Hello, you're speaking with the AI assistant of Hans & Partner. How can I help you?
An AI assistant that KNOWS YOUR WORKFLOWS.
The assistant takes requests, handles follow-up questions and completes agreed tasks. That way, it supports your team with incoming and outgoing calls.ab15 €mtl.
Listen to a conversation.A caller wants to reschedule their appointment. In this example, the assistant clarifies the request, verifies the caller's identity and finds a new appointment.
Hi, I'd like to reschedule my appointment next week.
What can be HANDLED BY PHONE?
We start with concrete tasks. Together, we decide what the assistant may handle on its own and when a person takes over.When customers call.
The assistant can take requests, ask for the information needed and initiate the right next steps. For example, handling an appointment request, recording a change or forwarding a request to the right person.Which information it may give out and which systems it needs for that are defined for your specific use case.When something needs following up.
Outgoing calls can also be part of an agreed workflow – for example, to coordinate an appointment change or to clarify missing information about an existing request.We define the reason for the call, recipients, approvals and handover together with you. The assistant acts within this defined framework.The assistant answers. Your team GETS THE DETAILS.
“Hi,
I have a question
about my order.”
“Can you help me
with an appointment?”
“I'd like to
arrange a callback.”
Request captured
Key details noted
Passed to the right team
Available, even when your team is busy.Speak naturally instead of tapping through phone menus.Take requests and pass them on, ready to go.
Good conversation design can WIN PEOPLE OVER.
76,5 %rated AI phone calls more positively after the test call.
5.46 out of 7
Even skepticism can change.
Among the participants with the strongest initial rejection, the average attitude toward AI phone calls rose from 1 to 3.67.+2.67 points after the test call1.00 before
3.67 after
12345
Strongly opposedVery open
End-user study with 51 participants on phone assistants designed with the Canvas, 2026.
Who takes over?
This is what a jointly agreed division of tasks can look like.AI handles it
For example, providing information from approved sources.AI prepares
Taking a request and collecting the details for a callback.Your team takes over
For difficult requests or when the caller asks for it.We plan for this too:What happens if no one is available?
Before the first call, THE PLAN IS IN PLACE.
First, we clarify who calls, which request should be resolved and when your team takes over.For this, we use our Conversational AI Design Canvas. It is based on a scientific publication by Pascal Slaghekke on the use of conversational AI in customer service. In three phases, it leads from analyzing your calls, through clear guidelines, to concrete conversation design.The result is a shared briefing, a system prompt and a prioritized task list for implementation – including the necessary integrations and any information still missing.Tasks and boundaries ARE SET IN ADVANCE.
What is the assistant allowed to do?
Reading information, preparing changes and carrying out an action are different permissions. We define which action is allowed in which conversation and when confirmation is required.What happens when it's unsure?
Not every request can be resolved automatically. We define when the assistant asks follow-up questions, which information it hands over to your team and what happens when no one is available – for example, an agreed callback instead of a transfer into the void.Where does the information come from?
Reliable information needs a solid basis. We discuss which documents, appointments and systems are connected and which data is actually needed in the conversation.Our method IS BASED ON RESEARCH.
Our Conversational AI Design Canvas is based on a scientific publication by Pascal Slaghekke.“Empirische Entwicklung eines Design Canvas für Conversational AI im Kundenservice” (2026) combines scientific literature with original studies. The result is a practical tool: it helps us ask the right questions before your assistant takes its first call.In the qualitative interviews, uncomplicated interaction, flexibility and appropriate technical terminology were among the things that mattered. We translate this into concrete decisions about language, follow-up questions and conversation flows.The guidance notes explain conversation design, tasks and responsibilities, with references to the underlying research. That way, decisions can be justified and reviewed later.
Analysis
What do your business and your callers need?1. Problem space & motivation2. Define the target group3. Tasks & evaluationGuidelines
Within what framework may the assistant act?4. General conditions5. Human–AI boundaryDesign
How does this become a helpful conversation?6. Task design & integration7. Persona & interactionIt all comes down to HOW YOU SPEAK.
A pleasant voice is a start. What really matters is how the assistant behaves in the conversation.Room to answer.
Assistant“Which day works for you?”Leave time for the answer.
Ask, rather than carry on wrongly.
Assistant“Sorry, I understood Tuesday. You mean Thursday – is that right?”
A clear next step.
Assistant“I've noted your callback request. Our team will get back to you.”
Design examples based on the guidance notes for “Persona & interaction” in the Design Canvas. We tailor the wording to your business.
After launch, KEEP IMPROVING.
Whether the assistant helps becomes clear in your customers' everyday lives.SERVPERF
How is the service experienced?
After the call, THE WORK CONTINUES.
A captured request only helps once it reaches the right place.We plan the phone assistant in the context of your work: who handles the case? What information is needed? Which steps can be automated? That way, the conversation becomes part of a traceable workflow.