At VB Transform 2026, Zillow’s engineering chief made a point that should matter to every real estate principal spending money on AI: ROI numbers only hold up if you measure before you build. The firms that skip baseline measurement can’t prove the technology saved anything, because they never documented what the work cost in the first place.

I see this pattern every week. An agency owner tells me they’re deploying AI to handle buyer enquiries or listing follow-up, and when I ask how many hours those tasks consume today, they guess. The PM who wants to automate maintenance triage doesn’t know whether it takes 40 minutes or four hours per property per month. Without that number, you can’t calculate ROI. You can’t tell whether the agent you built is working. You’re flying blind.

Real estate agencies leak between $60,000 and $250,000 a year to manual coordination work that could run on autopilot. Speed-to-lead loss when enquiries come in after hours. Listing follow-up debt when open-home attendees never get the second touch. Property management triage that consumes PM capacity every single day. But if you don’t measure the baseline hours before you deploy AI, you won’t know which leak you actually plugged, and you won’t be able to defend the spend when someone asks whether it was worth it.

This article walks through the specific tasks real estate agencies should measure, the data you need to collect, and what an AI agent doing that work looks like once you have the baseline. The output isn’t a dashboard. It’s a one-page summary that tells you where the hours go, what an agent would replace, and whether the economics make sense for your business.

Why Real Estate Agencies Skip Baseline Measurement

Most agency owners don’t measure baseline hours because the work feels impossible to track. Buyer enquiries come in at random times. Listing follow-up happens in the cracks between inspections. Property management triage is distributed across three PMs who all handle it differently. You can’t instrument every conversation, so you skip the measurement and deploy AI based on intuition.

The problem is that intuition doesn’t hold up when the finance partner asks whether the AI agent paid for itself. You’ll say it saved time, they’ll ask how much time, and you won’t have an answer. The ROI claim collapses because you never documented the before state.

Zillow’s engineering lead made this point explicitly at VB Transform. Firms that measure before they build can prove the value. Firms that don’t measure can’t. The difference isn’t the quality of the AI. It’s the discipline of the baseline.

Real estate agencies have an advantage here because the work is repetitive and the hours are visible if you look. Buyer enquiries follow a pattern. Listing follow-up happens on a predictable cadence. Property management triage consumes a fixed slice of PM capacity every month. You don’t need to track every minute. You need to sample a week, document the hours, and extrapolate across the year.

The Three Baseline Measurements That Matter for Real Estate

If you’re going to measure baseline hours, focus on the three areas where real estate agencies leak the most capacity: speed-to-lead response, listing follow-up, and property management triage. These are the tasks that consume agent and PM time every day, and they’re the tasks where AI agents deliver the clearest ROI.

Speed-to-Lead Response

Buyer enquiries come in at 9pm. The agent replies at 10am the next day. The buyer has already booked another viewing. First-responder agents win two to three times more often than agents who reply the next morning, but most agencies don’t have the capacity to answer enquiries outside business hours.

The baseline measurement here is simple: how many enquiries come in after hours, and how long does it take your team to respond? Track a week. Count the enquiries that arrive between 6pm and 9am. Document the response time for each one. Multiply by 52 weeks, and you’ll know how many leads you’re losing to response delay.

A typical agency with 15 active listings sees 30 to 50 enquiries per week. Half of those come in outside business hours. If your team replies the next morning, you’re losing 15 to 25 leads per week to speed-to-lead lag. That’s 780 to 1,300 leads per year. If your close rate is 2%, that’s 16 to 26 lost sales. At a $10,000 commission per sale, that’s $160,000 to $260,000 in annual leakage.

The AI agent that replaces this work is a Buyer Enquiry Agent running on Omni voice. It answers portal and phone enquiries within seconds, qualifies the buyer with a conversational script, and books the inspection directly into the agent’s diary. The agent doesn’t touch the enquiry unless the buyer is qualified and ready to view. You can see the full breakdown at the AI audit for real estate agencies.

Listing Follow-Up

Open-home attendees leave their details. Portal enquiries submit a form. Warm prospects say they’ll think about it. Most of them never get the second or third touch because the listing agent is already chasing the next lead. Listings die from neglect, not market conditions.

The baseline measurement here is how many warm prospects each listing generates, and how many of them receive follow-up. Track one listing from open home to sale. Count the attendees, the portal enquiries, and the phone calls. Document how many of them get a follow-up email or call within 48 hours. Document how many get a second follow-up within a week.

A typical open home generates 12 to 20 attendees. A typical listing receives 8 to 15 portal enquiries over its first two weeks. If your listing agent follows up with half of them once, and none of them twice, you’re leaving 10 to 18 warm prospects per listing on the table. Multiply by 15 active listings, and you’re neglecting 150 to 270 prospects at any given time.

