Retool — a company that provides a platform designed to streamline the process of building internal business applications — has published a comprehensive State of AI report that sheds light on the changing dynamics of artificial intelligence in the business world.
For this research, Retool surveyed over 1,500 technical professionals, including software engineers, business and engineering leaders, executives and designers to find out how they’re using and building with AI.
Let’s take a closer look at some of the findings from the report, to see what we can learn about how companies and individuals are integrating AI into everyday business practices.
Perceptions vs. reality in AI utilization
There is a popular belief that AI’s immediate impact is overstated, but Retool’s research indicates the opposite. The report highlights a significant shift in workplace practices due to AI.
A striking 68% of respondents reported a decreased reliance on Stack Overflow, with 94% attributing this change to the adoption of advanced tools like GitHub Copilot and ChatGPT.
Even among the most critical voices who rate AI's current hype as 10/10 on the “overrated” scale, only 14% believe their companies are overinvesting in AI.
Driving forces behind AI adoption
Businesses are primarily motivated to invest in AI for three key reasons: cost savings, staying abreast of industry trends and addressing customer needs.
This means AI isn’t just a technological fad — it has the potential to offer tangible benefits to companies.
AI in action: Internal vs. external applications
The landscape of AI applications within companies is rapidly evolving. Two-thirds (66%) of businesses have implemented at least one internal AI use case, and 43% have created external apps.
Prominent internal uses include programming, handling knowledge base queries and copywriting. Externally, AI has found a place in product features, customer service chatbots and external knowledge bases.
Challenges in AI development
The primary concern in AI application development is the quality and accuracy of model outputs — 39% of respondents noted this as a potential challenge.
Data security is a close second at 33%, with concerns scaling up with company size — peaking at over 40% in larger organizations.
The dominance of OpenAI's models
Not surprisingly, most of the survey respondents are using the industry leader in artificial intelligence — OpenAI.
In fact, OpenAI is the overwhelming leader in the popularity of AI models, representing 80% of all use cases. Among businesses using a single model, 89% opt for OpenAI.
Trends in model hosting and customization
There’s currently a strong trend toward hosted AI models, with nearly 70% of companies following this route and an additional 20% planning to implement open-source models soon.
Larger companies show a greater inclination towards self-hosting. Surprisingly, half of the companies Retool surveyed do not engage in any form of model customization.
AI development tools and databases
Retool also asked respondents questions about AI dev tools. Over a quarter of respondents are using Hugging Face, followed by Langchain and AWS Bedrock. Pinecone takes the top spot in vector databases, trailed by MongoDB and PGVector.
AI is becoming essential for operational excellence and innovation in business
Retool’s report indicates that companies aren’t just implementing AI because it’s trendy — they’re increasingly adopting it to address concrete operational and customer needs.
As AI continues to reshape the business landscape, companies must navigate challenges like model quality and data security, while strategically leveraging AI tools to stay competitive and innovative.
It is clear we are in the early innings of the AI investment supercycle. AI is a core investment area at Wing — find out more about how we are seeking investment opportunities with companies that are driving the forward motion of data and AI.
To stay on top of the latest news for founders and practitioners navigating the data and infrastructure software landscape, subscribe to my Substack.