How a 12-person real estate agency automated 80% of their admin in 8 weeks

The Agency Before Automation

When Meridian Property Partners first contacted PURIST in October 2025, they were a 12-person estate agency operating across three postcodes in South West London. Two senior partners, four sales negotiators, two lettings negotiators, an office manager, an administrator, and two part-time viewing coordinators. By revenue, they were performing well average monthly completions were solid, their repeat and referral rate was above the London average, and their Rightmove and Zoopla ratings were consistently strong.

The problem was invisible in the revenue numbers. It was visible in what their people were actually doing each day.

Sophia, one of the senior negotiators, described it plainly in our initial discovery call: "I spend the first 90 minutes of every morning updating listings. Rightmove, Zoopla, our own website, our internal spreadsheet. The same information, four times, manually. By the time I am done with that, I have missed half the morning's incoming enquiries. By the time I get to those, the best leads have already called another agency."

Their office manager, Marcus, had a different but equally revealing description: "Every Friday afternoon I spend two to three hours compiling the weekly report for the partners. Viewings booked versus viewings completed, offers made, pipeline by stage, lettings renewals coming up. Every single number I type manually from six different places. It is always the same information. Nothing about it requires a human. But it is always me doing it."

The administrator, Priya, spent approximately four hours per day on two tasks: sending initial property details packs to enquirers (email with attached PDF brochure, floorplan, and EPC), and chasing buyers and tenants for documents after an offer was agreed. Neither task required any judgment. Both required someone to do them, consistently, every day.

In our initial audit, we identified six pain points that were collectively consuming an estimated 48 staff hours per week in manual, rule-based work.

The Six Pain Points Before Automation

Pain Point 1 Manual Listing Updates Across Four Platforms

Every time a property status changed new listing, price reduction, under offer, sold, let agreed the negotiator responsible needed to manually update Rightmove, Zoopla, the agency website, and the internal spreadsheet. A single status change took 12 to 18 minutes and required navigating four separate interfaces. With an average of 14 status changes per day across the portfolio, this was 2.8 to 4.2 hours of daily manual updating. Over a week, that was 14 to 21 hours of senior negotiator time spent on data entry.

Pain Point 2 Lead Routing Without a System

Inbound enquiries arrived from five sources: Rightmove (email notification), Zoopla (email notification), the website contact form, direct phone calls logged manually in a notebook, and walk-ins noted on a physical sheet. Each enquiry needed to be routed to the appropriate negotiator based on property type, postcode, and the negotiator's current caseload. This happened through a group email where the office manager would reply-all to assign the enquiry, and the negotiator would then respond to the original enquirer. Average time from enquiry arrival to first response: 2.5 hours.

Pain Point 3 Viewing Scheduling Was a Phone-Heavy Process

Booking a viewing required the negotiator to call or email the enquirer to establish availability, call or text the vendor or landlord to confirm access, then call or email the enquirer back to confirm the slot. Three separate communications for every viewing booking. With 25 to 35 viewings booked per week, this was between 75 and 105 individual communications per week roughly 8 to 12 hours of negotiator time on viewing logistics alone.

Pain Point 4 Follow-Up Was Inconsistent

Post-viewing follow-up depended entirely on the individual negotiator. Some called every viewer the day after their viewing. Some sent a template email. Some did neither unless the viewer contacted them first. The inconsistency meant that the agency's conversion from viewing to offer was lower than it should have been and nobody could diagnose exactly why, because the follow-up data lived in individual email inboxes and not in any shared system.

Pain Point 5 Document Sending Was Entirely Manual

Two document-sending moments consumed disproportionate time. First: when a new enquiry arrived, Priya manually assembled and sent the property information pack (brochure PDF, floorplan, EPC, area guide). This took 4 to 6 minutes per enquiry, and there were 40 to 55 enquiries per week 2.5 to 5.5 hours per week for a single repetitive task. Second: when an offer was agreed, the solicitor referral, ID verification request, proof of funds request, and mortgage broker referral all needed to be sent to the buyer and to the vendor. Eight separate documents to four parties. Priya was spending 20 to 30 minutes per agreed offer on document assembly and sending, and there were 6 to 10 agreed offers per month.

