This is the story of a Central London letting agent managing 180 properties who was losing landlord instructions to a phone that rang out. At peak season, roughly 35 to 40 calls per day were going unanswered. After deploying Softomate's AI receptionist, the call answer rate moved from 62 per cent to 98 per cent within 30 days. Monthly revenue recovered by £3,200. Monthly cost: £299.
Everything in this case study reflects confirmed client data, shared with permission. The client is presented anonymously at their request.
Contents
The client is an independent letting agent based in Central London, managing 180 properties across a mix of long-let residential tenancies. The business operates with a small team: the principal, one senior negotiator, and one PA who handles all inbound calls, email enquiries, viewing bookings and tenancy administration.
For most of the year, this configuration works. Call volumes are manageable, the PA handles the flow, and the principal focuses on new landlord relationships and portfolio growth. But between May and September - the peak rental season in London - the model breaks down.
During peak months, the London rental market accelerates sharply. Properties come to market faster. Prospective tenants call multiple agents simultaneously. Landlords with voids call repeatedly for updates. Tenancies coming up for renewal require outbound contact. The same team that handles 30 calls a day in February is suddenly fielding 100 calls a day in July.
One PA cannot answer 100 calls a day while also processing tenancy documents, responding to emails, and managing viewing diaries. Something has to give. In this case, what gave was call answer rate. The agent estimated that 35 to 40 calls per day were going unanswered during peak - which, based on what we know about missed call behaviour, means 30 to 34 of those callers never tried again.
The letting agent was acutely aware of the problem. The principal had considered hiring a second PA for the summer months, but the economics did not stack up. A temporary PA for four months costs roughly £9,000 to £11,000 all in, requires onboarding time the team did not have, and creates a knowledge gap at the start and an awkward end-of-season conversation at the finish.
A live answering service had been trialled the previous summer. The results were disappointing. The external agents taking calls had no knowledge of the properties, the landlords, or the agency's processes. Callers asking about specific properties received vague answers and a promise of a callback that sometimes came hours later. Several Rightmove leads - prospective tenants who had enquired about specific properties - converted their enquiry into a viewing at a competitor agency because the callback came too late.
The principal identified three specific problems he needed to solve:
Problem 1 - New landlord enquiries: Landlords calling to discuss instructing the agency for the first time were the highest-value calls in the business. A missed call from a prospective new landlord is a missed instruction worth £400 to £600 in initial fees plus recurring management commission. These calls often come during business hours when the PA is already on another call.
Problem 2 - Rightmove applicant calls: Applicants calling about specific properties need to book viewings within hours of their enquiry. If they cannot get through, they move on. Every missed applicant call during a void period extends the void and costs the landlord rent - which in turn damages the agency's landlord relationship.
Problem 3 - Routine tenant queries: A significant portion of the daily call volume was tenants asking questions that did not require human judgment: confirming payment received, asking about inspection dates, querying a maintenance ticket status. These calls were consuming PA capacity that should have been directed at higher-value work.
The agent contacted Softomate after reading about AI receptionist solutions for letting agents. The initial conversation focused on whether an AI could handle the nuance of letting agent calls - not just take messages, but actually answer questions, book viewings, and handle the range of call types that come into a busy London agency.
The Softomate AI receptionist is not a phone menu. It is a conversational AI that answers calls in the agency's name, understands what the caller needs without them pressing numbers, and takes action: booking viewings, answering property queries, routing urgent calls, and taking structured messages for the team.
For this client, the solution was configured to handle four primary call types that together accounted for approximately 80 per cent of inbound volume:
New landlord enquiries: The AI greets the caller, identifies their intent, gathers the key information a property manager needs before a follow-up conversation (property address, number of bedrooms, current tenant status, reason for enquiry), and books a call with the principal at a time that suits both parties. The principal receives an email with the caller's details and a summary before the booked call.
Rightmove applicant calls: Applicants calling about a specific property are identified by the property address or Rightmove reference number. The AI confirms the property is still available, answers standard questions about the property (rent, deposit, availability date, pets, parking), and books a viewing directly into the negotiator's calendar. Applicants receive an SMS confirmation immediately.
Tenant maintenance calls: Tenants reporting maintenance issues go through a structured triage: property address, nature of the fault, urgency. Emergency issues trigger an immediate alert to the principal. Routine issues are logged with a confirmation message and expected response timeline sent to the tenant.
General tenant queries: Rent payment confirmations, inspection date queries, tenancy renewal status questions, and requests for document copies are handled directly by the AI using information pulled from the agency's property management system. No callback required.
