Smart Argument Logic
Creation of a smart AI agent
Specifications
Basic Settings
AI agents offer an innovative way to efficiently handle customer inquiries and automate standard processes.
Thanks to flexible configuration, agents can be precisely tailored to individual requirements. The creation is done in a few, clearly structured steps, described under the tabs listed above.
1. Define Specifications
This is the foundation for the AI agent. This includes:
-
Name: The agent’s name should reflect its remit, for example, “Instalment-Agent”.
-
Description: A short description helps grasp the purpose of the agent at a glance.
-
Topic: The topic arises from the available skill tags. For instance, an instalment agent could be assigned the skill tag “Payment Transactions”. The assignment ensures a clear organisational categorisation and, for example, simplifies access via the template catalogue.
-
Channels: Determine which communication methods the AI agent should operate – Email, Chat, Letter, Telephone, etc. An agent can easily be assigned to multiple channels.
2. Choose an Avatar
The avatar gives the AI agent a visual identity and emphasizes its individual character. Various options are available:
-
Style: Should the avatar appear more professional, friendly, or neutral?
-
Colour accents: Additionally, colours can be chosen that, for example, match the brand identity.
-
Recognition value: A uniform avatar across all channels enhances the visual appearance of the AI agent.
3. Configure Intelligence
This is where it’s decided how the AI agent solves tasks:
-
Smart argumentation logic: Rules in natural language without programming – ideal for standard processes.
-
Rule-based logic: Code-based workflows for more complex procedures.
4. Adjust Personality
The AI agent’s personality defines how it communicates – not just in content, but also in tone. The communication style can be fine-tuned using intuitive sliders:
-
Informal to formal: Should the AI agent be casual or businesslike?
-
Speaking style: From brief answers to detailed explanations.
-
Concise to comprehensive: Determines the level of detail in the feedback.
These adjustments ensure that the AI agent perfectly matches the company’s language and positively supports the customer experience.
All specifications – name, topic, intelligence, personality, and channels – can be adjusted flexibly at any time. This keeps the AI agent dynamic and always adjusted to current requirements.
Specifications
Basic Settings
AI agents offer an innovative way to efficiently handle customer inquiries and automate standard processes.
Thanks to flexible configuration, agents can be precisely tailored to individual requirements. The creation is done in a few, clearly structured steps, described under the tabs listed above.
1. Define Specifications
This is the foundation for the AI agent. This includes:
-
Name: The agent’s name should reflect its remit, for example, “Instalment-Agent”.
-
Description: A short description helps grasp the purpose of the agent at a glance.
-
Topic: The topic arises from the available skill tags. For instance, an instalment agent could be assigned the skill tag “Payment Transactions”. The assignment ensures a clear organisational categorisation and, for example, simplifies access via the template catalogue.
-
Channels: Determine which communication methods the AI agent should operate – Email, Chat, Letter, Telephone, etc. An agent can easily be assigned to multiple channels.
2. Choose an Avatar
The avatar gives the AI agent a visual identity and emphasizes its individual character. Various options are available:
-
Style: Should the avatar appear more professional, friendly, or neutral?
-
Colour accents: Additionally, colours can be chosen that, for example, match the brand identity.
-
Recognition value: A uniform avatar across all channels enhances the visual appearance of the AI agent.
3. Configure Intelligence
This is where it’s decided how the AI agent solves tasks:
-
Smart argumentation logic: Rules in natural language without programming – ideal for standard processes.
-
Rule-based logic: Code-based workflows for more complex procedures.
4. Adjust Personality
The AI agent’s personality defines how it communicates – not just in content, but also in tone. The communication style can be fine-tuned using intuitive sliders:
-
Informal to formal: Should the AI agent be casual or businesslike?
-
Speaking style: From brief answers to detailed explanations.
-
Concise to comprehensive: Determines the level of detail in the feedback.
These adjustments ensure that the AI agent perfectly matches the company’s language and positively supports the customer experience.
All specifications – name, topic, intelligence, personality, and channels – can be adjusted flexibly at any time. This keeps the AI agent dynamic and always adjusted to current requirements.
Detection
Methods and Configuration
In the Detection section, it is defined when an AI agent is assigned to a ticket. This can be based on keywords, ticket data or AI analysis.
