Why Autopilot asks before it acts

Most "AI for agents" pitches promise the same thing: sit back, and the software will work your pipeline for you. We built Autopilot on the opposite bet. On the Costa del Sol and the Costa Blanca, the difference between a warm buyer and a cold one is a single well-timed, well-worded message in the right language. That is judgement, and it belongs to you. So Autopilot does the reading, the drafting and the proposing. You do the deciding.

What "asks before it acts" means in practice

Every run ends the same way: with a proposal you can read, edit or throw away. Nothing leaves Fondaro until you approve it. In a normal morning, Autopilot will:

• Summarise where each active lead stands, in plain language, so you are not re-reading a week of notes.
• Draft the next follow-up for a buyer, in their language, using what they actually told you.
• Suggest which leads to work first, and say why.
• Flag the deals that have gone quiet and propose a way to reopen them.

Each of those is a suggestion sitting in a review panel. You approve the ones you like, tweak the ones that are close, and ignore the rest. The queue is the priority order; there is no separate scoring dial to babysit.

Grounded in your data, not in guesswork

An assistant is only as trustworthy as the context it reads. Autopilot does not invent facts about a buyer. Before it drafts anything, it pulls the real record: who they are, the language they want to hear from you in, the budget band they gave you, and when you last spoke.

{
  "lead": "Elin Karlsson",
  "language": "sv-SE",
  "budget": { "min": 450000, "max": 600000, "currency": "EUR" },
  "interestedIn": ["Estepona", "sea view", "2 bed"],
  "lastContact": "2026-07-08",
  "stage": "Viewing booked"
}

Because the draft is built from that record, the follow-up reads like you wrote it after checking your notes, not like a template. If the record is thin, the proposal says so rather than filling the gap with something that could embarrass you in front of a client.

Why we drew the line here

Two reasons, and both come from watching agencies work.

First, trust compounds. The moment an assistant sends one tone-deaf message on your behalf, you stop trusting all of them, and you go back to doing everything by hand. Keeping a human in the loop is not a limitation we will remove later. It is the product.

Second, the hard part of coastal property is not volume, it is language and timing. Your team might work in English, Swedish or Dutch while your buyers speak something else entirely. Autopilot is most useful exactly where a generic tool is most dangerous: writing something warm and correct in a language you may not speak. That only works if you can glance at the draft before it goes.

Where this goes next

Human-in-the-loop is the floor, not the ceiling. As the drafts earn your trust, we will add opt-in ways to let Autopilot handle more of the routine on a schedule you set. But the default will always be the same: it proposes, you approve, and you stay in control of every word that reaches a buyer.