The agent signed the Softomate contract on a Monday morning. The system was live on Wednesday afternoon. Here is what happened in those 48 hours.
Monday afternoon - Knowledge base build: Softomate's onboarding team worked through a structured briefing call with the principal and PA. The call covered: the agency's most common call types and how the team currently handles them, the tone of voice the agency uses with callers, the properties currently on the market, the property management system in use, and escalation rules for urgent or complex calls.
Tuesday - Script development and integration: Softomate built the conversation scripts for each call type and connected the AI to the agency's property management system for live property and tenancy data lookup. The AI was configured to use the agency's name throughout all calls and to match the professional but approachable tone the principal described in the briefing.
Wednesday morning - Testing: The principal and PA made a series of test calls covering every scenario in the briefing: new landlord enquiry, applicant booking a viewing, tenant reporting an emergency, tenant asking about a rent payment. Three small adjustments were made to the scripts based on the test results.
Wednesday afternoon - Live: The agency's existing phone number was redirected through the Softomate system. From that point, every inbound call was answered by the AI. The PA was available to take escalated calls but received no routine inbound calls for the rest of the afternoon.
The 48-hour deployment compares favourably with the live answering service the agent had previously trialled, which required a two-week onboarding period and still delivered inconsistent call handling. It also compares favourably with the cost and time of recruiting and onboarding a temporary member of staff. Explore the complete AI receptionist guide for a full breakdown of the deployment process.
The agent tracked performance for the first 30 days with the AI system live. The headline numbers were clear within the first week, but the full picture only emerged at the end of the month.
Call answer rate: Rose from approximately 62 per cent before Softomate to 98 per cent in the first 30 days. The two per cent of calls not fully resolved by the AI consisted of calls where the AI identified a complex situation - a landlord complaint requiring the principal's direct involvement, a tenant in apparent distress - and escalated immediately. These calls were answered by the PA or principal within minutes of escalation.
PA time on calls: Dropped significantly. In the week before deployment, the PA logged approximately 22 hours of inbound call time. In the first week after deployment, that figure fell to under six hours - all on escalated calls requiring human judgment. She used the freed time on tenancy renewals and email responses that had been accumulating.
Viewing bookings from AI-handled applicant calls: In the first month, the AI booked 31 property viewings from applicant calls without any human involvement. Previously, a proportion of these calls would have been missed or returned hours later, by which point many applicants had booked viewings elsewhere.
Landlord enquiries captured: Four new landlord enquiries were received via the AI in the first month that the principal believes would previously have gone to voicemail and been lost. All four resulted in booked introductory calls. Two of the four became new management instructions within the 30-day period.
Staff stress reduction: This is harder to quantify but was noted explicitly by both the principal and the PA in the 30-day review. The PA described the change as moving from reactive to proactive: instead of spending the day trying to keep up with an incoming call queue, she was working from a clear action list of escalated calls and follow-up tasks.
The £3,200 per month revenue recovery figure is based on a conservative, traceable calculation using the confirmed client data. Here is exactly how it was derived.
Before deployment, the agent estimated 35 to 40 calls per day going unanswered at peak. Not all of those were revenue-generating calls. Maintenance calls, routine tenant queries, and calls from existing landlords checking on known properties do not directly generate instruction revenue.
The agent identified that approximately 10 to 12 of the daily missed calls during peak were from prospective new tenants or new landlord enquiries - callers who had found the agency through Rightmove, Google, or word of mouth and were calling to either book a viewing or discuss a new instruction.
After deploying the AI, the agent tracked which previously-missed call types were now being captured and converted. In the first full month, the AI captured and converted 8 additional enquiries that resulted in new instructions - either new tenancies from applicant calls that were booked for viewings, or new landlord instructions from previously-missed landlord enquiry calls.
The agent's average instruction fee for a new management instruction is £400. Eight additional instructions at £400 each equals £3,200 in recovered monthly revenue.
The monthly cost of the Softomate subscription is £299. Net monthly benefit: £3,200 minus £299 equals £2,901. Return on investment: 970 per cent in the first month.
This calculation does not include:
On the basis of the direct instruction revenue alone, the system paid for itself within the first week of the first month. Use our AI receptionist ROI calculator to estimate the equivalent figure for your agency based on your call volume and instruction fee.
The agent has now been using Softomate for seven months. The system has been refined over that period based on monthly reviews of call transcripts, escalation patterns and conversion data.