Pain Point 6 Monthly Reports Were a Friday Afternoon Destruction

Marcus's Friday afternoon report took 2 to 3 hours every week. It covered viewings, offers, pipeline, lettings renewals, and vendor reports for properties on the market for more than 45 days. Every data point was manually gathered from a different source: the viewing log (a shared Excel file), the offers spreadsheet (another Excel file), the Reapit CRM for pipeline data, and a separate spreadsheet for lettings renewals. The report went out to the partners on Friday evening and was stale by Monday morning.

The 8-Week Implementation Timeline

We structured the implementation as four two-week sprints, each focused on a distinct part of the automation stack.

Weeks 1 and 2: Discovery, integration setup, and data normalisation. Mapped all existing tools (Reapit CRM, Rightmove feed, Zoopla feed, website CMS, Xero for billing), established API connections, and built the unified property data model that would serve as the single source of truth for all automation.

Weeks 3 and 4: Listing update automation and lead routing. Built and tested the property status sync automation across all four platforms, and the enquiry ingestion and routing system with Slack notifications.

Weeks 5 and 6: Viewing scheduling, post-viewing follow-up, and document sending. Built the Calendly-integrated viewing booking system, the automated follow-up sequences, and the document assembly and sending workflows.

Weeks 7 and 8: Monthly reporting automation and testing. Built the automated reporting workflow, ran full system integration testing across all seven automations, conducted team training, and monitored live operations with close support.

The full implementation ran over 8 weeks from kickoff to live operation of all seven automations. The agency was live on each automation sequentially as it was completed, so they started seeing time savings from week 3 rather than waiting until week 8.

The 7 Automations: Architecture and Tools

Automation 1 Property Status Sync Across All Platforms

Goal: eliminate manual multi-platform listing updates.

Architecture: the source of truth for property status is the Reapit CRM. A webhook integration fires from Reapit whenever a property record is updated (status change, price change, or property going under offer). This webhook triggers an n8n workflow that reads the updated property data, formats it for each target platform, and pushes updates simultaneously to Rightmove via the Rightmove Data Feed API, Zoopla via the Zoopla Listing API, and the agency website via its WordPress REST API.

For the internal master spreadsheet, the n8n workflow updates a Google Sheet row via the Google Sheets API, with a timestamp of last update.

Tools: n8n (orchestration), Reapit webhooks (trigger), Rightmove Data Feed API, Zoopla Listing API, WordPress REST API, Google Sheets API.

Time saved: 14 to 21 hours per week. Eliminated entirely. Property status updates now happen in under 90 seconds across all platforms after the Reapit record is updated.

Automation 2 Enquiry Ingestion and Smart Routing

Goal: route all inbound enquiries to the right negotiator within 5 minutes of arrival.

Architecture: we created a single unified enquiry ingestion point using n8n. Rightmove and Zoopla enquiry email notifications are parsed by an n8n email processing workflow via IMAP. Website form submissions trigger an n8n webhook. For phone and walk-in enquiries, we added a simple internal web form that the office manager completes in under 60 seconds.

All enquiries, regardless of source, flow into the same n8n routing workflow. The workflow extracts the property reference, enquirer name, email, and phone number, queries the Reapit CRM to get the responsible negotiator for that property, checks negotiator availability (queried from a Google Sheet that negotiators update when they are on viewings or out of office), and assigns the enquiry to the next available negotiator following the capacity rules.

Assigned enquiries trigger a Slack DM to the negotiator with all enquiry details, the enquirer's contact information, and the specific property they enquired about. The negotiator can respond directly to the enquirer from the Slack notification.

A Claude AI node analyses the enquiry message text for urgency signals ("looking to move quickly", "made an offer elsewhere", "chain free") and flags high-urgency enquiries in the Slack notification with an urgent indicator, prompting faster response.

Tools: n8n (orchestration and routing), IMAP email processing, Claude AI (urgency classification), Slack (negotiator notification), Google Sheets (availability tracking), Reapit API (property and negotiator lookup).

Time saved: first response time dropped from 2.5 hours average to 18 minutes average. The office manager's daily enquiry routing task was eliminated entirely.

