Simplain is very excited to introduce some exciting AI related features into our flagship product, Simplain Vendor Portal. We believe these features are grounded in the product and provide transformational benefits in terms of user experience and productivity.
Simplain Vendor Portal is a collaboration platform for retailers to streamline their supplier collaboration activities. It is a business to business application that facilitates productivity, transparency, collaboration so that retailers and their supplier partners can both be more agile in their response to supply chain disruptions and realize significant cost savings. We have added the following Generative AI features into our solution.
Product Search:
By fine tuning LLMs on retailer's item catalogue, we have enhanced the search algorithms to be more user friendly. As a simple example, searching for "DAIRY products" will show milk and other dairy items even if their product description and literature do not have the word "DAIRY". As simple as it sounds, we believe this will boost the collaboration between the retailers and suppliers as they be referring to the product using their own nomenclature. Historically, we have identified product identification as one of the biggest barriers for streamlining collaboration.
Similar Items:
When a category manager reviews a new item proposal from their vendor, a natural question is "Do we carry this item or a similar item already from another vendor?". In Simplain Vendor Portal, we leverage Large Language Models to create embedding that allow us to identify semantically similar items based on various attributes/dimensions. With this, Category Managers can quickly compare the new product to other products with similar attributes and make instant decisions.
Other AI Features in pipeline:
Simplain Vendor Portal is now enriched with OCR/image reading functionalities, powered by AI. With this, the end user can simply upload images of the product into the system rather than keying in. We are working on enhancing this to also include the scanning of descriptive product attributes such descriptive text in labels and nutritional tables.
Invoice management:
This is another module that leverages AI powered OCR features. As of now we are able to read expense invoices into the system using AI powered OCR libraries. We are currently working, reading and matching merchandising invoices into the system. Merchandise invoice matching can be little more involved, as the description and codes of the item in purchase order and invoice can be different. Moreover, different vendors can use different formats of invoices. However, we believe we can solve most of these issues by using a combination of AI powered OCR, product embedding (using their similarity in vector space) and supervised learning.
To conclude, the features mentioned above are just the tip of the iceberg when it comes to leveraging AI features into a data management application such as Vendor Portal. We can think of other techniques such as anomaly detection to identify data quality issues that are not identifiable using traditional algorithms (rule/logic based validations). Hence, we are excited thinking of the prospects of how this evolves Simplain Vendor Portal and how it helps our clients.