Knowledge

AI agent for frequently asked questions

An AI agent answers questions about opening hours, services and policies from your knowledge and leaves unanswered questions for your team.

Website

The guest gets precise arrival instructions

The AI agent uses information prepared by the property: check-in time, entrance and parking. The guest gets a specific answer for their arrival without waiting for reception.

Park Apartments Website
  1. Customer

    We arrive around 3 pm. When is check-in and where can we park?

  2. Telp

    Check-in starts at 2 pm, so you can go straight to the apartment. Parking is in the courtyard, entered from 2 Bocna Street. Are you arriving by car?

  3. Customer

    Yes, by car. Is reception at the same entrance?

  4. Telp

    Yes. After entering the courtyard, reception is through the first door on the right. Parking is included for apartment guests.

Telegram

The student knows what to prepare for the first class

The AI agent shares current school instructions through the channel the student already uses. Answers stay tied to the programme and rules in your knowledge base.

Lingua school Telegram
  1. Customer

    What should I bring to the first German A1 class?

  2. Telp

    Bring a notebook and pen. First-class materials are provided, so you do not need to buy a textbook in advance. Will you attend at school or online?

  3. Customer

    Online, from home.

  4. Telp

    For the online class, prepare a computer with a camera and headphones. The link is in your welcome message. Join five minutes early to check your audio.

The customer context

An AI agent answers questions about opening hours, services and policies from your knowledge and leaves unanswered questions for your team.

Uses approved business knowledge to answer recurring questions consistently and makes gaps visible to the team.

What this workflow can cover

  • Use only knowledge attached to the selected agent
  • Answer in the language and tone defined for the experience
  • Avoid inventing an answer when the approved source is insufficient
  • Escalate or collect a follow-up request when certainty is not enough

Knowledge

How it works

  1. 1

    The team prepares a focused set of approved documents or answers.

  2. 2

    The agent retrieves relevant content when a customer asks a question.

  3. 3

    Unanswered or escalated conversations show where knowledge or process rules need improvement.

What your team gets

More consistent customer answers

Less repetitive explanation for the team

A clearer list of missing or outdated knowledge

Telp

How your team keeps control

Business rules stay explicit

Instructions, required details and allowed actions are configured for the selected workflow.

Sensitive steps can wait for confirmation

The agent can prepare a structured request while a person remains responsible for the final action.

Handoff keeps the conversation

Operators receive the customer context, collected data and the point where help is needed.

Turn business information into a clear answer

Customers do not need to know which document contains the answer. The AI agent uses your approved service descriptions, opening hours and policies to answer a specific question. On a website, this is available through AI website chat.

Preparing knowledge means checking which information still applies and to which location, product or service. When an answer needs a decision or unavailable data, it belongs in the broader customer support workflow.

Two locations have different policies

The agent first clarifies which location the question concerns. One set of opening hours or conditions should not be presented as the rule for every branch.

Information is no longer current

Your team needs to update approved sources and check the resulting answers. An old document does not guarantee that a listed price or service still applies.

FAQ

Frequently asked questions

Does the AI agent only answer prewritten questions?

Customers do not need to use an identical phrase. They can ask in their own words, while the answer remains grounded in approved information.

What content should we prepare?

Clear service descriptions, opening hours, terms and common questions. Separate locations and policy versions so different sets of information are not confused.

Can it check today's stock levels?

A product document is not a stock feed. Current data requires a connected source that supports the particular check.

What if the knowledge does not contain the answer?

The question needs clarification or human follow-up. The agent should not present an assumption as an official company policy.

Can we change information after launch?

Yes, business knowledge is maintained through the platform. After an important policy change, check several realistic questions before relying on the updated answer.

How do we assess answer quality?

Prepare typical questions and a few exceptions, then check both answers and human handoff. Clear, current sources matter more than the volume of content.

Want to validate this workflow with your own examples?

Tell us the channel, current process and first result you want to improve. We will suggest a focused rollout path and the cases your team should test.

View pricing