Automation 3 Automated Viewing Scheduling with Calendly

Goal: book viewings without three-way phone-tag between negotiator, enquirer, and vendor.

Architecture: when an enquiry is routed to a negotiator via Automation 2, the notification includes a link to a Calendly booking page pre-configured for that property's available viewing slots. The vendor's availability is entered by the negotiator into a Reapit field at the point of listing, and an n8n workflow syncs these availability windows to a Calendly team schedule for that property.

When an enquirer books a viewing via Calendly, the booking triggers an n8n workflow that: sends a confirmation SMS to the enquirer with the viewing details and a one-tap cancel/reschedule link, sends a notification to the vendor via SMS with the viewing time and enquirer's first name, creates the viewing record in Reapit CRM, and sends a Slack notification to the assigned negotiator.

For viewings where the vendor has not set availability windows (approximately 30% of cases in the first month), the Calendly booking link falls back to a simple scheduling request form that routes to the negotiator manually. This hybrid maintained flexibility while automating the majority of booking interactions.

Tools: n8n (orchestration), Calendly (booking interface), Twilio (SMS to enquirer and vendor), Reapit API (CRM record creation), Slack (negotiator notification).

Time saved: viewing booking communications reduced from three to five interactions per viewing (phone calls and emails) to one (the enquirer books online, everyone is notified automatically). Negotiators reclaimed approximately 6 to 8 hours per week previously spent on viewing logistics.

Automation 4 Post-Viewing Follow-Up Sequences

Goal: ensure every viewer receives consistent, timely follow-up regardless of which negotiator handled the viewing.

Architecture: when a viewing is marked as Completed in Reapit CRM (updated by the negotiator immediately after the viewing), a webhook fires to n8n triggering the post-viewing follow-up sequence for that enquirer.

Hour 2 post-viewing: automated SMS from the assigned negotiator's virtual number "Hi [Name], thanks for viewing [address] today. What did you think? Happy to answer any questions." This short, conversational message gets a 40 to 55% reply rate dramatically higher than a formal email.

Day 1 post-viewing: if no reply to the SMS, an automated email from the negotiator's own email address (sent via Gmail API using the negotiator's OAuth credentials) a more detailed follow-up referencing the specific property and asking whether they would like to arrange a second viewing or have any questions.

Day 3 post-viewing: if still no engagement, a follow-up email sharing two alternative properties from the current listings that match the enquirer's stated criteria (pulled from the Reapit record of their search criteria). This is generated by Claude AI, which queries the available properties database for matching listings and writes a brief personalised description of why each alternative might suit their needs.

Day 7 post-viewing: a final follow-up asking whether their search criteria have changed or if there is anything the agency can do to help.

Tools: n8n (orchestration and sequence management), Twilio (SMS), Gmail API with negotiator OAuth (personalised email from negotiator's address), Claude AI (alternative property recommendation copy), Reapit API (viewer profile and property data).

Time saved: 3 to 4 hours per week of manual follow-up across the negotiator team. Viewing-to-offer conversion rate increased from 18% to 22% over the first 90 days attributable to the consistency of follow-up rather than its existence (most negotiators were doing some follow-up, but inconsistently and with varying quality).

Automation 5 Property Information Pack Auto-Send

Goal: eliminate the 2.5 to 5.5 hours per week Priya spent manually assembling and sending property information packs.

Architecture: when a new enquiry is ingested by Automation 2, the n8n routing workflow also triggers the information pack assembly workflow in parallel. This workflow queries the Reapit CRM for the property record, retrieves the associated documents (brochure PDF, floorplan PDF, EPC PDF) from the property's document storage folder in Google Drive, combines them into a single ZIP file using an n8n file handling node, and sends an email to the enquirer from a generic properties@ email address with the ZIP attached.

For properties where not all documents are available in Google Drive, the workflow sends what is available and creates a task in Reapit for Priya to complete the document set ensuring nothing falls through without creating a dependency that blocks the initial send.

Tools: n8n (orchestration), Reapit API (property record and document metadata), Google Drive API (document retrieval), Mailgun (email delivery with attachments).

