Sales Day: AI Visibility and Display Scoring

15/10/2026

The Misleading Metric of Productivity

Nearly half of a modern sales team's time is consumed by checking data from retail outlets. While this may seem reasonable, the figure often creates a significant misunderstanding about actual work efficiency. Many assume that reducing inspection time will directly increase sales. However, in practice, reducing inspection time does not equate to increasing selling time. It simply changes how data is processed. This is a critical point that many managers overlook when deploying new technology solutions.

The core issue is not a lack of information, but rather that the information is cumbersome and not properly standardized. A sales representative might visit 20 outlets in a day, but if they spend 15 minutes recording and cross-referencing each product on the shelf, they will have little energy left to persuade customers to buy more. This is a trade-off few want to admit: time spent on detailed observation is time taken away from human interaction.

The Reality of Retail Outlets

The retail market in Vietnam, and extending to neighboring markets like Thailand and the Philippines, shares a common characteristic: extreme fragmentation. Small outlets of varying scales have very low compliance with display regulations if relying solely on manual checks. When discussing AI Visibility in this context, we are not talking about passive remote monitoring. We are talking about transforming quick on-site photos into structured data.

What is often said is that an app is needed to take photos and send them to the office. However, reality shows that this process often gets stuck at the processing stage. Photos are often blurry, taken from incorrect angles, or subject to constantly changing lighting conditions. If the AI system is not sensitive enough to filter out junk images, the back-office team will have to sit down and classify them manually. In that case, the technology solution is merely shifting the work from the sales team to office staff, rather than solving it.

In a recent deployment for a fast-moving consumer goods brand, we found that the biggest barrier was not recognition technology, but the habits of the photographer. Sales representatives often take photos hastily, not adhering to standard angles. The AI system must be intelligent enough to automatically adjust or politely reject non-compliant images, accompanied by specific instructions, rather than simply reporting an error. This interaction creates the difference between a monitoring tool and a true assistant.

Deployment Flow and the Role of AI Vision

For AI Visibility to truly deliver results, the deployment process must be divided into clear steps, starting with redefining what constitutes a standard photo. We work with clients to define specific criteria: shooting distance, viewing angle, and elements that must appear in the frame. Subsequently, the AI Vision system is trained to recognize specific SKUs on shelves, cross-reference them with the approved planogram, and automatically score compliance levels.

The key here is that the display scoring algorithm must reflect the reality at the point of sale. Not all deviations are equally serious. A product partially obscured by a competitor's box has a different level of severity than a product that is completely missing. The system needs to prioritize errors so that the sales team can focus on issues that directly impact revenue. This is where the industry expertise of a consulting firm like AIVISION plays a vital role, ensuring that technical metrics are translated into specific business actions.

We also integrate the ability to detect fraudulent behaviors such as screenshots or duplicate images. In a highly competitive environment, sales representatives reusing old photos to submit reports is a potential risk. The AI system needs a mechanism to compare image matrices to automatically eliminate these cases, allowing management to trust the data without doubting every report submitted.

Measuring Effectiveness and Achievable Outcomes

After deployment, we do not measure effectiveness by the number of images processed, but by the time saved by the sales team and the quality of interactions at the point of sale. In an illustrative example, time spent on recording and cross-referencing data decreased significantly, allowing sales representatives to spend more time consulting and persuading store owners. Planogram compliance at outlets also improved markedly, not due to top-down pressure, but because data is updated in real-time, creating a closed feedback loop.

However, it must be acknowledged that no solution is perfect from day one. Initially, the rejection rate for low-quality images was quite high. This forced the sales team to adjust their photography habits. This adaptation process took about a month, but afterward, the overall team's productivity increased sustainably. This change did not come solely from technology, but from clarity in the workflow. When everyone knows exactly what is expected and how to achieve it, resistance to change decreases significantly.

True effectiveness lies in transparency. Management can see the overall market picture from standardized data, rather than having to compile it from hundreds of disparate reports. This leads to faster and more accurate decision-making. For example, discovering that a specific geographic region has an unusually low compliance rate can lead to targeted marketing campaigns or appropriate sales policy adjustments.

What Would Be Different in the Future

If we were to deploy from scratch, we would focus more on user training from the design phase. Instead of just handing them a tool, we would organize workshops to co-create photography rules and evaluation criteria. The involvement of the sales team in shaping the system creates a sense of ownership, helping technology adoption go more smoothly. Additionally, we would integrate predictive features to suggest next actions for sales at the point of sale, based on historical data. This transforms AI Visibility from a control tool into a strategic ally.

Returning to the initial figure regarding inspection time, we see that the ultimate goal is not to eliminate this process entirely, but to optimize it until it becomes invisible. When display scoring happens automatically and instantly, the sales team only needs to focus on what machines cannot do: building relationships and persuading people. That is the outcome we want to see in any AI Visibility project: technology serving people, not the other way around.

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.