Feature Request

While working with the Eddy AI Analytics page in the Knowledge base portal, I noticed that the citation details for responses are not currently visible. At present, Eddy AI displays references like [citation 1], [citation 2], but it doesn’t show which specific articles those citations are pulled from. Why does it matter? We regularly review the queries users submit and how Eddy AI responds to them. If the information provided is accurate, there’s no concern. However, when the response is incorrect or out of context, we need to identify the source of that information to make necessary corrections or improvements. Currently, to find out where Eddy AI fetched the content from, we often need to manually re-enter the same query in the Knowledge base site search. This process: Consumes Eddy AI query credits May return different results due to frequent content updates or backend changes in Eddy AI Sometimes fails to reproduce the original citation, making it harder to trace and verify the response Suggested improvement: It would be very helpful if the Analytics section could include the exact article titles or links from which the citations were generated. Providing this visibility within the portal itself will: Help us validate responses more efficiently Enable quicker identification and correction of outdated or misleading content Enhance trust in Eddy AI outputs for both the team and our end-users Thank you.

Eddy still looks at strikethrough text as if it is standard text. If text has a strike through it, it should be treated differently by the search...

We've been testing Ask Eddy using our content to see if it's something we'd want to add to our site. However, there are a few key issues we identified, As with other chatbots, Eddy has a major problem with hallucination. It often makes up answers, especially when leading questions are used. In one of our test questions (regarding deleting a category), Eddy makes up a response that could potentially cause huge problems for users with significant data loss. I checked the reference pages cited to make sure we didn’t include or allude to the incorrect information. Additionally, the reference articles are often unrelated to the answer provided (sometimes Eddy is able to provide the correct answer even though it is not included in the reference articles cited). In other instances, Eddy will pull an answer from one page, but miss that the correct answer was provided in another article, and will not include that page in the reference articles. Eddy also seems to have some difficulty pulling answers from HTML tables, which I also found when testing ChatGPT separately. For example, when I asked which versions of PHP are supported, the response ignored the version numbers provided in the table on the system requirements page, and instead generated a response based on the body text on this page. Finally, the responses are typically very short and give less detail than a standard GPT would - I assume this is done intentionally to reduce the risk of hallucination. This wouldn’t be much of an issue if the reference articles were more accurate. I don't know what the solutions are to these issues, but hopefully this feedback is useful. As others have mentioned in their feedback, having some control over the instructions provided to the GPT would help to alleviate some of these issues.

Contributors often create articles on similar topics across departments, which can lead to duplication or conflicting information. Using AI to analyze these interactions would be helpful. Suggestion: Add a tool to analyze content across articles to detect data duplication, conflicts, and potential interactions. This would help streamline content and improve consistency.

Enable AI assistive search for Brazillian portugese

It would be great if Eddy could detect the right English variant in the query language and the reply should be given in the same variant.

It would be great if the answers provided by Eddy could return links (links which are added to the text) in its response.

We have received a request from the customer to include a unique request/response ID in the exported CSV from Eddy Analytics. This would provide a unique reference ID for each user request, helping to avoid duplication and enabling better tracking of valid requests with ID.

Creating this on behalf of one of our customers: Having more than one workspace and when ask eddy is enabled, the customer should get an option to set generate answers from specific workspace(s) or generate answers from all workspaces.

Right now in the Analytics for Ask Eddy, we only see what the end user is asking and the metrics of their activity. It would be great if we can have a provision to see what Ask Eddy gives as a response. This will really help the tech writers understand where the gaps are in their content and what Ask Eddy is processing behind the scenes.