The AI agent that replaces this work is a Listing Nurture Agent running on Omni ops. It runs a per-listing follow-up cadence to every open-home attendee and portal enquiry until the property sells or they unsubscribe. The agent sends the first follow-up within an hour, the second follow-up three days later, and the third follow-up a week after that. The listing agent only touches the prospect when they reply with interest.

Property Management Triage

Maintenance requests come in by email, text, and phone. Tenant questions arrive at random times. Inspection scheduling consumes PM hours every week. Most PMs cap out at 80 to 120 properties because triage work scales linearly with portfolio size.

The baseline measurement here is how many hours per property per month your PMs spend on triage. Track one PM for a week. Count the maintenance requests, tenant questions, and inspection coordination tasks. Document the time spent on each one. Divide by the number of properties in their portfolio, multiply by four weeks, and you’ll know the monthly triage cost per property.

A typical PM managing 100 properties spends 15 to 25 hours per week on triage work. That’s 60 to 100 hours per month, or 0.6 to 1.0 hours per property per month. At a $60 hourly rate, that’s $36 to $60 per property per month in triage cost. Multiply by 100 properties, and you’re spending $3,600 to $6,000 per month on coordination work that could run on autopilot.

The AI agent that replaces this work is a Property Management Triage Agent running on Omni ops. It handles tenant maintenance requests end-to-end. It triages the request, schedules the trades, and updates the owner without PM intervention. The PM only touches the request if the tenant escalates or the trades can’t resolve it.

If you want a practical starting point for measuring speed-to-lead response, we’ve built a worksheet that walks through the exact questions to ask and the data to collect. You can download the Speed-to-Lead Script for Real Estate Teams and use it to document baseline hours for your team. It’s a single-page checklist that takes 20 minutes to complete.

What an AI Agent Doing This Work Looks Like

Once you have the baseline hours, you can design the AI agent that replaces the work. The agent isn’t a chatbot. It’s a system that handles the entire task from enquiry to resolution without human intervention. Here’s what that looks like for each of the three tasks above.

The Buyer Enquiry Agent answers the phone or portal enquiry within seconds. It asks the buyer what they’re looking for, qualifies their budget and timeline, and offers available inspection slots. If the buyer is qualified and ready to view, the agent books the inspection directly into the listing agent’s calendar and sends a confirmation SMS. If the buyer isn’t ready, the agent captures their details and adds them to a nurture sequence. The listing agent never touches the enquiry unless the buyer is qualified.

The Listing Nurture Agent runs a per-listing follow-up cadence to every warm prospect. It sends the first follow-up email within an hour of the open home, with a link to the property video and a question about their timeline. It sends the second follow-up three days later, with comparable sales data and an offer to answer questions. It sends the third follow-up a week after that, with a reminder that the property is still available and an invitation to book a private viewing. The listing agent only touches the prospect when they reply with interest.

The Property Management Triage Agent handles tenant maintenance requests from submission to resolution. It receives the request by email or SMS, triages the urgency, and schedules the appropriate trades based on the property’s vendor list. It sends the tenant a confirmation with the scheduled time, updates the owner with a summary, and follows up with the trades to confirm completion. The PM only touches the request if the tenant escalates or the trades can’t resolve it.

These agents run on Omni, the AI platform we built for mid-market service businesses. Omni voice handles the phone and enquiry automation. Omni ops handles the follow-up and triage workflows. Omni apps connects the agents to your CRM, calendar, and property management system so they can read and write data without manual intervention. You can explore the full platform at our learning hub.

How to Document Baseline Hours in One Week

You don’t need a month-long study to measure baseline hours. You need a week of disciplined tracking and a spreadsheet that captures the right data points. Here’s the process we use with real estate agencies during the Omni Audit.

Pick one week that represents typical volume. Don’t pick the week of a major auction campaign or the week between Christmas and New Year. Pick a normal week when your team is handling the usual mix of enquiries, listings, and property management work.

Track three data points for each task: volume, time per task, and who handles it. For buyer enquiries, count how many come in, how long each one takes to answer and qualify, and which agent handles it. For listing follow-up, count how many warm prospects each listing generates, how long each follow-up takes, and which agent handles it. For property management triage, count how many maintenance requests come in, how long each one takes to resolve, and which PM handles it.

Use a simple spreadsheet with one row per task. Columns for date, task type, volume, time spent, and owner. At the end of the week, sum the hours by task type and multiply by 52 weeks. That’s your annual baseline.

Most agencies find that speed-to-lead response consumes 8 to 12 hours per week across the team. Listing follow-up consumes 6 to 10 hours per week. Property management triage consumes 15 to 25 hours per week per PM. If you’re running a 10-person agency with two PMs, that’s 35 to 60 hours per week of coordination work that could run on AI. At a blended hourly rate of $50, that’s $91,000 to $156,000 per year in capacity cost.