Time saved: Priya's document assembly task was reduced from 2.5 to 5.5 hours per week to approximately 45 minutes per week (handling the incomplete document cases that fall back to manual). Her role shifted to the higher-value task of maintaining the document library rather than sending it repeatedly.

Automation 6 Post-Offer Document Collection

Goal: automatically send all required post-offer documents to buyers, sellers, and solicitors the moment an offer is agreed in Reapit.

Architecture: when an offer is marked as Accepted in Reapit CRM, a webhook triggers an n8n workflow that orchestrates a sequence of document sends to four parties simultaneously.

To the buyer: ID verification request (with instructions for their chosen ID verification method), proof of funds request, mortgage broker referral letter with the agency's preferred broker contact, and mortgage agreement in principle request if not already received.

To the vendor: offer summary confirmation, next steps guide explaining the conveyancing process, and solicitor instruction letter with the agency's preferred solicitor referral.

To the buyer's solicitor (if known at this stage): memorandum of sale draft.

To the vendor's solicitor: same memorandum of sale draft and title pack request checklist.

All letters are generated from templates stored in Google Docs, with the property details, buyer name, offer price, and solicitor details inserted via n8n's Google Docs template filling. The generated documents are converted to PDF and sent via Mailgun.

Tools: n8n (orchestration), Reapit webhook (offer accepted trigger), Google Docs API (template filling), Mailgun (multi-party document delivery), Google Drive (template storage).

Time saved: Priya's post-offer document task reduced from 20 to 30 minutes per agreed offer to 2 minutes (reviewing the automation confirmation in Slack before approving send for high-value transactions). The consistency of document sends also reduced the number of solicitor follow-ups requesting missing documents, which had been a recurring source of transaction delays.

Automation 7 Automated Weekly and Monthly Reporting

Goal: eliminate Marcus's 2 to 3 hour weekly report compilation and replace it with a real-time dashboard and automatically generated narrative.

Architecture: a weekly n8n workflow runs every Friday at 4pm. It queries Reapit CRM via API for: all viewings logged in the current week (total, by property, by negotiator), all offers made and their status, current pipeline by stage, any properties that have been on market for 45+ days without an accepted offer, and all lettings renewal dates in the next 60 days.

For the lettings renewals report specifically, a separate daily n8n workflow checks Reapit for renewals due in the next 60 days and sends a daily digest to the lettings team every morning.

The weekly data is passed to Claude AI with a prompt that generates a 400-word narrative commentary highlighting notable performance versus the previous week, flagging properties that need attention, and calling out any pipeline risks. The narrative plus the data tables are formatted into a Google Slides presentation using a pre-built template (filled via the Google Slides API) and sent as a PDF to the partners via email and as a Slack message to the management channel.

For vendor reports (a statutory requirement to update vendors on marketing activity after 45 days without an offer), the same weekly workflow automatically generates a personalised PDF for each qualifying property using the same Slides template system and sends it directly to the vendor via email. These had previously been written manually by the listing negotiator a 15 to 20 minute task per vendor. With an average of 8 to 12 properties per month reaching the 45-day mark, that was 2 to 4 hours of manual report writing eliminated.

Tools: n8n (orchestration and scheduling), Reapit API (all data queries), Claude AI (narrative commentary), Google Slides API (report generation), Mailgun (delivery), Slack (management channel posting).

Time saved: Marcus's weekly report task reduced from 2 to 3 hours to 15 minutes (reviewing the automated report before it goes to the partners). The vendor reports were eliminated as a manual task entirely. Total weekly admin time recovered: approximately 2.5 to 3.5 hours.

Results: 12 Weeks Post-Implementation

At the 12-week mark post-implementation, we conducted a formal outcomes review with the Meridian team.

Total weekly hours recovered: 34 hours per week across the team. This represents approximately 80% of the 48 hours of manual admin identified in the initial audit. The remaining 20% approximately 14 hours consists of tasks that require human judgment at a frequency that makes automation technically possible but not yet economically justified for this team's scale.

First response time to enquiries: reduced from 2.5 hours average to 18 minutes average. In the first two weeks after Automation 2 went live, three leads that would previously have received a first response the following morning were contacted within 15 minutes. Two of those three proceeded to viewings; one made an offer that was accepted. The partners attributed approximately £9,200 in commission directly to the improved response time in the first quarter.

