iGenius used a combination of Google Cloud AI machine learning APIs on a Google App Engine infrastructure to create the world’s first AI marketing advisor
Built a working prototype in just 30 days
- Delivers a virtual advisor experience with seamless integration of Google Cloud AI machine learning APIs
- Saves 50% on server costs with flexible pricing
- Reduces the need for DevOps, empowering developers to work on building the best possible product
With custom-built artificial intelligence and a tightly focused customer base, Italian company iGenius has built crystal, the world’s first virtual marketing advisor. Ask crystal a question in natural language, such as “How is my website doing?” and it provides immediate answers and clear insight within a couple of seconds. When iGenius founder and CEO Uljan Sharka first conceived of crystal his aim was to provide insight as quickly as possible, allowing customers to stay ahead of the game and make more effective decisions. To do that, iGenius turned to Google Cloud Platform (GCP) and its Google Cloud AI machine learning APIs.
“Our advisor, crystal, creates context from lots of different data streams,” says Uljan. “To make it work we had to integrate lots of different components and technologies, and Google offered us the best way of doing that.”
“Our advisor, crystal, creates context from lots of different data streams. To make it work we had to integrate lots of different components and technologies, and Google offered us the best way of doing that.”—Uljan Sharka, CEO and Founder, iGenius
Seamless AI integration
Digital marketing provides a large amount of data, such as website analytics, social media tracking, and conversion rates, so finding valuable information often takes significant time and effort. When iGenius began building crystal, it knew that it would have to create a system with powerful AI capabilities, combining several machine learning processes with a solid, high-performance infrastructure. In addition, as a startup, iGenius had limited IT resources and wanted to maximize the amount of time its developers could work on crystal, rather than maintenance.
After trialing other machine learning products on different platforms, iGenius adopted Google because of the quality of its products and the ease of integration. Voice recognition in particular represented a huge improvement over the competition. “We tried one of the leading voice recognition products and kept having to repeat what we were saying,” says Uljan. “Google Cloud Speech API was much better for us. It’s like you can speak while you’re eating or listening to a song.”
“Google App Engine was very easy to get going. We set it up with all the deploy scripts and it was just like writing a terminal line. We deployed our code, and Google App Engine handles all the traffic. It was very smooth because the load balancing is built in.”—Uljan Sharka, CEO and Founder, iGenius
For the AI processes, iGenius used a combination of Google Cloud Natural Language API, Google Cloud Speech API, and Google Cloud Translation API to help make spoken interaction with crystal as easy as possible in a number of languages. After experimenting with another cloud provider for its infrastructure, iGenius switched to Google App Engine for its ease of use, automation, and flexible pricing, while Google Cloud Pub/Sub helped coordinate and integrate the processes and components that make crystal work.
“Google App Engine was very easy to get going. We set it up with all the deploy scripts and it was just like writing a terminal line,” says Uljan. “We deployed our code, and Google App Engine handles all the traffic. It was very smooth because the load balancing is built in.”
High performance, low maintenance
Thanks to the ease integrating with Google technology, iGenius was able to create a fully functional prototype of crystal within 30 days, just in time for a high-profile demonstration at Cannes Lions Innovation festival in June, 2016. Since then, iGenius has iteratively built new versions of crystal and taken on more customers. With Google App Engine scaling capabilities, crystal now handles 5,000 visitors per day and has the capacity to handle several times more while maintaining high standards of performance.
“We can take our engineers out of the DevOps environment and work on the product, as well as other areas. For instance, we can send engineers to the marketing team to help boost engagement or mine public data, and generate thousands of leads for our sales team. The results have been great!”—Uljan Sharka, CEO and Founder, iGenius
iGenius has also saved over 50% of its infrastructure costs compared to its previous cloud solution thanks to GCP flexible pricing that helps ensure the company only pays for what it uses. Preparing for crystal’s next iteration, iGenius is currently evaluating Google Cloud Dataproc and Google Cloud Dataflow to enhance its data processing and improve its infrastructure even further. Meanwhile, GCP automation and its ease of use means that the iGenius team of developers can fully concentrate on building the best product possible, without spending valuable time and effort on infrastructure maintenance.
“We can take our engineers out of the DevOps environment and work on the product, as well as other areas.” says Uljan. “For instance, we can send engineers to the marketing team to help boost engagement or mine public data, and generate thousands of leads for our sales team. The results have been great!”
With offices in Milan, London, and Silicon Valley, iGenius offers digital solutions to make the complex simple.