AI & Machine Learning
AI agents: what to prepare before we build
a field-service company needs AI agents you can open tomorrow morning, not a demo. The agent sorts incoming messages, drafts a reply and marks what is urgent. It does not send anything to a customer before you approve it.
Before building AI agents for a field-service company, you do not need a fifty-page spec. You need the work as it is. Bring ten real messages from this week, without names if you must, and the reply you would have sent yourself.
Also bring one sentence on what is forbidden: what the dispatcher, the technician and the customer should not see, and what must not be sent alone. The agent does not close a deal, change a price or delete a record. A person keeps those.
The message lands in the information system, the agent suggests a plain reply, and you approve or correct it in one pass. If the people who actually do the work do not recognize those steps, you are not ready to build. You are ready for a conversation.
Do not bring a brand, a logo and colors before you bring the customer card. The look comes after the work is clear.
The agent sits on the information system and on the WhatsApp bot. Without the system it answers in a vacuum. Without the bot the customer stays in a separate chat. In the morning you see a queue of drafts, not an endless inbox. What you approved yesterday already went out. The rest is waiting.
For a field-service company, “what to prepare before we build” is a practical test of AI agents, not a headline. the dispatcher, the technician and the customer need to know what is going on without asking in the group. In the morning you see a queue of drafts, not an endless inbox. What you approved yesterday already went out. The rest is waiting. Bring ten real messages from this week, without names if you must, and the reply you would have sent yourself. The agent does not close a deal, change a price or delete a record. A person keeps those. The example stays the work of a field-service company: jobs in a group, addresses in a voice note, and job completion on the technician’s phone. If you cannot point to one step that enters the system and one step that stays with a person, you do not yet understand their day, and there is no point building a screen. The call is booked only after there is a real example from this week, not a general description, and the system stays the business’s.
If a field-service company wants AI agents as part of one software house, the next conversation is about the work itself: AI agents, a secure information system and a WhatsApp bot as an official Meta provider, with human approval before anything goes to a customer. We do not add clients or numbers we do not have.