Matan Kleyman

Analytics for AI products ?

16
•

Hi everyone,

What are you using to track user behavior in your AI product?

If I have a conversational AI assistant, what tools can I use to monitor user satisfaction, intentions, or frustrations?

Add a comment

Replies
Best
Hassan Sajjad

An intent heatmap helps me spot gaps in responses.

Matan Kleyman
@m_hassan_sajjad Interesting . How do you generate this heatmap?
Tania Bell / Product Delights

oh that's clever, @matan_kleyman . ask questions that your product helps solve. gg, man

Matan Kleyman

@taniabell I'm building a product because I honestly care about ppl dealing with these questions. Don't you think that's a worthy cause?

Tania Bell / Product Delights

@matan_kleyman i don't have a view on the cause.


you could have mentioned that you've got a product that solves these problems. that's all

jonny vince

AI analytics can be great for refining my product

Matan Kleyman
@jonny_vince what is the kpi you are trying to optimize ?
John Davis

For AI products, tracking metrics like model performance, error rates, and usage patterns is key

Matan Kleyman
@john_davis29 model performance and error rates are easier to track than usage patterns. How do you measure usage patterns automatically ?
Muhammad Saleem

The beauty of AI analytics lies in its ability to provide insights

Matan Kleyman
@muhammad_saleem27 what kind of insights are you looking for ?
Hazel Mathew

Session recordings reveal pain points that raw data can't.

Matan Kleyman
@hazel_mathew what if you could convert session recording into structured thing you can query for insights ?
Gilad Barnea

At my last workplace we'd given in-house eval tests a try a few times

Matan Kleyman

@giladbarnea What evaluations did you run ? Your custom business logic evaluations?