this by incorporating a wide range of real-world data while building the product and by building solutions that are easy to both track and update. Lastly, companies should plan to stress-test their products periodically in order to catch problems early.


According to Lux Research, to be successful in deploying AI, companies will have to challenge traditional processes. With the right changes, companies can adapt meaningful AI solutions successfully. Lux Research predicts that the digital transformation of physical industries will evolve more and more toward a framework that emphasizes the simultaneous development and deployment of software solutions.

To learn more, download the report’s executive summary here.



Stories continue below

Print this page

Related Stories