AI Agent ROI Calculator
Free training: Chatbot crash course
Whitepaper: The acceptance of chatbots
Guided Communication
--> Go to BOTwiki
Guided communication refers to a style of dialogue used in chatbots in which users do not type freely but are guided through a conversation using predefined buttons, carousels, quick replies, or drop-down menus. Instead of open-ended text input, the AI agent presents clearly defined options from which the user can choose.
In this way, Guided Communication deliberately distinguishes itself from purely NLU-based free-text dialogs. It is frequently used in scenarios where use cases are highly structured, a fast time-to-market is required, or reliable user guidance is a priority.
How Guided Communication Works
Technically, Guided Communication is based on a rule-based dialogue tree. In addition to the text, each of the bot’s responses includes a list of possible next steps, such as buttons displayed below a chat message. This makes the dialogue path deterministic: each selection leads to a clearly defined next step.
A simplified example from a service chat: After the greeting, users can choose between “Track Order,” “Change Appointment,” or “Speak to an Agent.” The user taps a button in the chat, and the agent guides them through the next steps. This structure reduces recognition errors because the system does not have to interpret open-ended phrasing.
Advantages and Limitations
Guided Communication significantly speeds up projects because it eliminates or at least reduces the need for time-consuming NLU training phases. At the same time, it increases reliability because users can only select valid paths.
- Faster implementation, as no extensive training is required for training phrases is required.
- Lower error rate, because there are no misunderstood free-text entries.
- Clear management of expectations, as users only see what is actually available.
- Limited flexibility when dealing with complex or open-ended issues.
- The risk of long menu tree structures.
In practice, however, it has become clear that simple button-based dialogs are not effective as the range of functions expands, and they place too many restrictions on the user. To ensure long-term scalability, it is best to combine guided communication with AI-based free-text processing, so that the AI agent can respond optimally to as many topics as possible.
Implications for Voice and Chat
Guided communication works differently in voice channels than it does in chat. A traditional IVR is essentially just a menu-driven system. It follows a “body without a brain” logic and does not allow for real conversation. A voicebot based on AI-Native Voice, on the other hand, can combine guided selection with natural language understanding, thereby combining the advantages of both approaches.
In chat, buttons, carousels, and quick replies are key elements of guided navigation. They are particularly well-suited for guiding users through form sequences, self-service processes, or multi-step selection questions.
Frequently Asked Questions (FAQ)
Guided Communication is a dialogue style in which users are guided through a conversation using predefined buttons, carousels, or dropdown menus. Instead of typing their own responses, they select from clearly defined options. This makes the dialogue path largely deterministic.
It is always a good choice when use cases are highly structured and precision is more important than conversational breadth. Typical examples include self-service processes, authentication steps, or brief selection dialogs in hotline triage. It is also a good option for projects with tight deadlines because it does not require extensive NLU training.
Simple button-based dialogs scale poorly as use cases become more diverse. Long menu trees frustrate users. Complex scenarios that fall outside the intended paths cannot be accommodated without resorting to a fallback.
> Back to BOTwiki
Trainings Phrases / Utterances
--> Go to BOTwiki
Utterances are training phrases that an AI agent uses to understand what users mean by comparing the examples stored in the system with the live query. This concept originates from traditional, intent-based dialogue architecture and has been a central element in the development of chatbots and virtual assistants for many years.
In modern agent-based systems based on large language models, the manual maintenance of utterances is becoming less and less important. These systems understand free-form text in context and can handle linguistic variation without having been trained in advance on specific example phrases. However, for those who work with intent-based NLU models or maintain existing systems of this type, the following section covers the essential basics.
Each utterance is a specific example phrase associated with a particular intent. By learning from many such examples, the NLU model learns to recognize linguistic variations of a query. The broader and more realistic the utterances, the more stable the classification in production.
What an utterance is and how it works
An utterance is a short example input formulated in natural language. For the "Book_appointment" “Book_Appointment,” these might include: “I need an appointment,” “Can you schedule an appointment for me?” or “I’d like to come in next week.”
