- Supported languages
Automatic knowledge discovery:
- Questions and answers in your dialogues
- Useful facts in your messages
- Expertise of your teammates
- Automatic search for answers in Slack channels
- Asking Memonia to search
Create knowledge manually
- A Q&A pair
- Memorising facts
- Knowledge collections
- Memonia Slack commands
- Conversational Memonia bot
Memonia supports only English at the moment.
We're working on support other languages but it takes time to implement new language models.
Automatic knowledge discovery
Questions and answers in your dialogues
Let us show you how it works, providing you some example:
Choose a public channel.
Automatic knowledge extraction works only in public channels in DM. Invite the Memonia bot to any channel with:
/invite @Memonia #your-channel
Ask a question in this channel.
Memonia tries to ignore contextual, personal or small talk questions like this:
Try to use instead more generalised/factual forms of questions:
Hey, where to find an instruction manual for our office printer?
Kindly ask any of your teammates to answer your question.
Memonia detects the answers replied in the Slack threads, using mentions or textually related in some time frame and not too distant from each other.
As well as for questions, Memonia ignores contextual, personal or small talk answers. So answers like
might not work. Try to use instead more generalised/factual forms of answers like:
We have a corporate Wiki. There are a lot of manuals for many devices. I'm sure you'll find the printer manual there.
Wait for a few minutes.
Memonia intentionally waits for any possible outcome of your conversation (other replies may come, etc). It usually takes 2-3 minutes.
You will receive a notification that the Q&A pair has been indexed.
If the previous steps were done as expected, you should receive a notification with the message from Memonia.
In case you want to edit this new Q&A pair, you will easily find the 'Edit' button in the message from Memonia.
Ask a similar question in the channel.
A new question must be similar or close to the question from the original Q&A pair:
Where to find an instruction manual for our office printer?
Wait for a few minutes (again).
Additionally to doing some voodoo to find answers, Memonia intentionally waits again for any possible outcome of this conversation as well.
It usually takes 2-3 minutes.
You will receive an answer from Memonia.
You should receive a notification with the original answer from Memonia.
Useful facts in your messages
Using built-in reading comprehension engine Memonia memorises some of useful messages describing some facts.
Those factual messages are identified not just by their sentences structure, but also it depends on channel,
terms frequency and other statistical models. The same message text might be memorised in one channel as factual,
but filtered out in another. The more messages coming in your channel, the more Memonia understands what is really important.
This function is work completely automatic, and, occasionally, you just might receive
useful answers for your questions in your channels.
Expertise of your teammates
Analysing messages from your channels, Memonia identifies some of specialisations related to
your teammates in your particular workspace, which we call user expertise.
The discovered user expertise could be used to find your teammates in some areas related to the questions.
This discovered user expertise might be requested with the @memonia bot or the search command.
This kind of knowledge is extracting entirely automatically, and you don't need to do anything to enable it.
However, it only works if you have thousands of messages in your public channels.
Memonia might also automatically show a private hint with a short list of people who might help for some of your questions.
With Memonia, your team may easily access to the shared knowledge, and your teammates get answers fast and work more productively.
Memonia is able to find you answers using various approaches:
Automatic search for answers in Slack channels
Memonia automatically recognises common forms of questions in Slack channels and looks up for if there is anything related to those questions.
You should receive a message from Memonia if there is anything useful:
Asking Memonia to search
There are several ways to search answers to your questions:
Ask Memonia bot privately using direct messaging, or mention it with @memonia in public channels:
Where is our office printer manual?
Use /memonia search command:
The manual search finds semantically relevant knowledge entities as well as supports the unified keyword-based search in all of your integrations.
Creating knowledge manually
Sometimes you want to save useful message or you'd like to make series of Q&A pairs (e.g, FAQ).
Use "Create knowledge" in your message context menu.
Knowledge is linked to a Slack message, so you shouldn't create multiple Q&A pairs or knowledge items based on the same message.
A Q&A pair
Provide any question and modify the answer for this question.
The answer will be already filled in from a message, but the question has to be provided to proceed.
Make up the most popular and relevant question to the answer.
When you finish, click on the 'Create' button. There's some validation process to determine if the question is suitable for Memonia automated search.
For instance, avoid using contextual/time-constraints (like 'it', 'this', 'that', 'at the moment'), people nicknames in questions.
You would see the updated status message if anything went wrong.
In case you want to create multiple Q&A pairs (FAQ, for example) you should notice that you have to create series of separate messages and each of these messages should contain some answer.
Then you could create a Q&A pair from each of those messages individually. You cannot create multiple Q&A pairs from single Slack message.
In case you just want to save useful information from some message, but don't want to make up a question,
use "Factual text" in a "Create knowledge" dialogue and Memonia would gather all facts from it and try to semantically answer any questions related to the text.
For example, for text like:
Programming in Scala is the definitive book on Scala, written by Martin Odersky.
This is a comprehensive tutorial and reference for Scala, this book covers the entire language and important libraries.
Memonia could answer for some questions about it like:
Do you know any good books related to Scala?
You also might want to consider to use knowledge collections to share knowledge with other channels.
By default, automatically discovered knowledge from one channel isn't available in other channels.
With the knowledge collections, you could group your channels knowledge into some named areas/topics.
The knowledge in one collection is available to automatic and manual search in all of the channels from the same collection.
Use the bot or the collections command to manage your collections and channels.
Memonia can show you the insights and statistics related to your knowledge base extracted in channels and workspace.
Just ask a bot (e.g. using DM):
For a whole workspace:
show my digest
For a particular channel:
show my digest for #some-channel
Conversational Memonia bot
Memonia provides a conversational chat bot for Slack which is available for direct messages or mentions (to @memonia).
It supports free-form dialogues, so instead of Memonia commands you can just ask or request what you'd like to do.
Inviting Memonia to public channels:
/invite @memonia #my-channel-1
Searching in knowledge base:
Where is our office printer manual?
Searching in the knowledge base related to the channel:
search in #my-channel-1 for "Where is our office printer manual?"
Creating knowledge manually:
create a Q&A pair
Knowledge collections management:
show my collections
show my digest
show my digest for #my-channel-1
If you have any questions, suggestions or need help, don’t hesitate to get in touch!