Over the past year, our customers have expressed a growing interest in building custom AI-powered applications. As more and more developers and businesses look for ways to use AI, it’s a great opportunity to come together to discuss strategies and lessons learned from actually building and implementing AI apps. Last week, we had the pleasure of hosting an event in London with OpenAI and 50 members of the European tech community to collectively discuss our experiences building custom AI applications.
We heard from several speakers about how they’re leveraging AI within their organizations, and saw some demos of AI applications currently used by customers and internal business units. Here are some highlights.
Dom Grillo, Head of Technical Success at OpenAI, shared that while most companies recognize the potential of AI to change how they do business, few have a concrete strategy in place. Many companies start by automating their operations using a combination of chatbots, customer support agents, and AI-powered internal tooling. A common theme is that automating operations often requires custom development on top of the OpenAI API.
Dobo Radichov, CDO at Holland and Barrett, shared his vision and strategy for using AI to make his team more efficient at finding answers to data-driven questions. The user enters a question into the application, and the LLM routes the user to the appropriate dashboard or document which contains the answer. Dobo’s solution stands out because it combines LLM capabilities to both summarize documents and perform search across a tabular dataset.
The Retool team presented two AI applications currently being used to improve the efficiency of customer-facing teams:
- : Retool uses AI to deflect low-risk customer support questions and allow our support engineering team to focus on challenging questions. The support team also uses a RAG chatbot to inform their responses to customers.
- Natural language query engine: A customer’s infrastructure team uses an AI application to allocate resources to customers. The app translates a natural language question into a SQL query, allowing operators to query a large data warehouse which can otherwise be challenging to navigate.
The overall consensus was that custom-built AI applications are often more useful than out-of-the-box solutions because they can be tailored to your company’s data model and business requirements. Speed of development and flexibility are also key to developing useful AI applications, since both the technology and business requirements are constantly evolving.
We’re excited to continue working with our customers on creative AI applications. We hope you’ll join us at future events as we build together.
Cheers!
Keep your toolkit up to date with the latest insights, how-tos, and news from Retool.
Reader
