In order to support your employees, Didask AI begins by assessing user requests. Sometimes these are very easy to understand (“I need to practice answering customer questions about data hosting”), while other times they require more context (“customer with DX problem 402: what to do?”). As an administrator, you have several tools at your disposal to provide context to Didask AI so that it can more accurately assess your employees’ needs.

📚 How does Didask AI search for relevant information in your knowledge base?

When the AI receives a message from the learner, it will make an hypothesis about the learner's needs, which will enable it to search your company's knowledge base for the information it needs to respond to those needs.

To make this assessment, the AI will rely on:

Your business context: What is the name and mission of your company? What are your challenges and the environments in which you operate?

Your company's lexicon: if you use words that have a different meaning from their common usage (e.g. a “lead” in marketing).

The learner's context (if they have provided one while interacting with Didia): what is their role? Their responsibilities?

And based on this information, it will make an educated guess about the learner's likely needs.

Let's consider the following example: an employee on a call with one of your company's customers writes the message “how do I activate cf?”. This message alone will be very puzzling for the AI, as it does not know what 'cf' is or where the employee wants to activate it... and so the AI is going to have a hard time finding relevant sources in your database.

Now let's imagine that Didia, the Didask AI, has been given the following information:

🏢 Company context: “LogInnov is a company specializing in business software for logistics. Its main product is Flowtrack, a web app designed to track locations, stages, and delivery dates in real time. We have two types of customers: international delivery providers and large retailers. Our goal now is to expand into the passenger transport sector.”

💬 Lexicon: “Customized transport flow that allows the company to add its own delivery provider, available to customers with a silver or higher license / Custom flow / CF.”

👷 👩‍💻 Learner context: “Customer success manager, supports key accounts during regular meetings. They guide their strategy for using Flowtrack, answer both strategic and operational questions, act as advisors, and prepare contract renewals.”

Didia will then be able to make the following request: “The user is looking for the conditions for activating Custom Flow (CF), probably to set it up for their customer. Probable resource titles: ”Custom Flow: Trigger Conditions,“ ”Custom Flow Tutorial,“ etc.” and find articles in the company's help center related to CF.

📝 How to write a good business context

The company background describes the environment and challenges faced by all your employees. The information contained in the company background must be valid for all employees, regardless of their team or role.

✅ To add:

the name of your company,
the services it provides,
the type of products it sells

❌ Do not add:

information that is only relevant to a single group of employees
information that changes regularly (and will therefore need to be updated): number of customers, number of employees, etc.
generic figures that do not provide information about the real challenges facing your company (“we operate in 7 countries”)

✏️ How to write a good lexicon

The lexicon describes the specific terminology used in your company. For example, a “lead” does not have the same meaning when used in marketing as it does in general conversation. Without a lexicon, artificial intelligence will use the “classic” meaning of words or make assumptions about their probable meaning, at the risk of making mistakes. By creating a lexicon, Didask AI will better understand your users' specific requests and be able to find the right information on the subject.

✅ To add:

The terms specific to your company, typically those that you cannot use with someone who works in a different sector without having to explain them. Pay particular attention to words that have a different meaning in everyday language (for example, if you use “offline” to mean something other than “not connected to the network”). Also include acronyms and their translations if you use them internally.

Some examples:

An acronym invented by your company or industry to refer to a process, user group, etc.
A feature of your business software
A specific status

❌ Not necessary to add:

Acronyms used by all companies, or specific terminologies widely documented on the internet.

Some examples:

The European General Data Protection Regulation (GDPR).
SEO: search engine optimization.

🕵️ How to inform your learners of the benefits of adding learning context?

Each learner is free to add information about their job and the tasks they perform. They can edit this information at any time if they wish.

However, each learner has a specific context. Which means that the same request from different people will be interpreted in different ways, according to their context.

For example, a developer and a customer success manager do not need the same information when they ask a question about one of your company's products.

When deploying Didask AI, we recommend that you inform your learners of the importance of taking the time to specify their context.