Create your AI Bot scenario
Table of Contents
Available for AlcmeonChat, WhatsApp, RCS, Facebook Messenger, Instagram Direct, Apple Messages for Business, and LINE channels
Thanks to generative AI, Alcmeon bots can generate near-human responses to customer queries, providing accurate information or even connecting to your systems to retrieve personalized data or perform actions automatically, thus reducing the need for intervention from your advisors.
To produce these responses, the AI bot evaluates the context of the conversation and uses your Knowledge Documents to provide accurate and relevant answers.
It can also leverage its AI skills to retrieve or send data to or from your information systems.
Create your AI Agents
In the “Settings / Alcmeon AI” section you have the option to configure the AI Agent(s) that will be involved in your bot scenario.
For more information, see our detailed article: Creating your AI Agent .
Configure your AI Bot scenario
Here is the classic chain of steps that will allow you to integrate AI responses into your Alcmeon bot scenario, which we will detail:
1. Retrieve the customer's question
The first part of this scenario branch involves asking the customer to ask a question and then passing that question, typed by the customer, to the AI.
For their very first interaction , greet the customer with a welcome message. Optionally, you can present them with a link to your data privacy policy. Then, ask them to submit their question. A "Waiting for a message" step will retrieve this question, and you can link it to your "AI Bot Connector" step. Above, a message asking them to wait has also been integrated: this is good practice because AI bots sometimes take ten seconds (or even longer depending on the complexity) to respond.
For subsequent conversations , the customer has already typed a message. Therefore, you can link this step directly to your AI Bot Connector. In the example above, we decide to greet them with an interim message and then have them wait before transferring their question to the AI.
In the event of an unexpected message , i.e. a client typing a message instead of clicking a button or option anywhere in the scenario, then we decide to transfer this message to the AI.
2. (Optional) Introduce the AI assistant
This step is optional, but we recommend it. It allows you to inform the customer that an artificial intelligence will now respond, and that this may take a few seconds, which is often the case with generative AI.
Connect the output of this step to the AI Connector step.
3. Handing it over to AI
This step will transmit the customer's question to your AI Agent who will be responsible for finding an answer, indicating that they could not answer, transferring to an advisor, or even detecting a specific intention if you have defined "custom exits".
Configure the AI Connector
Click on the "pencil" icon in this step to open its settings screen. You will find only one setting there:
AI Agent : In this dropdown menu, choose the agent that should respond. This is configured in the “Settings / Alcmeon AI” section.
Note : Regarding the maximum number of characters for a message, Alcmeon forces it to 1000, which ensures good readability for your customers and has the advantage of not exceeding the limit imposed by the channel, such as WhatsApp. Generative AI sometimes doesn't strictly adhere to this type of specification; if the limit is exceeded, Alcmeon will split the AI response into two messages.
Management of different exits
Below for each exit you will find a description of it, the bot event (1) which is triggered when this exit is taken, and our recommendation on how to link this exit to the next step.
Exit name |
Description |
Bot event |
Recommendation |
|---|---|---|---|
AI (end of discussion) |
The AI provided a valid response which was sent to the client. It considers this the end of the discussion. |
AI - answer (No need for an advisor) or AI - no answer (No need for an advisor) |
Link to an “AI Satisfaction” step |
| Continue with AI | The AI needs the client to send a new message, usually because the AI has responded with a question for the client. This creates a feedback loop with the AI. |
AI - answer (Continue conversation) or AI - no answer (Continue conversation) |
Link to a “Waiting for a message” step, which is itself linked to the AI connector. |
| Handover to an advisor | The AI detected that it needed to hand the question off to a human rather than answering it. Either because the customer requested it, or because the question was irrelevant… |
AI - answer (Advisor needed) or AI - no response (Advisor needed) |
Link to a message that passes the handle to an advisor + associated action and routing to the correct team. |
| Error | The AI encountered a technical error that prevented it from responding. | AI - no response (Technical error) | Link to an apology message and/or pass the matter on to an advisor. |
| Your Custom Exits | The AI has detected the intent you defined for this exit in your AI agent. Even if the AI decides to respond, this response is not sent to the client. | Your Custom Bot Event | Depends on your use case. |
(1) Bot events are internal tracking elements of Alcmeon, which allow you to obtain statistics on the usage of your bots in the “Analytics / Volumes / Bot Events” section
4. Measure satisfaction
When the AI Connector step takes the output “AI - response (no advisor needed)”, the AI response has been sent to the client, and it is therefore appropriate to ask the client what they think of this response. To do this, connect the output to an “AI Satisfaction” step.
In this step, you can configure:
- The text of the question asks the customer if the AI response seems relevant.
- The text of the “Yes” answer button
- The text of the "No" reply button
This stage has 3 exits:
- A “Yes” button to be linked to a message box with a menu to take leave of the customer, and possibly to offer them via a button to return to the beginning of the scenario.
- A "No" button to link to the scenario branch to transfer the customer's message to an advisor
- An output labeled “User Input” indicates that if the client doesn't click either of the two buttons above but instead types a new message, it's very likely a continuation of their previous question. We therefore suggest linking this output to the “Capture Message Content” Sub Bot step earlier in the scenario. Alcmeon will then signal to the AI that this new question is a continuation of the previous one, allowing the AI to respond with the correct context.
5. Handling cases of AI non-response
Here is an example illustrating our recommendations from the table above; you can use it as inspiration to build your own scenario!