#LaunchWithAI – Benefits of being cloud-native on Azure

| Nandita Jaya

Zammo header

In the initial stages of launching Zammo.ai, CEO and founder Alex Farr and chief engineer Guy Tonye faced a familiar problem. The team had exclusive access to an impressive propriety technology to power a revolutionary conversational AI engine, but existing cloud platforms like Google, AWS, and Microsoft were also rapidly advancing their AI. A classic tech conundrum.

In assessing the opportunities and challenges, Farr questioned the long-term defensibility of the advantage if they built the platform in-house. “Even with solid differentiators like voice-first architecture that competitors would not be able to match,” Farr explained, “I decided to move forward with fully cloud native,” as in building Zammo’s conversational AI solution on an existing cloud platform.

Having decided to build on the cloud, the next key decision was selecting the cloud. As chief engineer, Tonye had initially planned to build in the Google Cloud Platform, given his prior knowledge and experience with those services. However, after comparing it with Azure, he found that Azure performed better, across a range of evaluation criteria, for their ambitious SaaS product vision.

Born in Cameroon, educated in France and living in Toronto, Tonye was already building one-off voice-first conversational experiences for top brands. “My passion for voice technology started with being born into a tribe with very low literacy,” Tonye shared. His personal vision is to “globally democratize voice AI to expand everyone’s access to information.” Since making the decision to build on Azure in 2019, Guy has been able to leverage Azure’s AI, from conversational to language, to help scale toward that vision faster than he’d imagined.

The full picture

Farr and Tonye knew that to succeed they had to deliver a superior solution in this crowded market, which required being better at both speed and user experience:

  • Speed. Their solution needed to perform with minimal latency and generate production-ready proofs-of-concept (POCs) to new customers within 24 hours for fast go-to market.
  • User experience. The voice-first, omni-channel experience needed to provide the end user a consistent experience delivering the very best natural language processing across many languages, with unified analytics, and across many channels: through text, voice, or digital-voice assistants like Google Assistant, and Amazon Alexa.

Zammo’s platform needed:

  • Broad API integration with SaaS platforms like ServiceNow, Salesforce, Cisco, Genesys, Avaya, Zendesk and others
  • Ease of use without requiring specialist developer skills to create, deploy and update conversational experiences
  • Scalability for composite AI use cases, which combine conversational AI with other AI services, including Azure Cognitive Search, Intelligent Recommendations, Personalizer, Document Translation and more

Building an in-house solution to achieve all of this would have taken several years. Zammo discovered that Microsoft’s applied AI services would help build the product while prioritizing these needs.

Tonye said, “Many enterprise SaaS startups are simply unaware that starting cloud native can address their unique and specific needs. Building in-house may seem to offer more control, but it can also add a great deal of time and expense. There are also hidden costs to in-house development. For example, if startups can’t respond nimbly to market demands, such as the need for omnichannel availability in conversational AI, they may find that the market takes off without them.”

How I built this?

Connecting Azure Language Services, Speech Services, Text Analytics, QnA Maker and Azure Bot Framework Composer, Zammo created their architecture off ready-to-deploy and customizable AI models. Reference architecture below:

Zammo.ai architecture
Fig. 1: Zammo.ai architecture, built from off-the-shelf Azure Cognitive Services components to get up to speed quickly

The final tally?

Zammo’s investors are taking note of the capital efficiency resulting from leveraging Azure AI as well as the extensive support from Microsoft engineering, sales and marketing. Zammo investor and advisor Kevin McQuillan is a venture capitalist who has been selected five times as a Forbes Midas List VC and has been involved with 26 IPOs. “Across all the companies I have invested in and advised over 30 years as a VC, Zammo.ai has shown one of the most-efficient capital burns that I have ever seen,” McQuillan said.

Getting to market fast is better for business… and sometimes it’s better for the world, too. If you’re already part of Microsoft for Startups Founders Hub, click here to get started with Conversational AI services in minutes.

Tags: ,