The NLU model derives linguistic patterns from these examples. It then recognizes phrases that were not included verbatim in the training data. Utterances are therefore the training material that determines the quality of recognition.
Requirements for Good Utterances
For each intent, several utterances should be stored to cover different phrasings. The following are important:
- Linguistic variations: short and long phrases, with and without polite expressions.
- Synonyms: “appointment,” “meeting,” “consultation.”
- Realistic input from the target audience's perspective, not artificially constructed sentences.
- Coverage of common typos, colloquial language, and dialects, as they occur in practice.
It is equally important to ensure clear boundaries. Utterances with different intents must not overlap too much semantically. Otherwise, the confidence score will drop because the model can no longer make a clear decision.
Frequently Asked Questions (FAQ)
Utterances are training phrases that an AI agent uses to learn to recognize an intent. Each utterance is a specific example input, such as “I need an appointment” for the “Book_Appointment” intent. The NLU model derives linguistic patterns from many such examples, enabling it to recognize variations that were not trained on verbatim.
For Pro Intent, ten to thirty utterances are generally a good starting point. What matters is not the sheer quantity, but the linguistic variety. Short and long phrases, synonyms, and realistic typos improve recognition stability more effectively than additional similar sentences.
The Confidence Score indicates how confident the model is in assigning an intent. Well-formulated utterances increase the score because the model learns clear patterns. If utterances from different intents overlap, the score decreases and the system falls back on a fallback more frequently.
> Back to BOTwiki
Human Handover
--> Go to BOTwiki
A human handover (also human takeover, human handoff) is the forwarding of a conversation from a human to a real person. Chatbot to a real human being.
The term Human Takeover is usually used when the conversation is not handed over, but the person actively takes over a conversation.
Trigger for the Human Handover
A human handover can be triggered by different scenarios:
- Explicit question of the beneficiaries for a person
- The chatbot doesn't know the answer to a certain question (default fallback intent is hit)
- The chatbot is not confident enough (low confidence level)
- The sentiment of the users shows a negative value (Sentiment Score)
- A specific intent is made where human intervention is desired or required
- Certain metrics, such as the shopping cart of an online shop, contain products worth > 1,000 €.
Warm/ Cold Human Handover
A warm handover refers to the immediate forwarding of the user to a staff member. The human response is played out to the user promptly and in the same channel.
A cold handover, on the other hand, interrupts the flow of conversation and/or changes the channel. A common example of this is a handover from Facebook Messenger to the email channel.
Tools for Human Handover
A handover can be integrated into various tools:
- Social Media Engagement Tools
- Call Center Software
- Slack / Facebook Messenger
- Human Handover Tools
> Back to BOTwiki
Tone of voice
--> Go to BOTwiki
The tone of voice (or tonality) represents the specific way in which an AI agent in the form of a chatbot or voicebot. It encompasses the choice of words, sentence structure, and general communication style that shape the personality of the digital assistant. A consistent tone of voice strengthens brand identity and significantly influences the user experience in conversational AI.
Elements of tone of voice
The tone of voice of a chatbot or voicebot is determined by several components that together form its linguistic identity. Conscious design of these elements is necessary for coherent communication.
The choice of words and style determine whether formal or informal expressions are used and to what extent technical jargon is employed. Sentence structure is also taken into account, with a decision being made as to whether short, concise sentences or more complex formulations are preferred.
Emotionality and empathy emotionality and empathy also play a role, especially when dealing with user emotions, complaints, or errors. The use of pictograms and emojis can further complement the style and must be handled consistently.
Significance for conversational AI and AI agents
A well-defined tone is crucial to the success of conversational AI applications and AI agents . It strengthens brand identity and promotes recognition. This contributes significantly to establishing a consistent brand message.
In addition, the user experience (UX) is positively influenced. A pleasant and consistent tone of voice promotes user confidence in the digital assistant and increases its acceptance. Misunderstandings can be reduced through clear communication that is appropriate to the context. Consistent application of tone of voice across different channels and AI agents, including chatbots, voicebots, and within workflows, is essential.
