Release notes

What changed in Inquio, newest first. The version your account is running is shown at the bottom right of every screen.

5 October 2026

Compact filtering on Conversations and User utterances. Both screens have a new, smaller filter bar, so the filters take up less room and leave more of the screen for what you came to read.

Share exactly what you are looking at. Copying the URL of a filtered view now carries the time window as well as the filters. Send the link to a colleague and they open the same conversations over the same period, instead of the current period with your filters applied to it.

On-premise deployments extension for Enterprises. Inquio can now run against Dell EMC ECS S3 storage and Azure OpenAI deployments, with internal back-end calls that need no token. Contact sales (sales@inquio.ai) to discuss it.

Unsuccessful attacks no longer count against Safety & Compliance. An attack your assistant repelled is evidence that it behaved correctly, so it no longer reduces the score. This changes existing scores: if your assistant has faced attempted attacks and handled them, its Safety & Compliance figure will be higher than before, without anything on your side having changed. Successful attacks and incidents still count. See Safety & Compliance.

Charts rebuilt around Issues and Incidents. Conversations over time and Frequent topics now follow the current Issue and Incident categories, so the colors in those charts match the flags you see on individual conversations.

Conversation metrics now use Inquio's own measures. Containment Quality and Abandoned conversations on that screen are now calculated the same way as everywhere else in Inquio, so the same conversation is counted the same way whichever screen you read it on. See Conversation metrics.

Counts are clearer on report cards and insight details. Both the number of issue types found and the number of conversations affected are now shown. A single conversation can carry several findings, so these two numbers are different and both are useful.

Signing in from a shared link. The login flow, session checks and the redirect that follows sign-in have all been corrected. Opening a shared conversation link while signed out now takes you to that conversation after you sign in, rather than dropping you on the home screen.

Full-text search leaked into other screens. A search typed on Conversations also filtered Issues, Sales Radar and Defender, so those screens showed less than they should have and gave no sign why. Search now applies only to Conversations.

Defender incidents and the Safety & Compliance calculation. Recently added incident types were not always recorded or scored correctly.

Links from a report card now set the right period. Following a finding from a report card through to the conversations behind it landed on the wrong time window.

Let an AI reformat your data for you. If your export does not match the format Inquio expects, you no longer have to study the specification and rebuild it by hand. When an upload is rejected, the error window now includes a ready-made prompt: copy it into ChatGPT, Claude, Gemini or whichever assistant you use, hand over your file, and get back a file in the format Inquio accepts. This is aimed at the most common reason a first upload fails — data that is perfectly good but shaped differently. See Prepare your data.

Upload errors now explain themselves. When a file is rejected, the screen lists what was wrong with it instead of simply refusing the upload.

Help opens the documentation. The Help link in the left panel now leads to this site.

Report card overview rebuilt. The list of report cards now shows a clear state for each one: live data, processing, validation failed, or processing failed. Previously an unfinished card was hard to tell apart from a finished one.

Bots can be opened during an upload. You no longer have to wait for processing to finish, or for a failed upload to be cleared, before opening the bot. The bots list also works before a first report card has been calculated, so a new bot is usable straight away.

Industry is shown when editing a bot. The Edit dialog now displays the industry you chose at setup, as read-only. It still cannot be changed. See Manage bots and report cards.

Plan descriptions corrected. The subscription screen now describes each tier accurately, and the Starter conversation limit is shown as 1,000, which is the figure that has always been enforced. See Plans and limits.

Assistants overview simplified. The +n badge next to a bot's score has been removed. It counted the bot's other report cards, but sat directly above the score and read as a change in score. The number of cards is still shown inside the bot, as a total.

Opportunities are now called Leads. The Sales Radar box on a report card used the word Opportunities while the Sales Radar screen itself said Leads. Both now say Leads. See Sales Radar.

Growth accounts could not reach Live data. Access was incorrectly blocked. Live data works on Growth as intended. See Live data and report cards.

Issues and recommendations: wrong counts and charts. The severity totals and the trend chart did not always match the issues listed beneath them.

Report card showed wrong Sales Radar figures. The Leads and Churn counts in the Sales Radar box on a report card did not match the Sales Radar screen. They now come from the same source.

Conversation display. Overlapping headers when scrolling, and long incident summaries that overflowed their panel.