When a client rings, the agent answers, asks who is calling and what it is about, and checks the answer against what you have approved it to say. A question about a filing date gets the date. A request for time with an adviser gets a slot offered from your calendar.
Anything the agent should not handle goes to your team with a short written summary, so the client does not have to repeat themselves. It does not give tax or financial advice, and it does not guess at a client's personal situation.
Before launch we test the script on real call examples from your own phones, including the awkward ones: a worried client, a wrong number, a caller who is not a client at all. You hear the recordings and adjust the wording. Only then does it answer live calls, and your team can still change the rules any time.
The most useful setup work is deciding what the agent may say and what it must pass on. We draft this with you in the blueprint and test it on real call examples before it goes live. We keep the list short at first and grow it.
Look at call logs after the first weeks. How many calls ended with a booked meeting? How many were passed on, and were they right to pass? Those two answers show whether the boundaries need moving. It is not a fit for firms that want every call answered by a named person, or for clients who need a considered view on their own position. Those calls belong with your qualified staff.
Yes. The agent introduces itself as an AI assistant, and callers can ask for a person at any time.