Factors for determining tonality
Various factors determine the appropriate tone to use. These include the company's current communication style on other channels, the direct customer approach (e.g., "you" or "you"), the specific use case of the AI agent, and the target audience to be addressed. The company's values, beliefs, and ethical guidelines are also taken into account to ensure stylistic consistency across all forms of communication.
Questions regarding the determination of a good tonality may therefore include the following:
- How does communication with customers currently take place on other channels?
- How are my company's customers addressed? (You)
- Which expression suits my use case?
- Which target group do I want to address with the chatbot or voicebot?
- Which values, beliefs, and ethics does my company embody in terms of communication?
- Which guidelines and guidelines already exist with regard to external presentation?
- Should the chatbot use emojis ?
Control and adaptation in AI systems
The tone of voice of generative AI is shaped by the complex training process and the underlying database. In pre-training, AI systems learn general language structures and contextual relationships.
In the BOTfriends X platform, you can create and use your own AI agent personas, within which the language style is defined in detail. Parameters for this include, for example, the tone of voice, the language characteristics, or the length of the responses.
This AI agent personas are used to generate individual communication patterns and adapt responses to the respective application purpose. Retrieval Augmented Generation (RAG) is often used for company-owned knowledge databases to deliver more accurate and reliable information in the desired tone.
Frequently Asked Questions (FAQ)
How is the tone of voice of a chatbot or voicebot determined?
The tone of voice of a chatbot or voicebot is determined by the company's communication strategy, the target audience, the specific use case, and the style of expression established on other channels. Key design questions include the type of customer approach, the company's values, and compliance with existing guidelines for external communication. These considerations enable the creation of a consistent and brand-appropriate interaction style.
What role does tone of voice play in the user experience?
The tone of voice is very important for the user experience, as it influences how users perceive and accept information. An empathetic, factual, or informal tone of voice can improve interaction and increase trust in the AI agent. An inappropriate tone, on the other hand, can lead to misunderstandings or a negative user experience, especially when dealing with complex or emotional topics.
Can the tone of voice of AI systems be customized?
Yes, the tone of AI systems can be customized. In our BOTfriends X platform, the various aspects of tone are defined in separate AI agent personas that can be used flexibly.
> Back to BOTwiki
(Default) Fallback
--> Go to BOTwiki
In conversational AI, the (default) fallback is a predefined dialogue path that is triggered when the AI agent cannot process a user’s query or the confidence score is too low.
Back when chatbots still used fixed utterances and user inputs were matched to an intent based on predefined example sentences, the fallback served as a crucial safeguard when none of the predefined topics were recognized. The user was then directed to a special dialogue where a predefined response was displayed to prevent hallucinations.
Today, intent recognition has become significantly more flexible thanks to agent-based logic and large language models. An AI agent can respond contextually, resolve ambiguities on its own, and provide meaningful responses even when a query does not fit exactly into a predefined pattern. The concept of a fallback does not lose its importance, but rather shifts. Instead of a rigid intent handler, it is often sufficient today to specify in the instructions prompt how the agent should proceed in unclear situations.
When the fallback is triggered
If the confidence score for an AI agent’s intent recognition falls below a specified threshold, the request is considered unclassifiable. This triggers the fallback dialog, or the AI agent uses the action instructions stored in the prompt to respond to the situation. Typical triggers include incomplete sentences, topics outside the use case scope, or faulty speech recognition in the voice channel.
Even ambiguous inputs, where multiple topics are equally likely, can trigger a fallback. In modern AI agent platforms, the fallback is not viewed as a dead end, but rather as an active dialogue step that guides the user back onto a productive path.
Creating a Good Fallback Response
An effective fallback response follows a clear structure and avoids generic phrases. A three-part structure has proven effective, as it helps guide the user and keeps the conversation going rather than ending it.
- Clarifying the misunderstanding without blaming the user for the failure.
- A reminder of the AI agent's capabilities, to clarify which topics are covered.
