AI Display Scoring in the Philippines: Process Breakdown

05/10/2026

Display Scoring Process: Where Does It Break Down When Expanding to the Philippines?

When rolling out in-store display scoring in a new market like the Philippines, most sales management teams assume the challenge lies in the software. However, based on experience consulting for FMCG companies in Vietnam and Southeast Asia, the reality is more complex. Nearly half of the time and budget is often consumed by manual data reconciliation and resolving edge cases that AI cannot yet handle fully automatically. This figure can be misleading, suggesting we need a more "magical" algorithm, when the core issue actually lies in the human operational processes surrounding the technology.

To bring AI Visibility into practice, we need to dissect each stage of the current workflow. Imagine a supermarket chain in Manila. Today, a sales representative (saché) goes to check the stock. They photograph the shelves. The issue arises in the next stage: the data goes to the central office, where logistics staff input or cross-check it. The bottleneck is not in taking the photo, but in turning that image into a business decision: is the stock running low, is the price wrong, or is a competitor taking up space? AI fits into the step of converting images into structured data. But deciding "what to do next" still has to be handled by humans.

Crossroads #1: Standardize Input Data or Prioritize Collection Speed?

This is the first and most painful decision. Do you choose to require staff to take photos according to an extremely strict standard (45-degree angle, sufficient lighting, no obstructions) to ensure AI accuracy, or do you choose a flexible approach so that staff can complete their route faster?

The trade-off here is clear. If you choose a strict standard, SKU recognition accuracy will be high, but you will face resistance from the sales team. They will take fewer photos, or take them carelessly to avoid the process. If you choose flexibility, you get a large volume of data but with noisy quality. With current solutions like VTraks, the system is capable of processing real-world store images with a claimed accuracy of 99.97% (according to vtraks.com), but this requires the input data to have a certain level of sharpness and viewing angle. Standardization does not mean taking photos in a studio, but building the right photo-taking process directly on the smartphone. If the Philippine market has many small, independent stores with poor lighting, you may need to consider an additional image quality check step on the mobile device before uploading. Do not let AI bear the burden of fixing human errors. Let humans do their job well: take good photos. Let the machines handle the rest.

Crossroads #2: Full Automation or Keep the Human Loop?

The display scoring process should not be a one-way street. Many companies make the mistake of thinking that AI will provide a score and that’s the end of it. In reality, AI is a decision-support tool, not a decision-maker. At this stage, we must choose: will AI automatically generate reports and send alerts, or will AI provide recommendations and wait for manager confirmation?

My recommendation is a hybrid approach. Let AI handle the high volume of tedious work: SKU recognition, comparison with planograms, and compliance score calculation. But keep a "confirmation" step for edge cases. For example, if AI detects a product not in the current catalog, or a price tag that is half-obscured, have the system flag it as "needs review." Regional managers will review these cases. This builds trust. If you let AI automatically penalize or reward based on minor recognition errors, you will lose the sales team's trust very quickly. Let humans work with cleaned and structured data, instead of having to sift through hundreds of photos. That is the essence of AI Vision in management: it does not replace the brain; it frees the brain from processing raw information.

Crossroads #3: Focus on a Few Metrics or Measure Everything?

When expanding to a new market, the appeal of AI is its ability to measure everything: stock availability, pricing, placement, shelf cleanliness, and even staff attitude (if cameras are present). But be careful. Measuring too many metrics without strategic significance creates "information noise." Staff will not know what to improve first.

You need to select 2-3 key performance indicators (KPIs) that directly reflect revenue. For example: % of SKUs in stock on the shelf (Out-of-Stock detection) and Planogram Compliance. These metrics are the core of display scoring. Do not try to cram in sentiment analysis or auxiliary metrics in the early stages. Keep the process focused. Once you have mastered these two main metrics in the Philippines, then expand to other metrics. This helps the team focus on the right improvements, and the input data will be more consistent, thereby increasing the efficiency of the AI model.

The Outcome of the AI Visibility Journey in the Philippines

Returning to the initial estimate that nearly half of resources are wasted. When the process is redesigned according to the crossroads above, that waste does not disappear completely, but it shifts from "technical failures" to "operational processes." Instead of spending time arguing about AI misidentifications, the team focuses on how to increase stock availability at key retail points. AI Visibility is not a magic spell that turns a weak sales team into a strong one. It is a magnifying glass that helps you see exactly where the problem lies, and how quickly.

Successful deployment in a new market does not depend on how many cameras you have or how complex the algorithms are, but on whether you dare to cut out unnecessary manual steps. Let the machines do the machine's work, and let people do the people's work. That is the most sustainable path to bringing display scoring into practice, whether you are in Hanoi, Manila, or Mexico City. AIVISION and partners like VTraks are supporting many major brands in the region to achieve this, but the operational challenge ultimately always belongs to your business.

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.