Over the past year we’ve watched as artificial intelligence systems have wowed us with amazing capabilities. Here’s just one example:

AI obviously has the potential to be revolutionary in many fields. For Agilebase, we’ve gathered lots of great ideas, not least from speaking to customers, and tried out some fun tests.

We’ve also heard about some of the potential pitfalls to watch out for when relying on AI.

That means we’ve been able to be thoughtful about how to take advantage of AI capabilities and consider what the purpose would be.

It has to help us in our aim of changing who gets to build software, making it easy and accessible for all.

To that end, we’re releasing the first beta test of our AI Assistant feature of Agilebase, one of many that will help people get to grips with building systems more quickly.

Prompt-based system building

Given the name or description of the system you’d like to create, e.g. ‘CRM’ or ‘stock control’, the AI Assistant will create a set of tables which will comprise that functionality. It will then connect them all together and flesh out each with relevant fields. In short, you get a complete system built in a couple of key-strokes, which can then be customised to your heart’s content.

Here’s a quick demonstration of the process:

The idea came to us after talking to someone using Agilebase for the first time, who said they’d like to see templates for building common tables, e.g. contacts or sales enquiries, for getting started faster. Now, you in effect have an infinite amount of templates – you can create anything the AI knows about!

What’s to come

We’ve many more ideas, gathered from customers and partners, which we look forward to introducing soon.

We will be mindful of some of the things to watch out for when working with AIs. For example, there’s their propensity to make up ‘facts’ such as Australia doesn’t exist. More insidiously, when large language models are trained on vast quantities of data, they can reinforce and amplify biases, for example race and gender stereotypes. For an AI-powered counter to that, check out Bristol company https://www.inclued.ai/

Also, there’s the inconvenient point that training AI systems takes a massive amount of energy, which in our global goal to work towards net-zero carbon emissions, could be severely problematic.

Keeping these in mind will be important as we progress beyond beta testing.

If you’ve any AI ideas you’d like to contribute, please let us know – leave a comment on our LinkedIn page.