Hoping someone can give us some advice/guidance on why we're getting a difference in the knowledge base articles surfaced in web messaging with a Digital Bot Flow, compared to the articles surfaced by the same search in the Knowledge App.

E.g.  The Knowledge App "Search for a topic" is returning 3 relevant article and it looks like these articles are based on the phrases for the article.  This is what we would expect.

For the same search, our Digital Bot Flow "Send a message" is returning 5 articles and it looks like these articles are based on key words in the heading of the article and not based on phrases at all.  Some of these articles are not at all relevant to the search.

We have the same variations turned on for both touchpoints.

We would like the Digital Bot Flow to recognise and respond to the phrases within the articles - any ideas on how we can do this?

Hi Lee,

I am not sure why its different, can you confirm what you mean by Knowledge App?

In the meantime, hopefully someone in the community might be able to answer you better, but I will look into it.

Hoping someone can give us some advice/guidance on why we're getting a difference in the knowledge base articles surfaced in web messaging with a Digital Bot Flow, compared to the articles surfaced by the same search in the Knowledge App.

E.g.  The Knowledge App "Search for a topic" is returning 3 relevant article and it looks like these articles are based on the phrases for the article.  This is what we would expect.

For the same search, our Digital Bot Flow "Send a message" is returning 5 articles and it looks like these articles are based on key words in the heading of the article and not based on phrases at all.  Some of these articles are not at all relevant to the search.

We have the same variations turned on for both touchpoints.

We would like the Digital Bot Flow to recognise and respond to the phrases within the articles - any ideas on how we can do this?

Hi Sam,

What I mean by "Knowledge app" is the Knowledge app for Messenger: A self-service outlet that, paired with the Messenger Homescreen, enables customers to navigate between a bot or agent conversation and your public knowledge base (as per the Resource Center).
My understanding is this is the part of Web Messenger that has the functionality behind the "Search for a topic" part in Web Messenger.

When I click on "Search for a topic" and put the search in that I used as an example above, I get back 3 relevant articles.
When I put the same search text into "Send a message" (i.e. the digital bot flow), I'm getting back 5 articles of which 2 are not relevant.

Any help will be much appreciated :-)

Hi Lee,

I am not sure why its different, can you confirm what you mean by Knowledge App?

In the meantime, hopefully someone in the community might be able to answer you better, but I will look into it.

Hoping someone can give us some advice/guidance on why we're getting a difference in the knowledge base articles surfaced in web messaging with a Digital Bot Flow, compared to the articles surfaced by the same search in the Knowledge App.

E.g.  The Knowledge App "Search for a topic" is returning 3 relevant article and it looks like these articles are based on the phrases for the article.  This is what we would expect.

For the same search, our Digital Bot Flow "Send a message" is returning 5 articles and it looks like these articles are based on key words in the heading of the article and not based on phrases at all.  Some of these articles are not at all relevant to the search.

We have the same variations turned on for both touchpoints.

We would like the Digital Bot Flow to recognise and respond to the phrases within the articles - any ideas on how we can do this?

Hi Lee,

I see, that's clarified it for me.  Let me have a look into this and see what I can find.

Hi Sam,

What I mean by "Knowledge app" is the Knowledge app for Messenger: A self-service outlet that, paired with the Messenger Homescreen, enables customers to navigate between a bot or agent conversation and your public knowledge base (as per the Resource Center).
My understanding is this is the part of Web Messenger that has the functionality behind the "Search for a topic" part in Web Messenger.

When I click on "Search for a topic" and put the search in that I used as an example above, I get back 3 relevant articles.
When I put the same search text into "Send a message" (i.e. the digital bot flow), I'm getting back 5 articles of which 2 are not relevant.

Any help will be much appreciated :-)

Hi Lee,

I am not sure why its different, can you confirm what you mean by Knowledge App?

In the meantime, hopefully someone in the community might be able to answer you better, but I will look into it.

Hoping someone can give us some advice/guidance on why we're getting a difference in the knowledge base articles surfaced in web messaging with a Digital Bot Flow, compared to the articles surfaced by the same search in the Knowledge App.

E.g.  The Knowledge App "Search for a topic" is returning 3 relevant article and it looks like these articles are based on the phrases for the article.  This is what we would expect.

For the same search, our Digital Bot Flow "Send a message" is returning 5 articles and it looks like these articles are based on key words in the heading of the article and not based on phrases at all.  Some of these articles are not at all relevant to the search.

We have the same variations turned on for both touchpoints.

We would like the Digital Bot Flow to recognise and respond to the phrases within the articles - any ideas on how we can do this?