Once you have the baseline, you can calculate the ROI of an AI agent. If a Buyer Enquiry Agent saves 10 hours per week, that’s $26,000 per year in capacity cost. If it also increases your close rate by converting 10% more after-hours enquiries, that’s another $80,000 to $120,000 in annual commission. The total ROI is $106,000 to $146,000 per year. The agent costs a fraction of that to build and run.

The next step is to book a 60-min Omni Audit where we walk through your baseline data, map the AI agents that replace the work, and calculate the ROI for your specific business. The audit delivers three outputs: a one-page baseline summary, a per-agent ROI model, and a 90-day build plan. No deck, no sales pitch. Just the numbers and the plan.

Why ROI Claims Collapse Without Baseline Measurement

Zillow’s engineering chief made the point at VB Transform because he’s seen it happen. Firms deploy AI without measuring the before state, then can’t prove the technology delivered value when the finance team asks for the ROI. The AI might be working perfectly, but without a baseline, you can’t quantify the savings.

Real estate agencies face this risk every time they deploy AI based on intuition instead of data. You know listing follow-up consumes agent time, but you don’t know how much time. You know property management triage caps PM capacity, but you don’t know the hourly cost. You deploy an AI agent, it handles the work, and you can’t prove whether it saved $20,000 or $200,000 because you never documented the baseline.

The fix is simple: measure before you build. Track a week, document the hours, and calculate the annual cost. Once you have the baseline, you can design the AI agent that replaces the work, calculate the ROI, and prove the value when someone asks whether it was worth it.

Most real estate agencies we work with find that baseline measurement takes less than a week and delivers a 10x return on the time invested. You spend five hours tracking the data, and you unlock $100,000 to $250,000 in annual capacity savings. The measurement isn’t overhead. It’s the foundation of the business case.

What the Omni Audit Delivers for Real Estate Agencies

The Omni Audit is a 60-minute working session where we walk through your baseline data, map the AI agents that replace the work, and calculate the ROI for your specific business. It’s not a discovery call. It’s a structured audit that delivers three outputs you can use the same day.

The first output is a one-page baseline summary that documents the hours your team spends on buyer enquiries, listing follow-up, and property management triage. It shows the volume, time per task, and annual cost for each one. This is the number you’ll reference when someone asks what the AI agent saved.

The second output is a per-agent ROI model that shows the capacity savings, revenue uplift, and payback period for each AI agent. It uses your baseline data to calculate the annual value, then compares it to the cost of building and running the agent. Most agencies see a 4 to 8 month payback for speed-to-lead and listing follow-up agents, and a 2 to 4 month payback for property management triage agents.

The third output is a 90-day build plan that sequences the agents by ROI and maps the integration work required to connect them to your CRM, calendar, and property management system. It’s not a roadmap. It’s a build plan with milestones, dependencies, and go-live dates.

You can see examples of what other real estate agencies have built at the AI audit for real estate agencies. The audit itself costs nothing. It’s a working session where we walk through your data and deliver the three outputs. If the ROI makes sense, we’ll talk about the build. If it doesn’t, you’ll have the baseline data and the model to use internally.

The agencies that get the most value from the audit are the ones that come prepared with baseline data. If you’ve tracked a week of buyer enquiries, listing follow-up, and property management triage, we can calculate the ROI in the first 30 minutes and spend the rest of the session designing the agents. If you haven’t tracked the data yet, we’ll walk through the measurement process and schedule a follow-up session once you have the numbers.

The Real Cost of Skipping Baseline Measurement

The cost of skipping baseline measurement isn’t the time you save by guessing. It’s the ROI you can’t prove when someone asks whether the AI agent was worth it. You deploy the agent, it handles the work, and you can’t quantify the savings because you never documented the before state.

Real estate agencies leak $60,000 to $250,000 a year to manual coordination work. Speed-to-lead loss when enquiries come in after hours. Listing follow-up debt when open-home attendees never get the second touch. Property management triage that consumes PM capacity every day. But if you don’t measure the baseline hours before you deploy AI, you won’t know which leak you actually plugged.

Zillow’s engineering chief made the point at VB Transform because it’s the difference between an ROI claim that holds up and an ROI claim that collapses under scrutiny. Firms that measure before they build can prove the value. Firms that don’t measure can’t.

The fix is simple: track a week, document the hours, and calculate the annual cost. Once you have the baseline, you can design the AI agent that replaces the work, calculate the ROI, and prove the value when someone asks whether it was worth it. The measurement takes less than a week. The ROI lasts for years.

If you want to see what that looks like for your agency, book a 60-min Omni Audit and we’ll walk through your baseline data together. You’ll leave with a one-page summary, a per-agent ROI model, and a 90-day build plan. No deck, no sales pitch. Just the numbers and the plan.