AI Implementation Risks Accelerating Faulty Business Processes, Warns Expert

Noam Cohen, founder of CritiqeIL, a consultancy specializing in AI integration for Israeli organizations, warns that implementing artificial intelligence on flawed business processes can amplify errors and threaten organizational stability. He compares using AI on broken workflows to accelerating a powerful race car on a pothole-filled road, which only leads to faster crashes rather than progress.

Cohen recounts an example where a company deployed AI for customer service responses, resulting in quick but incorrect answers. The AI did not create the problem but accelerated an already broken process, producing a misleading appearance of professionalism to clients. He emphasizes that AI acts as an accelerator, speeding up existing processes, so if the underlying process is flawed, the errors will spread more rapidly and widely.

Before integrating AI, Cohen advises managers to ask five critical questions: whether the process works manually, if it can be clearly explained to new employees, whether they are accelerating a good process or masking a broken one, how errors are handled mid-process, and if the process has been measured beforehand to assess improvement. He stresses the importance of mapping and fixing processes before applying AI technology.

Cohen concludes that organizations that repair their workflows before AI adoption gain a powerful accelerator to advance their operations. Those that skip this step risk creating accidents that are difficult to recover from. He highlights that while AI tools are impressive, they should only be used to speed up well-functioning processes, not to cover up existing faults.