An atHome enquiry arrives while the team is in viewings, on calls, or dealing with the work already in front of them.
That is normal. A busy agency should not have to stop everything for every incoming email.
The risk is quieter: the enquiry remains an item in an inbox until the right person sees it, has time to read it, checks what is available, works out the prospect's criteria, chooses a language and writes a reply. None of that is difficult in isolation. Repeated through a week, it becomes a fragile part of lead handling.
The useful question is not whether a real-estate agency needs more AI. It is whether the first response to a potential client should depend on an inbox being noticed at the right moment.
An enquiry should create visible work.
The point is not to automate the relationship. It is to make sure a potential relationship has a clear first step.
A shared inbox can receive a lead, but it does not automatically make the next step visible. If there is no clear flow after the email arrives, the team still has to reconstruct the same information manually: what the prospect wants, which active properties may fit, whether someone has already replied, and what should happen if the request is incomplete or unusual.
A dependable workflow gives those questions a home. It does not need to become a large new system. It can be a small, controlled sequence around the tools the agency already uses.
Keep the first workflow bounded
The strongest first automation projects have a narrow job. They are easier to test, easier for the team to trust, and easier to improve when a real case does not fit the expected path.
Read the incoming enquiry
Start with the message already arriving in the agency inbox.
Identify useful criteria
Use the details that are clear enough to guide an initial match, not guesswork about a prospect's unstated preferences.
Compare against active listings
Check one reliable catalogue rather than accelerating uncertainty across several partly correct sources.
Reply or route the exception
Respond in the prospect's language when the agreed rules are met. Send no-match, unclear or unusual cases to the team for a decision.
Four decisions to settle before building anything
Before connecting an inbox to a catalogue, agree the practical rules that give the workflow a safe boundary.
Trust one property source
Decide where active property information is reliable enough to use. A workflow should not have to reconcile three competing versions of the catalogue.
Define a useful match
Agree which criteria are dependable enough to use. Local knowledge and judgement still matter where a property or a prospect needs a closer look.
Name the exception owner
No match, incomplete request, uncertain availability or a language issue should have a visible person and next step. Stopping for a person is part of the design.
Measure what matters
Track the enquiries received, matched and escalated, the time to first response and the manual handling removed. Measure this workflow, not a generic benchmark.
Pilot the flow before expanding it
A sensible first step is a bounded pilot rather than a broad promise to automate the agency. Early examples reveal the details that are invisible in a process diagram: conflicting criteria, a listing that is technically active but unavailable, or a prospect who needs a different response from the standard path.
Check the inputs
Verify the incoming email, catalogue, languages and exception paths.
Use real agency cases
Validate each rule with the team before relying on it in day-to-day work.
Decide what to add
Expand to confirmations, reminders or follow-up only if the first workflow proves useful.
A pilot makes the commercial decision easier too. The agency is not committing to replace every process at once. It is testing one repeated piece of admin where a clear first response has obvious value.
What the Groupe Iris case study shows
Flowly's public Groupe Iris case study is useful evidence that this type of workflow can remove meaningful repeated work in a Luxembourg agency.
Median response
Lead emails handled
Active matches found
Manual work avoided
The useful lesson is not that every agency should expect the same numbers. It is that a narrow first-response workflow can be measured, inspected and improved. That makes it a stronger starting point than a vague plan to “use AI in the business”.
Start with last week's enquiries
Before choosing a platform or discussing a wider automation project, take ten recent atHome enquiries and trace what happened after each one arrived.
- •How quickly did the agency know who owned the next step?
- •How often did someone have to search for matching properties?
- •Which cases were straightforward, and which ones needed judgement?
- •Where did the team lose time copying, checking or following up?
A good outcome is not an agency that feels more automated. It is an agency where an interested prospect gets a dependable next step, while the team keeps control of the cases that need human attention.