Script refinements: In month two, the agent identified that applicants asking about pet-friendly properties were being given a generic response rather than a definitive answer. The knowledge base was updated with pet policy information for each active property, and the AI began giving property-specific answers. Viewing bookings from pet-owner applicants increased.
Seasonal calibration: In October, as the rental market quietened after peak season, the escalation thresholds were adjusted. During peak, urgent landlord calls were escalated immediately. In the quieter winter months, the same calls are handled by the AI with a same-day callback rather than immediate escalation - giving the principal more focused working time.
Outbound use: In month four, the agent began using the outbound calling capability for tenancy renewal outreach. Instead of the PA manually calling 15 to 20 tenants in the renewal window, the AI makes the initial contact call, gauges intention, and books follow-up calls with the principal for tenants who want to discuss terms. The renewal pipeline is now managed more consistently and without the PA time overhead.
Expanding to out-of-hours: From month three, the agent enabled 24/7 coverage after previously running the AI only during business hours. Two emergency maintenance calls during evenings in month four - both involving heating faults - were handled immediately with contractor notifications dispatched. Both tenants received same-night confirmation of action being taken. The principal received a summary notification at 8am the following morning.
The agent's view at seven months is that the AI receptionist has become a permanent part of the business model rather than a seasonal solution to peak call volume. The combination of 24/7 coverage, consistent call quality, and CRM-linked action logging has improved both the client-facing service and the team's internal capacity in ways that compound over time.
If your letting agency is managing similar call volumes and experiencing similar capacity constraints, contact Softomate to discuss whether the same approach would work for your business. There is no obligation and no hard sell - the numbers either stack up for your portfolio or they do not.
Three findings from this deployment are directly transferable to other letting agencies facing similar call volume pressure. First: the majority of the revenue recovery came from prospective tenant calls, not existing tenant calls. The agency had assumed tenant maintenance calls were the dominant inbound category; the actual data showed new tenant enquiries accounted for 41% of total call volume. Any letting agent under-resourcing inbound prospect handling is likely leaving a similar gap open.
Second: the Rightmove and Zoopla speed-to-respond metric was the single biggest behavioural change the AI produced. The agency moved from an average first response time of 47 minutes to under 3 minutes for portal-sourced enquiries. This change alone was responsible for a measurable increase in viewing conversion from portal leads, independent of the missed call recovery. Agents evaluating AI reception purely on call-handling miss this second benefit.
Third: the implementation caused no disruption to existing workflows. The team did not need retraining and the PA role was not eliminated - the PA moved from reactive call-answering to proactive tenant management and landlord relationship development. Productivity per staff member increased because administrative interruptions decreased. The agency is now considering extending AI handling to cover WhatsApp enquiries alongside phone.
The case study data is available for discussion with prospective Softomate clients under NDA. The agency has given consent for aggregate figures to be shared but not named in public materials.
The £3,200 figure is calculated from 8 additional instructions per month captured by the AI that would previously have been missed calls. Each instruction is worth £400 in initial fees. The AI captured and converted those enquiries in the first full month of operation. The calculation is conservative and excludes the value of avoided voids and retained landlord relationships.
The Softomate subscription costs £299 per month. Based on the confirmed revenue recovery of £3,200 in the first month, the system recovered its cost within the first week of operation. The net benefit in month one was £2,901. Even if results in your agency are half this figure, the system still delivers a 435 per cent return on investment in the first month.
The AI works alongside the PA, not instead of her. It handles inbound calls autonomously - routine queries, viewing bookings, maintenance reports. The PA focuses on escalated calls, tenancy administration and landlord relationship management. Her call-handling time dropped by around 16 hours per week. She is doing higher-value work, not fewer hours. No one lost their job.
The feedback gathered by the principal in the first month was positive. Callers appreciated getting a fast answer and immediate action rather than voicemail. Most callers did not ask whether they were speaking to an AI. A small number did ask, and the AI confirmed it clearly. None of those callers complained - they had their query resolved and received confirmation by SMS.
This result is at the stronger end of what we see, driven by the high instruction fee value and peak-season call volume. That said, any letting agent missing 15 or more calls per day during peak season will see material revenue recovery. The exact figure depends on your instruction fee, conversion rate and call volume. Use the ROI calculator to estimate your own figure.
We protect the real names of all clients featured in examples and case studies. Every testimonial is from a real client.
More from the blog
Work with us
Every project we take on has a measurable outcome. Talk to our London team and we will show you exactly how we would approach your challenge.
Softomate Solutions Limited
Deen Dayal Yadav
Online