- A specific recommendation for action, such as a sample question, a menu item, or transferring the matter to a colleague.
It is also important to vary the wording. If the same phrases are used every time a fallback occurs, the dialogue comes across as robotic, and trust in the agent drops noticeably. Having several alternative responses that rotate randomly or based on context keeps the conversation feeling natural.
Implications for Voice and Chat
In the voice channel, for example, with a voicebot in hotline triage, the fallback is particularly critical. Speech-to-text consistently produces recognition errors, and background noise and dialects further degrade the input signals.
In chat and email channels, user inputs are generally more concise but longer and cover a broader range of topics. In such cases, additional suggestions—such as topic overviews or quick reply buttons—can be displayed to quickly guide the user to the appropriate feature.
Frequently Asked Questions (FAQ)
A fallback is a specific response triggered by an AI agent when it cannot map a user’s input to a specific intent or when the confidence score is too low. It ensures that the agent responds in a controlled manner rather than guessing or remaining silent. This keeps the dialogue stable and coherent even when dealing with unclear queries.
The fallback is always activated when the user’s request cannot be clearly mapped to a defined topic. Typical examples include incomplete inputs, topics outside the scope of the use case, or recognition errors in the voice channel. Even if several topics are nearly equally applicable, a fallback with a follow-up question may be triggered.
A good fallback response clarifies the misunderstanding, briefly reminds the user of the AI agent’s capabilities, and provides a specific recommendation for action. It is important to use a respectful tone that does not belittle the user. Providing several alternative phrasings prevents the response from sounding robotic if the same issue arises repeatedly.
> Back to BOTwiki
Voice Bot / Assistant
--> Go to BOTwiki
A voice bot is a specific type of conversational user interface and is the counterpart to a chatbot. Conversational user interfaces enable users to communicate with computer systems using natural language. With a voice bot, input and output are handled through spoken language.
The technology of Voice Bots
The computer is able to convert the incoming speech into text using a Speech-To-Text Converter. The converted text is then interpreted and processed by the system using Natural Language Processing. The output of the speech is done by a Speech-To-Text technology.
For example, the Cloud Services Speech-To-Text and Text-To-Speech from Google and other providers can be used to convert spoken language into text.[1]
Application areas of Voice Bots
Classic voice bots are the Virtual Assistants Alexa from Amazon, Siri from Apple and the Google Assistant, which are mainly operated with voice. However, these can also be addressed by text input. Voice bots can also be found in the smart home area, where they can be used to control the lamps or the heating, for example, using voice commands. [2]
> Back to BOTwiki
Sources
[1] https://cloud.google.com/speech-to-text/
[2] Gartner IT Glossary, 2019, "Virtual Assistant".
Conversational Map / Conversational Flow
--> Go to BOTwiki
The conversational map, often referred to as the conversational flow, is a visual representation of the dialogue between a user and an AI agent. It maps out all possible paths a conversation can take: from the initial interaction through content-based branches to error cases and the handoff to a human agent. The map thus serves equally as a design tool, a basis for stakeholder alignment, and a blueprint for technical implementation.
Especially in more complex conversational AI projects, it serves as the foundation for ensuring that voice, chat, and email dialogues are not created haphazardly, but are deliberately designed. Without this visualization, the dialogue design remains hidden within text documents, making it difficult to coordinate effectively between business units, design, and development.
Why a Conversational Map Is Indispensable
A conversational map provides clarity on which issues an AI agent handles and how it responds to typical inputs. Stakeholders from business units, IT, service, and marketing gain a shared view of all conversation threads and can identify gaps, inconsistencies, or unclear wording early on. This allows them to add missing content, address weak points, and eliminate potential barriers to use before they arise in live conversations.
A good map clearly distinguishes between happy paths and edge cases. The happy path describes the ideal scenario, in which a request is immediately understood and resolved. Edge cases show how the agent responds to unexpected inputs, misunderstandings, or missing information. This distinction is essential for ensuring that conversational AI remains reliable in real-world scenarios and doesn’t just perform well in demos.
