New AI Data Platform Focuses on Accuracy

A new platform developed by UK firm Askalytics allows users to ask questions about their business data in plain English, and makes multiple checks on its own answers before showing them.

The firm says it is 'entering a market where conversational analytics has become common and accuracy has not.' AI's answers are fluent but not necessarily accurate. 'A wrong number and a right one look exactly the same on screen,' says founder Colin Fagras. 'Same confident sentence, same tidy chart. If you already knew the answer you would not have asked, so you are the last person equipped to catch it. That felt like the actual problem worth solving.'

In addition to tracing figures back to a source, Askalytics' platform verifies calculations independently rather than trusting them implicitly. Written insights are read a second time by a different AI model, from a different company, and where the two disagree the platform shows the disagreement instead of a clean result. These checks are performed by default with every query.

Askalytics connects to a business's Google Analytics 4, Search Console and Google Ads accounts, alongside HubSpot, Mailchimp and Microsoft Clarity, with more integrations being added. Questions can be asked from the platform, from a Slack channel, or from a Microsoft Teams conversation, where an answer comes back in seconds. The platform also delivers a weekly email containing insights drawn from the customer's own data; monitors for unusual changes and flags them; and provides dashboards covering audience, content, conversions, ecommerce, organic search and paid performance.

The firm is on the web at www.askalytics.io .

All articles 2006-23 written and edited by Mel Crowther and/or Nick Thomas, 2024- by Nick Thomas, unless otherwise stated.