Authentication
If this option is activated, requests from non-authenticated customers will be ignored by the AI agent. This setting is particularly suitable for processes involving sensitive data or requiring customer verification.
Detection Methods
Various methods are available for assigning AI agents:
AI Detection
AI Detection
AI-based detection relies on machine learning. Instructions assist the AI agent in identifying patterns in customer queries. A clear and precise instruction improves the detection accuracy of the AI agent.
Example
Instruction to the AI agent for customer concerns on the topic “SEPA Reversal”:
Condition Detection
Condition Detection
The AI agent assignment is based on ticket, customer, and/or contract attributes. Conditions can be combined as desired.
Example
Linking conditions for customer concerns for marketing purposes:
AI and Condition Detection
AI and Condition Detection
AI agents are assigned when both methods apply.
Example
Here, the AI agent is assigned when the AI training data applies, AND the customer’s message does not refer to cancellation or moving:
AI or Condition Detection
AI or Condition Detection
AI agents are assigned based on either of the two methods.
Example
Here, the AI agent is assigned if the AI training data applies, OR the customer’s message does not refer to cancellation or moving. The assignment would probably apply to a great many cases here:
Manual Detection Only
Manual Detection Only
AI agents are assigned exclusively manually by an operator.
Instructions
Prompt, Tools and Instruments
Instructions for smart AI agents are precise prompts defining how to respond to customer concerns. They guide the behavior of the AI agent, specifying which tools are used and which actions are taken.
For each channel, an individual prompt can be set up. If no individual prompt exists, the standard prompt automatically applies:
Example Prompt of a Tariff Consultation Agent with Custom Tools
You are an expert in customer retention for energy suppliers. As a tariff consultant, you want to politely but effectively encourage the customer on the phone to opt for a favorable tariff instead of going through with the termination. Here are the steps:
-
If the customer says the termination was incorrect, note this in and revoke the termination.
-
Otherwise, ask why the customer has terminated. Record the reason in .
-
Then ask if you can give the customer a prepared switch offer to save money. Take into account the reason for termination where applicable. If rejected, accept it and end the call.
-
If the customer is interested, fetch the offer with . Briefly present the offer by focusing on the highlights. Formulate numbers in words (e.g., 1.5 → one point five). Use natural language, not JSON variables or acronyms such as kWh. If the customer doesn’t have any interest, end the call.
-
Answer customer queries regarding this tariff upon request. If the customer wants the tariff to be sent by email, use and confirm the dispatch by stating the mentioned email address.
-
Ask if the customer would like to accept the offer. If agreed, initiate the tariff change with . Otherwise, accept the decision and end the call.
-
Confirm a successful switch to the customer and end the call.
-
If the customer wants a calculation of the savings, use the tool .
Focus only on this process and don’t handle other topics. You already know the customer, identification is unnecessary. Use formal address for the customer.
Custom Tools
In the area AI Customization → AI Tools you can create custom tools and add them to a smart AI agent at a later time.
These tools help the AI agent to efficiently and purposefully carry out processes and instructions. Depending on the configuration, the AI agent’s actions are triggered, adjusted based on the identified customer needs, and completed in the defined process.
- A custom tool describes a single action – for instance, “revoke termination” – that can be later called up by the smart agent.
- The information required by the tool (e.g., a contract number) is determined in advance and is automatically transferred when used.
- The actual action can either be defined directly by code or forwarded to an external interface.
Integrated Tools
Under the “Instructions” of a smart AI agent, Integrated Tools can be added. They extend the capabilities of the AI agent by providing access to relevant data sources and supporting standardized processes. They are pre-configured and ready to use immediately.
Testing and Publishing
Operating Mechanism
Test cases simulate real scenarios to ensure that the AI agent is functioning correctly. Each test case is based on a ticket ID and reflects the entire processing process. Regular tests ensure that the AI agent operates stably and reliably, even with changed requirements or system updates.
Test Procedure
-
A real ticket ID is selected and added.
-
The AI agent processes the ticket based on the defined logic.
-
The result is displayed as successful or failed.
Result Interpretation
-
Successful: The test confirms that the AI agent is responding as intended. The AI agent can now be published.
-
Failed: In this case, the existing settings need to be examined. Instructions, detections, input parameters, business logic, and/or output handling should be investigated for possible inconsistencies. After making the adjustment, the test should be run again.