Structure and typical stages
The creation of a conversational map follows typical steps that span from the greeting to the end of the conversation. This process takes into account the media used, the features, and the desired style: a click-based bot, a free-text dialogue, or a hybrid of the two.
- Welcome message including a greeting, tone of voice, and introduction of the agent, along with a brief explanation of how the system works and the topics covered.
- Content levels with subject-specific branches, such as by intents.
- Error messages and prompts for entries with a low confidence score.
- Back button and navigation links to return to previous steps.
- Handover to a human employee if Knowledge AI or AI workflows are insufficient or if the issue cannot be assigned.
Based on this, dialog logic, buttons, dynamic content, and backend calls are linked together. This creates a consistent blueprint that is later directly implemented in the platform.
Implications for Voice and Chat
In the voice channel, for example, with a voicebot or a phonebot in hotline triage, the conversational map must be designed with particular care. Voice inputs are shorter, often ambiguous, and contain recognition errors from the speech-to-text step. Traditional IVR systems create rigid menu trees without semantic understanding. AI-native voice with multi-agent orchestration, on the other hand, can flexibly switch between specialized agents based on the map, maintain context, and ask targeted follow-up questions without forcing the caller through a series of options.
In chat and email channels, the focus is more on structure and readability. Buttons, cards, and structured lists complement the free-form text, and longer responses can be broken down into several steps. The Conversational Map ensures that the same technical logic remains consistent across all channels. Only the presentation adapts to the respective medium.
Conversational Map in Multi-Agent Setups
In modern conversational AI architectures featuring multiple specialized AI agents, the conversational map serves as the overarching choreography. It displays not only individual dialogues but also the handoffs between agents: from the triage agent to contract management, and from self-service to escalation. Hybrid intelligence is explicitly modeled here by clearly marking the points at which the agent makes autonomous decisions and the points at which reliable business logic is used.
The map is transferred from the design phase to the platform, serves as a reference for optimizations during operation, and provides the foundation for seamlessly adding new use cases without disrupting existing workflows.
Frequently Asked Questions (FAQ)
A conversational map is a visual representation of the dialogue flow between users and an AI agent. It shows all relevant paths, from the greeting through subject-specific branches to error handling and handoffs, and serves as a shared working document for the design, business, and development teams.
These terms are often used interchangeably. In practice, the conversational map provides a general overview of all possible dialogues, while conversational flow typically describes the specific progression of a single interaction. Both perspectives complement each other and are combined in a structured document.
Typical examples include welcome messages, onboarding, levels of technical content, error and query logic, back navigation, and the handoff to a human agent. In addition, media, buttons, and integrated features such as email sending or backend queries are highlighted.
The map is particularly important for voice and phonebot projects because voice dialogs are less forgiving than chat. It helps define concise prompts, clear follow-up questions, and logical escalation paths. This results in a dialogue that feels natural and can also be reliably orchestrated between specialist agents and human teams in multi-agent setups.
> Back to BOTwiki
Entity
--> Go to BOTwiki
Entities are used to extract user information from natural language.
A distinction is usually made between System Entities and custom entities. System entities are entities already contained in the system for addresses, times and numbers, for example.
Custom entities, on the other hand, can be defined by the user and contain, for example, product information or a staff directory, depending on the use case.
Entities in NLP Services
The Entity Recognition is already integrated in the current NLP Services [1][2][3] and is, besides the Intent Matching, the main component of the Natural Language Processing for chatbots.
Example for an Entity Extraction
I would like to order a small Pizza Margherita to Berliner Straße 1.
