Excellent take here! Love it. Thanks for sharing. 15:40 Regarding the sale period that it did not fetch, I suspect it has something to do with cache and not the LLM's ability to interpret the data. You requested for the data immediately after updating it, so I guess the cache was not updated (if you have some server cache thing setup). Thanks again for sharing! I appreciate the insights.
Very nice, we were playing also with it with real live data the last weeks, but experiencing issues with real live data e.g. namings with brands, categories, or simple when a product doesnt have any necessary key word in the description...we are thinking to use maybe tags or something to get better results...maybe also a follow up video idea?
Intersting. It seemed for me the ai agent was also reading the image (depending on the model used) - which means if the image is sufficient, and has a category, maybe no descriptive text is necessary? i'd recommend adding something though, like tags as you mentioned.
for the WooCommerce part? check the n8n docs, they have a check list of exactly what you need. If its a new site, you need to change permalinks to not plain, otherwise I think everything else was from scratch in this video. heres docs docs.n8n.io/integrations/builtin/credentials/woocommerce/#using-api-key
Excellent take here! Love it. Thanks for sharing.
15:40 Regarding the sale period that it did not fetch, I suspect it has something to do with cache and not the LLM's ability to interpret the data. You requested for the data immediately after updating it, so I guess the cache was not updated (if you have some server cache thing setup).
Thanks again for sharing! I appreciate the insights.
Hey Tobi, thanks for watching. Yes you're probably right that i was fust going too fast for the model to see the update. Cheers
Awesome! Is it possible to style the output (messagee) so that it looks better with images?
there probably is. the nodepackage page for the chat widget has some styling options, i forgot to show that.
Very nice, we were playing also with it with real live data the last weeks, but experiencing issues with real live data e.g. namings with brands, categories, or simple when a product doesnt have any necessary key word in the description...we are thinking to use maybe tags or something to get better results...maybe also a follow up video idea?
Intersting. It seemed for me the ai agent was also reading the image (depending on the model used) - which means if the image is sufficient, and has a category, maybe no descriptive text is necessary? i'd recommend adding something though, like tags as you mentioned.
@@BrendanOConnellWP yes probably image file name, it reads all fields from the woocommerce API that you get from the product endpoint....
Hey man I having trouble with the credenciales, I do everything docs says, even turn off the wordsafe plugin and N8N still tells my I’m unauthorized
for the WooCommerce part? check the n8n docs, they have a check list of exactly what you need. If its a new site, you need to change permalinks to not plain, otherwise I think everything else was from scratch in this video. heres docs docs.n8n.io/integrations/builtin/credentials/woocommerce/#using-api-key
i just ran into this on a site. identical to this video but it just don't take. no wordfence either. LMK if you figure out also
👍
niceeeee
how to we know all conversation conversation in our website?
not sure what I know what you mean?