The most important points at a glance
SMEs improve accessibility without additional after-hours staffing by automating simple inquiries and forwarding only defined matters to staff. This is particularly sensible when outside core hours many standard questions come in, but adding staff would not be economical. The solution does not replace the team; it bridges gaps in after-hours periods and records callback requests in a structured way.
From our perspective at baselnetgroup, this is a typical case for small and medium-sized enterprises: The telephone switchboard should be reliably reachable, but in the morning, at lunchtime, or late afternoon no one can be permanently on the line. This is exactly where an AI-assisted telephony solution comes into play when it is properly prepared.
When is better accessibility without an extra shift worth it?
The trigger is usually not a big call center problem, but many small inquiries at the wrong time. In practice, it’s about callback requests, opening hours, simple forwards and short information questions that no one has to handle manually. If these calls increase but the workload for an additional after-hours shift is too small, there is exactly the gap that can be closed automatically.
AI is not a direct replacement for staff, but targeted for standardized inquiries and accessibility gaps.
What prerequisites does the solution in telephony need?
- Microsoft Teams and Teams Phone must be available.
- For staff handling calls, appropriate Teams Phone licenses are required.
- Central numbers need Teams Phone Resource Accounts.
- The Microsoft 365 tenant must be prepared for Nimbus, and a Microsoft Entra Global Administrator needs the necessary permissions.
- The public phone number must be reachable via Microsoft Teams Phone, e.g., via Direct Routing, Operator Connect or Calling Plan.
- For the AI Virtual User, a Nimbus service with Contact Center Enterprise Routing and a separate license per Virtual User is required.
What we see in projects: the technology alone is not the critical point. What matters is to clearly define in advance which issues the AI answers itself, which it captures, and when it forwards to a human. Without this boundary, accessibility improvement quickly becomes only a new phone menu.
How does it work in practice at baselnetgroup?
In a first project with AI telephony, a solution with Teams Phone and Luware Nimbus was deployed so that during core hours calls are distributed to the responsible staff. Outside core hours a Virtual User handles first contact, recognizes requests in natural language, and answers or captures simple inquiries. This is exactly why the solution is interesting for SMEs that already build their IT infrastructure and ICT infrastructure on Microsoft 365.
For example, a caller may call at 5:40 PM, want a callback after a conversation with Mr. Müller, and is asked for name, company and phone number by the Virtual User. Another caller immediately receives the opening hours for the coming Friday. This keeps the switchboard accessible without blocking additional personnel capacity in the after-hours.
| Situation | What the solution does | Benefit for the SME |
|---|---|---|
| Callback request in late afternoon | Name, company and phone number are captured | The callback lands in a structured way with the right team |
| Question about opening hours | The information is answered directly | Simple standard questions do not bind staff |
| Call outside core hours | The issue is recognized and forwarded or captured | Accessibility remains even without extra shift |
What are the biggest cost pitfalls in implementation?
- The AI is used too broadly, although only standardized inquiries are sensible.
- Lack of a clean separation between self-service, capture and forwarding.
- Public telephony is not cleanly connected to Teams Phone.
- Tenant administrator approval and the necessary roles are not clarified in advance.
- The company expects a replacement for staff rather than a supplement for after-hours.
Contrary to the assumption that automation makes service colder, it can actually relieve the workload in this case. Recurrent standard questions no longer bind humans, and staff have more time for calls that truly require expertise or personal clarification.
Do-it-yourself or implement with professionals?
For internal IT teams, the basic idea is comprehensible, but clean implementation requires experience with Microsoft Teams, telephony, permissions and call logic. At baselnetgroup we clearly see the boundary where multiple systems interact and the first contact becomes business-critical. It is then sensible to personally handle core hours and automate after hours specifically, rather than scheduling additional personnel.
Frequently asked questions
How many calls should you have for a solution like this to be worthwhile?
Can the solution also capture callbacks without anything getting lost?
What happens if a request is not standardized?
Is a completely new telephone system required?
For SMEs with 50 or more employees, the decisive point is usually not the technology but the operation logic: who should be reachable when, and which calls may be automated? This is where the lever lies to improve accessibility without occupying the switchboard with extra after-hours.
If you want to pragmatically relieve your company's telephony, it is worth looking at the existing Microsoft 365 environment and the types of calls that can be standardized. Experience from the first AI telephony project shows that small, recurring inquiries make the biggest difference.
Möchten Sie Ihre Telefonie gezielt entlasten?
baselnetgroup ag, Basel