Viewing-to-offer conversion rate: increased from 18% to 22% over the first 90 days, attributable to the consistency of Automation 4's post-viewing follow-up sequence.

Document-related delays in transactions: the lettings team reported a 40% reduction in the number of solicitor chasing emails requesting missing documents from buyers. Post-offer document automation (Automation 6) is credited with this improvement.

Negotiator job satisfaction: a qualitative outcome that was not in the original brief but was clearly visible by the 12-week review. Sophia described it directly: "I now spend my mornings on calls and viewings rather than updating portals. That is what I was hired to do. The admin was making me feel like a data entry clerk, not an estate agent."

Annualised value of time recovery: 34 hours per week at an average fully-loaded cost of £28 per hour (blended across all staff grades) equals £952 per week, or £49,504 per year. This is a direct cost saving the same work is being done without additional headcount, and the recovered time is being reinvested in viewings, vendor care, and business development.

Total implementation cost: £2,400 for the PURIST engagement (discovery, build, testing, training, and 30-day post-live support). Monthly ongoing automation infrastructure cost: £85 (n8n hosting, Twilio SMS, and API subscriptions). Payback period: 10 days of recovered time value.

What They Would Do Differently

The Meridian team's honest reflections at the 12-week review produced three observations that are worth sharing for any agency considering a similar implementation.

Start with the data. The single biggest delay in the implementation was discovering in week 2 that the Reapit CRM had significant data quality issues missing property IDs, inconsistent field usage across negotiators, and duplicate contact records that caused the routing logic to assign the same lead to two negotiators simultaneously. We spent three days in week 2 on data remediation that was not in the original scope. A data audit before automation build would have surfaced this earlier and cost less. Marcus now says this emphatically to other principals he speaks to: clean your CRM before you connect it to anything.

Training takes longer than expected. The technology was live in week 7. Full team adoption took until week 10. The gap was not reluctance it was habit. Negotiators who had managed their own enquiry routing via a shared inbox for five years needed time to trust that the Slack notification system was reliable and that they did not need to check the shared inbox as a backup. Sophia described it as "trusting the system to catch everything." This trust developed over three weeks of seeing the system work consistently before the old habit fully faded.

Measure before you build. The Meridian team had estimates of their time costs, but not measurements. They thought the Friday report took "about 90 minutes." It actually took 2 to 3 hours. They thought viewing scheduling took "maybe 5 hours a week." It was closer to 8 to 12. Better baseline measurement would have produced a more accurate ROI projection and made the case to the wider team more concretely.

Lessons for Other Agencies

The patterns at Meridian are not unique to a 12-person South West London agency. We see the same six pain points across estate agencies of all sizes. The specifics of the tools vary (some agencies use Jupix or Alto rather than Reapit, some use different portal feeds), but the operational problems are structurally identical.

The listing update burden is universal. Every agency using multiple portals manually updates the same information in multiple places. The portal data feed APIs exist precisely to solve this they are just not commonly used by agencies without a technical partner to connect them.

The enquiry response speed gap is universal. Every agency has enquiries arriving from multiple channels without a unified routing system. The first agency to respond wins the viewing more often than not. The response time gap between manual routing (hours) and automated routing (minutes) is a meaningful competitive advantage in a market where buyers and tenants are simultaneously enquiring about multiple properties.

For agencies earlier in their automation journey, the highest-ROI starting point is the same as Meridian's: automated enquiry routing with fast Slack notification to negotiators. This single automation reliably recovers the most time, creates the most visible competitive improvement (faster response), and requires the least complex integration to build. Start there, measure the impact, and use the recovered capacity to fund the next automation.

If you want to understand what a similar implementation would look like for your agency, our automation ROI calculator approach can help you build the business case. And if you would like to discuss a scoped engagement, book a free discovery call we will map your specific pain points against proven automation patterns before recommending anything.

The agencies that will dominate their local markets in 2027 are not the ones with the most negotiators. They are the ones whose negotiators spend their time on negotiations rather than administration and whose systems ensure every lead is contacted within minutes, every viewing is followed up consistently, and every transaction moves without manual friction slowing it down.