Case study

Rapidly integrated recommendations.

AI for product recommendations with seamless integration.
Share:

Challenge

Creating customised, AI based recommendations for products. Dynamic recommendations typically require using protected and/or confidential data. We needed to do without. The team needed to create a solution that knows the popularity of specific products, their connection to other products and based on that provide most relevant purchase advice; whilst sticking to the data limitations.

Solutions

A recommendations dynamic was created in a way that no confidential or protected data were required. A team deployed a solution that was easy to integrate and improved the search results in a snapshot of time. Additionally, a way to monitor profit improvement from these was provided.

Technology & Tools

Python
ScikitLearn
Pandas
GCP
Keras
TensorFlow
PyTorch
Numpy
AWS
S3

Client

Depict.ai

Opportunity

Integrating product recommendation solutions is usually a time consuming and problematic process. Normally the best customisation requires access to sensible data. Additionally, the created solution is usually tightly interwoven with the product.

Delivery

Requirements and a business case were obtained from the clients. The team enquired about the product portfolio, type of recommendations preferred and any particular needs. Based on that a recommendations’ dynamic was created. One that was created to not require confidential or protected data and can do with the information on clients that was permitted (which varied vastly from client to client). To substitute that, the team deployed a solution with the use of Google Analytics. Simultaneously data was scraped from websites and the front-end solutions were prepared. Based on data, recommendations were prepared and a demo was created. Improvements were made based on feedback received from the trial and a final solution was prepared. After integration, a portal access was provided, so that the client could easily see the profit from the recommendations.

For this case our developer worked with the client’s in-house SCRUM  team as a team extension. They delivered this to the client of the company that hired us. 

Effect

The solutions were implemented seamlessly and customers were satisfied with the results. The key advantages were the ease and speed of integration that took a few days instead of months.

 

More case studies

Case study

Urgent project rescue.

Re-inventing the faulty solution for managing machine learning models.

learn more.
Case study

Fresh approach to an old problem

Monitoring the EU roaming – the application.

learn more.

Get in touch to see
how we can generate
value together

let's talk.
Join our clients