How do I know what help articles to write?

Look at what customers ask that your current articles don't answer. AskBase does this automatically: every unanswered question is grouped with similar ones and ranked by how often it comes up.

Where do the questions come from?

From the AI assistant on your website and help centre. Customers ask in their own words, which is far more telling than what they type into a search box.

What counts as a gap?

A question the assistant couldn't answer from your articles — either because nothing relevant exists, or because what exists wasn't clear enough to answer from.

Do I need the chatbot to use this?

Yes — the questions are what the assistant was asked. It works on your help centre too, so you get gaps even before it's on your website.

Knowledge base analytics

See what customers ask that your help centre can’t answer

Every question your AI assistant couldn't answer from your articles is collected, grouped with others that mean the same thing, and ranked. The top of the list is your next article.

Get startedSet up in an afternoon. No credit card.

A to-do list, not a search log

Search logs show you thousands of typos and one-word queries. This shows you “14 people asked how to set up a custom domain”, with a button to write that article.

Grouped by meaning

“How do I cancel?”, “close my account” and “stop my subscription” are one gap, not three.

Ranked by demand

The questions asked most often come first, so your writing time goes where it helps the most people.

Real questions, in customers’ words

See exactly how people phrase things, which is often the best title for the article you write.

Why answers fall short

Content health separates questions where nothing relevant was found from ones where something was found but wasn't clear enough — each needs a different fix.

Articles that aren’t pulling their weight

Content health flags articles the assistant keeps reading but never uses, usually because they're long or cover too much.

Feedback on every answer

Customers can rate each answer, so you see which topics get thumbs down even when an answer was given.

From question to article

1
Customers ask
On your website or your help centre, in their own words. The assistant answers what it can.
2
Gaps collect
What it couldn't answer is grouped and ranked over the last 30 days, updated as questions come in.
3
You write the article
Start a new article from the gap in one click. As soon as it's published, the assistant starts using it.

Also on the dashboard

How many questions were asked, and how many were answered
The questions answered most often, and which articles answered them
Every conversation, with the articles each answer used
Where customers asked for a person instead

Questions people ask

How do I know what help articles to write?: Look at what customers ask that your current articles don't answer. AskBase does this automatically: every unanswered question is grouped with similar ones and ranked by how often it comes up.
Where do the questions come from?: From the AI assistant on your website and help centre. Customers ask in their own words, which is far more telling than what they type into a search box.
What counts as a gap?: A question the assistant couldn't answer from your articles — either because nothing relevant exists, or because what exists wasn't clear enough to answer from.
Do I need the chatbot to use this?: Yes — the questions are what the assistant was asked. It works on your help centre too, so you get gaps even before it's on your website.

Works together with

Find out what your help centre is missing

Put the assistant in front of customers and the list starts filling itself.

Get started