Entities:
small (Custom Entity pizza_size)
Margherita (Custom Entity pizza_type)
Berliner Straße 1 (System Entity street_address)
> Back to BOTwiki
Sources
[1] https://dialogflow.com/docs/intents
[2] https://docs.microsoft.com/en-us/azure/cognitive-services/luis/luis-concept-utterance
[3] https://cloud.ibm.com/docs/services/assistant?topic=assistant-intents
BOTwiki
A
Agent Tool
AI Task
AI Agent Persona
AI Workflows
Aleph Alpha
Artificial General Intelligence (AGI)
C
Channel Connector
Chatbot
Chunking for RAG
Clustering
Collected Data
Confidence Score
CCAI (Contact Center AI)
AI Context
Context Window
Contextual Awareness
Conversational Analytics
Conversational AI Platform
Conversational Copywriting for AI
Conversational Design
Conversational Map / Conversational Flow
Conversational Testing
Custom GPT
Custom Voice
E
F
G
Generative AI
Guided Communication
H
AI Hallucinations
Happy Path
Human Handover
Human in the Loop
Hybrid Human Chatbot
I
AI Instructions
Insult Rate
Intent
K
L
LangChain
LangSmith
AI Latency
M
Machine Learning
Markdown Prompts
Messaging Services
Natural Language Generation
Natural Language Processing
Natural Language Understanding (NLU)
O
On-Premises
OpenAI
AI Agent Operations
P
Filler
AI Agent Persona
Phonebots / Voicebots
AI Agent Prebuilts
Prompt Engineering
Prompt Injection
Prompt Jailbreaks
R
RAG (Retrieval-Augmented Generation)
RCS - Rich Communication Services
Rich Media Elements
Reasoning
S
Semantic Search
Sentiment Analysis
Session Initiation Protocol (SIP)
Speech-to-Speech
Speech-to-Text
Streaming Response
System Entities
T
AI Temperature
Text-to-Speech
AI Tokens
Tone of Voice
Agent Tools
Chatbot Training
Training Phrases / Utterances
Transformers
Triage Agent (Routing Agent)
V
Vector Database
Phonebots/Voicebots
Voice Cloning
W
Wizard of Oz Experiment
Word Error Rate (WER)
Haven't found what you're looking for? Maybe you'll find the right answer in our webinars.
Wizard of Oz Experiment
--> Go to BOTwiki
In the field of human-computer interaction, the "Wizard of Oz" method refers to a research experiment in which participants interact with a computer system that they perceive as autonomous, but which is in fact operated or partially controlled by an unseen person.
Chatbots and Wizard of Oz
Chatbots are enormously suitable for the Wizard of Oz experiment. In this way, a use case can already be examined for its"chatbot suitability" before implementation. The findings can then be used to iteratively expand existing flows and define new communication strands. In addition, the collected data, such as utterances, can be used directly for the Chatbot Training be used. A mature human handover tool is even able to automatically convert the tested data into a chatbot.
Recommended procedure
- Define different chatbot flows
- Integration of a live chat or an empty chatbot that only triggers a human handover.
- Manual response to user queries
- Derive chatbot flows
- Answering the queries based on the chatbot flows
- Iterative revision and expansion of the chatbot flows
- Manual or automated transfer of the flows into a chatbot builder
> Back to BOTwiki
The Wizard of Oz experiment is a UX research method in which users interact with an apparently autonomous system whose responses are formulated by an unseen human. In the context of conversational AI, the method is used to realistically test the dialogues of a planned AI agent before the system goes live. This allows use cases to be validated and real user data to be collected without the NLU model having to be trained yet.
A WoZ test is particularly worthwhile when a new use case for an AI agent is being planned and the requirements for dialogue management or tone are still unclear. The method also helps identify risks early on in sensitive channels such as voice hotlines, where there is little room for error. It is also useful when there is internal disagreement about whether a topic can be automated at all.
The experiment yields both qualitative and quantitative data: typical user comments, response patterns, escalation rates, common misunderstandings, and knowledge gaps. This information is incorporated into intent models, dialogue flows, and Knowledge AI content. As a result, the experiment serves as a direct precursor to the productive training of an AI agent.
While a traditional prototype test evaluates a system that has already been implemented, the Wizard of Oz experiment examines a use case whose logic does not yet exist in technical form. The Wizard replaces the model and provides flexible responses, enabling dialogues that are much more realistic. This yields insights that can be addressed during the early conceptual phase, rather than only becoming apparent after technical implementation.