How kea, an AI startup, uses DigitalOcean to pivot and scale

kea AI is a voice-based artificial intelligence platform that enables restaurants to take orders over the phone through kea’s virtual cashiers, freeing up employee time and creating upsell opportunities that result in larger orders. After launching their first product in 2018, kea has been improving their platform and gaining new customers, which resulted in a $10M Series A funding round in 2020. Along the way, kea relied on DigitalOcean to assist in their scaling and help them pivot as they worked to continuously improve their technology solutions. Learn more about their startup story below:

Background

kea was born in 2017 out of a realization that the manual phone ordering system for restaurants was inefficient and unsustainable. Restaurants struggled with hiring and training employees to take phone orders, which often caused them to lose out on money by missing potential orders. kea sought to solve this by automating the phone ordering system and partnering with restaurants so they could take orders over the phone without hiring more staff. This saved restaurants time and money, and kea’s artificial intelligence system had the additional benefit of resulting in higher ticket prices by intelligently suggesting additional items to customers. The value of kea’s solution became even more clear in 2020, when the COVID-19 pandemic forced many restaurants to turn to take-out orders; The use of phone-in orders quickly grew, and by using kea, restaurants were able to handle the increase in capacity during a time when hiring additional staff was challenging.

Challenge

When kea was first built they used a large cloud provider for their cloud services. As they grew and started to improve their platform, they were having challenges around the lack of transparency their cloud provided around their AWS Lambda setup; turning off one function could make others stop working, which was challenging as a lean organization trying to maintain their stack. They also were looking for a platform that would support their rapid growth and make it easy to scale their system.

Around this time, kea discovered DigitalOcean through our startup program, Hatch, which would give them access to infrastructure credits and dedicated support. For a fast-growing startup, moving to DigitalOcean as part of the Hatch program was a logical choice, as it provided kea with a long runway for their cloud services and would reduce the complexity of their setup, enabling them to spend more time improving their platform. kea made the switch to DigitalOcean, using Droplets and our Managed PostgreSQL services.

“DigitalOcean was super easy to understand and use, and instead of having to focus on correctly configuring everything, we were able to spend our time on writing code.” -Patrick Wingo, Head of Product, Kea

Once switching to DigitalOcean, kea found the simplicity of DigitalOcean’s solutions, the large library of documentation and tutorials, and the support provided to them enabled them to focus more on their business. They found it was easy to set up Droplets with just a few clicks in DigitalOcean’s user-friendly interface, and quickly got support whenever it was needed.

“The support team is incredible - every time we would have an issue we’d send out a request and somebody really pleasant and nice would get back to you quickly and give you exactly what you need. I’m very pleased with the level and quality of service we get.” Patrick Wingo, Head of Product, Kea

Before raising their Series A round, kea was scaling quickly and running into challenges related to their growth. They also were working on rebuilding parts of their platform in order to further expand their capabilities, and found that DigitalOcean was the ideal partner for a startup like theirs which was iterating and pivoting quickly. kea scaled up with Droplets, first adding more resources to each Droplet, and then moving to a horizontal scaling model where they added more droplets and containers, and then used a load balancer to move to a larger set of Droplets as they built their new platform. They also experienced first-class support for Kubernetes.

“A lot of times with a startup you’re experimenting, and what is great about DigitalOcean was that it pivoted with us. There were enough technology choices that it was easy for us to transition from one way of operating to another while feeling seamless for our customers” Patrick Wingo, Head of Product, Kea

kea currently leverages DigitalOcean Managed Kubernetes, Managed PostgreSQL, Managed Redis and Spaces as core part of their technical infrastructure:

Architecture

kea uses CloudFlare to manage their inbound traffic, which lands on a worker pod in kea’s DOKS cluster. There are various services such as recording and transcription running on the DOKS cluster. These services leverage other DigitalOcean products, such as DigitalOcean Managed Postgres for managing state, and Spaces to store transcriptions long term. In addition to the processing work occurring locally to the cluster, the cluster also works with external services, such as an external logging facility.

Impact

After raising their Series A in 2020, kea has been focused on growing their team and continually improving their product. Their current team is about 40 people, many of which are in the engineering and product organizations, and in the past year kea has brought on new customers including large fast-food chains while looking to the future of their artificial intelligence platform. As an innovative and growing company, kea is focused on making incremental improvements to their products which demonstrate their progress to investors, and DigitalOcean has been an integral part in their success.