Debunking Display Scoring Myths for ASEAN Expansion

10/11/2026

In my experience deploying AI for retail chains, I have observed that nearly half of international expansion projects face delays due to errors in the data preparation phase. While this figure may seem modest, the consequences are severe. Many businesses believe that simply installing software and uploading photos is sufficient. In reality, the barriers lie not in the technology, but in the differences in supply chain operations and market standards. This is why preparing for AI Visibility and display scoring requires a practical perspective rather than just presentation slides.

"As long as photos are taken the same way, the system will recognize them accurately"

Many regional managers believe that if employees in Bangkok take photos using the same template as those in Hanoi, the display scoring results will be equivalent. This is a common misconception because people tend to trust the consistency of processes. However, real-world deployment shows that store lighting in Indonesia is often more complex, or shooting angles are limited by narrow aisles. AI systems can handle these variations well, but if you expect "identical" results without reconsidering the local context, you will receive misleading reports. For example, a shelf in the Philippines might have products stacked differently than global standards, requiring the algorithm to be recalibrated to avoid misjudging compliance scores.

"The system automatically removes all incorrect photos"

This is a blind faith in absolute automation. While VTraks can detect duplicate or fake images, it cannot "know" that an employee is intentionally obscuring a competitor's product with an object not in the existing database. When expanding into new markets, you will encounter competitor display tactics that the system has never seen. Believing that AI will automatically filter out all noise will cause you to overlook critical anomalies. Instead of waiting for the system to handle everything, you need a cross-checking mechanism by humans or periodic updates to the image library so that the AI Vision system can recognize items more accurately.

"Expansion is just replicating the Vietnamese process to Thailand"

This approach sounds logical but is actually a trap. The Thai and Indonesian markets have their own cultural specifics and local regulations. Directly applying the Vietnamese scoring process can cause resistance from the sales team or skew data due to different food safety standards or labeling regulations. For instance, some products in Indonesia have mandatory local language notes; if these are not accounted for in the scoring process, the system will flag them as missing, even though they are compliant with local laws. There must be flexibility in software configuration to adapt to the new context.

"Data from other countries can be mixed into a single dashboard"

Many CIOs believe that aggregating all data in one place will improve management. However, when combining data from multiple countries with different time zones, currencies, and measurement standards, a common dashboard can become cluttered and difficult to read. If data is not clearly separated by geographic region before aggregation, performance reports will become meaningless. For example, planogram compliance rates in Jakarta cannot be fairly compared with those in Bangkok without adjusting for store complexity. A data middleware layer is needed to standardize information before it is presented to leadership.

"Technology is the deciding factor; people are not important"

The most dangerous mistake is underestimating the human element. No matter how powerful the AI is, if the sales team does not understand why their scores change, they will stop cooperating. When expanding into new markets, retraining employees on how to use display scoring tools is mandatory. They need to understand that AI is not a punitive supervisor, but an assistant that helps them optimize their time. If you focus only on the software and ignore human behavioral changes, the project will fail no matter how good the technology is. AIVISION always emphasizes that technology is merely a means; people are the ones who operate the system.

Returning to the initial figure of nearly half of projects facing delays. If you overcome these misconceptions, that number will no longer be a concern. Expanding into Thailand and Indonesia is not just a technical problem, but a problem of adaptation. Start by re-evaluating your assumptions about data and processes. Thorough preparation in the early stages will determine the success of your future AI Visibility campaign.

Every company hits this differently, and the hard part is usually the data rather than the model. To pressure-test your case quickly, talk to AIVISION - or first see how we deploy and what we have written before.