Redgate Software is previewing an artificial assistant (AI) tool that makes use of generative artificial intelligence (AI) to create, analyze and modify SQL using a natural language interface.
IT teams via that interface will be able to craft a query in natural language that Prompt+ will then convert into suggested SQL code optimized for the existing database schema that can be used to launch a query. That code is generated by the large language models (LLMs) developed by OpenAI without sharing data with those LLMs.
Tom Hodgson, innovation tech lead at Redgate Software, said Prompt+ will provide IT teams with an AI assistant that can be applied to multiple databases versus requiring IT teams to one day deploy and manage different AI assistants for each database they deploy.
A wave of AI assistants are on their way to the enterprise. The issue is IT teams might soon find themselves inundated by AI assistants they will need to orchestrate. Each IT team will need to decide for itself how many AI assistants it wants to employ versus relying on a smaller number capable of working across multiple platforms.
Of course, most IT teams should test the code generated by any LLM, especially a general-purpose platform that was trained using examples of code aggregated from multiple sources across the Web. Not each piece of code used to train an LLM model is likely to have been fully vetted for quality, so the output generated might not work at all or be especially efficient in terms of the how much infrastructure resources might be consumed to run it. In addition, IT teams need to also make sure there are no vulnerabilities being inadvertently introduced.
Overall, however, the potential productivity gains in terms of how much code can be written will largely outweigh the potential risks. In fact, itโs worth remembering that not every developer writes highly optimized code, so in many cases the output from a generative AI platform might be better.
The one thing that is certain is writing, for example, SQL queries, will become a lot more accessible to IT teams and end users alike because the need to understand how to construct them will be much less than it is today. In fact, a survey conducted by Redgate Software finds nearly two-thirds of respondents (64%) are already using AI for query and code optimization. Just under two-thirds (65%) are also using AI tools to help with tasks involving database schema, while more than half (55%) are generating sample data or code snippets.
In the meantime, IT teams should start considering how AI will change their roles as more processes become automated. Itโs not likely AI will replace the need for IT professionals any time soon, but the way IT teams are structured is going to evolve as it becomes easier to manage more applications at higher levels of scale. In fact, in many cases databases are already being managed by DevOps teams rather than traditional databases administrators. Advances in AI will only make it easier to consolidate the management of IT platforms that previously required dedicated administrators.
In effect, IT teams, rather than manually managing all those platforms, will soon find themselves supervising a series of highly automated processes to make sure whatโs expected actually happens as they planned with the help of an AI agent.