New AI workflows and GDPR

New AI workflows

Agilebase is now able to call on AI to both create and update data in your system. You can use existing data to feed it prompts, and have those act, automatically, on data in bulk.

If that seems a bit abstract, an example should help bring it to light.

You can think of the generic question as ‘given we have data X, how do we generate data Y’?

So thinking about some typical uses from the sales world:

In many sales systems (including our own), we have a complete conversation history for every enquiry, made up of emails to and from key contacts, our own notes and potentially call transcriptions.

We could do many things with this data – provide a quick summary for reviews, generate a task list or work out a likelihood of conversion for example.

Let’s take the task list idea. The way we’d accomplish that in Agilebase is to create an AI workflow.

Start by creating a standard ‘records creation’ workflow, then add an ‘AI prompt’ calculation, something like

“We have a conversation history with {company name} as follows: {comment feed [auto]}. What would be the most immediate tasks we’ll need to do to progress the sales opportunity?”

We then choose a table in which to store the generated tasks and fields in which to save the data, for example “task name” and “assigned to role”.

The exact process to set this up is described in our documentation.

We could then let that workflow run manually, on a button press in the enquiry record, or automatically e.g. every week.

Unlimited uses

The advantage of a platform like Agilebase of course is that this isn’t limited to a few specific use cases designed by an application developer. You can use these AI workflows for your own data and cases, whatever field you’re in. A food manufacturer can help create recipe variations, based on the ingredients they already purchase. A membership organisation can suggest upcoming events to members based on past attendance etc.

Data Privacy

Once you start feeding your own data into the AI, whether that’s commercially sensitive or private conversations with individuals, you’ll want to know that it’s not going to leak out and be used for the AI provider’s gains, not yours.

To that end, we can announce that by default, Agilebase will use the Mistral AI. Mistral is a French company bound by strict GDPR regulations. You can review their terms of service and privacy policy here: https://mistral.ai/terms#data-processing-agreement

Mistral is also the developer of some of the leading open-weight large language models, an alternative to proprietary models like those powering ChatGPT.

As previously announced, if you have your own AI model which you wish to use, you can easily override the default, supplying details to let Agilebase use a privately hosted AI solely for your organisation, or just another alternative.

Pricing

Pricing starts from just £20/month for AI workflow use. During the early adoption phase, we’ll be reviewing the volume of usage and related costs to tweak the pricing structure.

There are still a limited number of spots to be part of our beta testing programme, for which the cost is zero.

Please contact us for more information.