AI usage at Cuvva

Note: This article was published in Cuvva’s Linkedin page.

Cuvva has released many AI-powered features in the past. From customer support chatbots to fraud detection analysis, many of our features have leveraged AI throughout the years. Although these features were usually developed by a very technical group of engineers, the emergence of LLM tools in recent years has attracted a more diverse audience to what AI actually is and how we can all benefit from it.

With the public release of models like DALL-E and GPT-4, an ever-increasing number of Cuvva staff began to experiment with these AI products and followed closely how the industry was moving. Around mid-2023, when questioned about the application of these new AI models for Cuvva, Freddy, our own CEO and founder said:

‘…I’m extremely excited about giving you (access to) these tools, with the goal of making a big impact and seeing what happens.’ Shortly after, employees got permission to expense any AI tools that would support their workflow.

Very soon, the early adopters in the company started sharing new tools, use cases, and techniques to leverage LLMs on a variety of Slack channels. In engineering, for example, when a new model or coding tool was released and someone wanted to use it, the legal team would review it to ensure it complied with our data security and privacy concerns, and generally approved it shortly after.

This exposure slowly led to a widespread adoption that we had suspected was happening but couldn’t quite quantify. So, earlier this month, we gathered the opinions of over half the company, evenly distributed across departments, and our findings were even more surprising than what we expected.

All Cuvvarati were using AI tools in their day-to-day work, with 93% feeling positive about AI’s impact on their work, and over 60% using it daily across all sorts of activities. Research, brainstorming, and writing/editing were mentioned by most respondents, but a surprising 62% use it to learn a new topic! This aligns with what one Cuvvarati said, ‘ChatGPT helps me do tasks and projects that I have absolutely zero skills or knowledge in, so I can take on anything.’

Beyond the obvious personal development and growth, the impact on collaboration between departments is what is most impressive. Just recently, we had a member from the legal team reaching out to engineering with a new process, based on LLM prompts, to answer and self-serve trivial (and non-Cuvva related) questions. A few months ago, we also saw our fraud specialist demo a working prototype built with AI, hosted in his GitHub account, to illustrate what functionality we should integrate next.

These are all anecdotal examples, but we certainly didn’t need a Chief AI Officer or organisational redesigns to adopt AI, as mentioned by Riviera’s latest ‘Future of the tech leadership in 2025’ report. It is true that having a relatively small and agile company helps, but companies are made of people. When you have bold leadership with an optimistic view of technology and very little red tape, enthusiastic and curious employees will help drive any change.

To be honest, after all these years it still surprises me that a company in an apparently ‘boring’ industry can still be such a lively workplace full of ideas, opportunities, and modern technological challenges. Our main limitation has been time, with the quality we aim for.

But fortunately, AI is starting to help us with that.