Will customers know they're talking to an AI?
Yes, if you want them to. The agent can introduce itself as an assistant on the first reply, and every AI message is tagged in your inbox.
Before and after Loomi
Before
Replies start when the shift starts
After
Replies in seconds, at 2pm and 2am
Before
The same ten questions answered by hand all day
After
Routine questions handled, the team takes the rest
Before
A bot that loops "I didn't understand"
After
A handover with the context attached
The honest answer
What does the agent get wrong, and when does it hand over?
It hands over when the customer asks for a person, when it has no grounded answer, when the customer is upset, and on anything you mark as human-only, such as refunds or complaints.
It can misread very short or ambiguous messages. It asks a clarifying question before it answers.
It doesn't negotiate, promise delivery dates it can't see, or give medical, legal or financial advice.
Every agent reply is logged, so you can read what it said and correct the source it used.
A good agent isn't the one that answers everything. It's the one that knows what it shouldn't.
[NEED: real handover rules from the product.]Questions
Can it take an order or a booking by itself?
Yes.
How do I train it?
Connect the sources it should answer from: your catalogue, price list, FAQs, policies and Shopify stock. It answers only from those. Edit a source and the next reply changes.
Does it cost extra?
Your plan includes 5,000 AI credits. An average AI chat uses about 20 credits, so the included credits cover roughly 250 chats. A $100 top-up adds 10,000 credits, enough for roughly 500 more chats at that average. Actual usage depends on conversation length and complexity.
Can it get my WhatsApp number banned?
It replies inside conversations customers started, which is the use WhatsApp's rules are built for.
