AI Can't Fix a Business System with Broken Processes
October 2, 2026 - 10:15 pm
AI (explore this topic) cannot fix a business system with broken processes — this is a misconception that can be costly. While 88% of organizations use AI, only 7% have successfully scaled it (McKinsey). Dessy Pavlova argues the gap arises when companies apply AI to fragmented operations instead of redesigning the process.
She creates systems where:
- A single human decision triggers automatic updates across websites, scheduling, and finance.
- Stopgates allow for human oversight.
NBER research shows that AI boosted support-agent productivity by 14% on average, but it assists people in performing work, not eliminating the need to understand it.
The issue lies in many businesses' operational fragmentation. A customer interacts with a website, data is entered into spreadsheets, moved to platforms, and updated back onto the website—creating numerous opportunities for delay, duplication, or error.
Pavlova suggests building systems around the business itself, integrating tools seamlessly while allowing humans to remain the architects. AI becomes the operational engine. Stopgates ensure human oversight despite automation.
Research supports this approach. A study of 5,179 customer-support agents found that access to generative AI increased productivity by 14% on average, benefiting particularly less experienced workers.