AI call answering for after-hours, weekend and overflow calls
A call assistant that answers outside office hours and at busy times, records caller details and sends a daily summary.
The problem
A local services business, such as a solicitor's office, needed calls covered outside normal hours and during busy periods. It wanted this without hiring more staff and without an impersonal experience for callers.
Challenges
Covering only the hours that need it
The business needed cover for weekends, evenings and busy periods, not around the clock. The assistant had to be active only in defined windows, so the calls it takes are the ones the office cannot, and staff handle the rest as normal.
Working with whatever phones the client has
Local businesses run very different phone setups. Depending on the existing system, calls reach the assistant through scheduled VoIP rules, manual divert codes or full number porting, so routing had to be chosen to suit each client rather than assumed.
Keeping a personal feel for callers
A solicitor's office cannot afford calls that feel impersonal. The assistant had to answer, capture who is calling and why, and log everything in a way that lets the business follow up properly with each caller.
How the system works
Framing choice
We deliberately presented it as an AI call assistant rather than voice AI. That wording builds trust with non-technical, price-conscious buyers.
Scheduled coverage
The assistant is active only in defined windows such as weekends and after hours. Depending on the client's existing phones, routing uses scheduled VoIP rules, manual divert codes or full number porting.
Call handling
It answers incoming calls, captures who is calling and why, and logs everything systematically.
Daily reporting
A summary email of the day's calls is generated automatically, alongside a simple dashboard the client can check.
Commercial model
Pricing is a modest setup fee plus a per-minute usage rate, set with a margin above the underlying telephony and AI costs.
Design decisions
Calling it an AI call assistant
We present it as an AI call assistant rather than voice AI. For non-technical, price-conscious buyers the plain description builds trust faster, and it describes what the service does for them: answering calls when nobody else can.
Daily email summary and a simple dashboard
The client receives an automated summary of the day's calls by email, plus a simple dashboard. Busy owners rarely log in to another tool, so a short daily message gives visibility with almost no effort on their side.
Usage-based pricing
The commercial model is a modest setup fee plus a per-minute rate, set with a margin above the underlying telephony and AI costs. It keeps entry cost low and means the client pays in line with the calls they actually receive.
Result in production
After-hours and overflow calls are answered and logged without adding headcount, and costs scale with actual usage.
After-hours and overflow calls are answered and logged without adding to headcount.
The client sees who called and why through a daily summary and a simple dashboard.
Costs scale with actual usage instead of a fixed staffing commitment.
What we would do differently
We would build error handling and observability in from day one rather than adding them once the system was in production.
Who this pattern fits
Although built for local services, the same architecture suits other sectors with comparable call volume or coverage gaps. The stack is stable, self-hosted where critical and light to operate.
Tools used
ElevenLabs
Speech synthesis for the assistant's voice, chosen for quality and production latency.
Twilio
Telephony carrier for inbound and outbound calls, SMS and recording.
Automated email reporting
Daily summaries of calls handled.
Simple dashboard
Gives the client visibility of calls.
Related
Frequently asked questions
How do calls get to the assistant?
It depends on the client's existing phones. Routing can use scheduled VoIP rules, manual divert codes or full number porting, and the assistant is active only during the windows the client defines, such as weekends and after hours.
What does the client see afterwards?
The system answers, captures who is calling and why, and logs every call. The client then receives an automated summary email of the day's calls, and a simple dashboard gives visibility into call volume.
How is an AI call assistant priced?
In this build, pricing is a low one-off setup fee plus a per-minute usage rate, set with a margin over the underlying telephony and AI costs. Because it is usage-based, the client pays for the call volume they actually receive.