Real estate

Customer Follow-Up Agent

Customer Follow-Up Agent

Case Study Image

Business Challenge

Business Challenge

A pre-sales and sales function handling high enquiry volume loses deals to silence more than to competition. Every enquiry, site visit, price discussion and objection requires a follow up, personalised enough to be worth reading, sent quickly enough to matter.

At scale this becomes the single largest consumer of a salesperson's week, and it is also the first thing to slip when the week gets busy. The leads that go cold are rarely the ones that were never going to close. They are the ones nobody got back to on Thursday.

Overview Image
Overview Image

What We Built

What We Built

A follow up agent connected to the CRM that reads what has actually happened with each customer, determines who is due for contact and why, and drafts the personalised message for the salesperson to review and send.

It does not send. A human sends, every time.

How It Works

How It Works

Model and surface
Claude, delivered through a prioritisation dashboard where each salesperson sees their own queue, with role-based access so a manager sees the team and an individual sees only their own leads.

Connectors
Salesforce, with live sync, for lead records, interaction history, call notes, site visit records, stage and last contact date. Gmail or Outlook for the send action, which stays with the human. WhatsApp Business where the client's follow up runs on that channel.

Skills
Two encoded layers. The first is the prioritisation logic: how the client scores a lead, what constitutes overdue for each stage, how site visit recency weighs against budget fit, and which combinations escalate to a manager rather than sitting in a salesperson's queue. The second is the communication layer: the client's own message structures for each situation, post site visit, post price discussion, objection follow up, silent lead revival, with the tone rules and the compliance boundaries on what may be committed in writing regarding price, possession or inventory.

Data flow
The agent reads the day's CRM state, applies the prioritisation skill to rank the queue, then for each lead reads the full interaction history and drafts a message specific to what was actually discussed. The draft references the real objection raised, the real unit type shortlisted, the real date of the visit. The salesperson opens their queue, reads, edits, sends.

Human review
Every message. The agent's role ends at the draft.

Governance
Role-based authentication. A salesperson cannot see another salesperson's leads. Nothing the agent drafts can commit the company to a price or a date not present in the connected approved source.

One Cycle End To End

One Cycle End To End

A salesperson opens the dashboard at 9:30. Fourteen leads are queued, ranked. The top one visited a site nine days ago, raised a specific concern about a floor plate's kitchen orientation, and has not been contacted since.

The draft addresses that concern directly, references an alternative unit on a different plate that resolves it, and proposes two slots. The salesperson changes one line, sends it, and moves to the next.

Fourteen follow ups that would have taken most of the morning take twenty minutes, and the ones that would have been forgotten are not.

Solution Image
Solution Image

Outcome

Outcome

8,784 hours saved

Automates follow-up drafting and surfaces dropped leads to improve conversion.

The Cutting Edge School

Enquire for Custom AI Trainings

We design and deliver practical, enterprise-ready AI training that turns everyday teams into high-output, AI-powered operators.

The Cutting Edge School

Enquire for Custom AI Trainings

We design and deliver practical, enterprise-ready AI training that turns everyday teams into high-output, AI-powered operators.

The Cutting Edge School

Enquire for Custom AI Trainings

We design and deliver practical, enterprise-ready AI training that turns everyday teams into high-output, AI-powered operators.