- Prepared for
- The owner of a property management firm with 15 people and about 400 rental homes
- Prepared by
- Guy Pavlov, Rel-AI-able
- Date
- October 2026
1.At a glance: three changes cut 105 hours of monthly work to 42
Your team can make two of the changes alone. The third needs a small AI intake, so I recommend a short pilot first.
- 63 hoursa month freed across three workflows
- US$2,500a month in staff time, at about US$40 an hour
- +US$36a month in software, after cancelling a tool your team barely uses
- 1 of 3changes needs new software
The 63 hours are staff time, so your payroll stays the same. They pay off when your team spends them on work that's waiting, or when you take on more rental homes without hiring.
- Now
- 105 h: emails 35, statements 20, maintenance 50
- After
- 42 h: emails 13, statements 6, maintenance 23
| Workflow | Recommendation | Hours a month, now → after | First step |
|---|---|---|---|
| Tenant and landlord emails | Use a tool you already pay for | 35 → 13 | Write 12 reply templates |
| Monthly landlord statements | Ordinary automation, no AI | 20 → 6 | Set up the statement template |
| Maintenance requests | Build a small AI intake | 50 → 23 | A pilot of 2 to 4 weeks |
Your decisions
- Should your team go ahead with the reply templates and the statement template?
- Should we run the maintenance pilot?
- When should we review the results?
2.Tenant and landlord emails: reply templates free 22 hours a month
Use a tool you already pay for
Your team can set this up with the AI assistant you already pay for.
- Now
- 35 h
- After
- 13 h
Today
- 300 emails a week reach your shared inbox
- 120 of them ask one of the same 12 questions, such as when rent is due or how to report a repair
- 4 minutes each, mostly spent finding the last answer and rewriting it
- 35 hours a month in total
What I recommend
- Your receptionist and one property manager list the 12 questions.
- Your AI assistant drafts an answer to each one from a handful of past emails.
- The property manager checks each answer against your leases and office rules, then saves it as a reply template in your team's email.
- Your receptionist picks the right template, adds the tenant's details, answers anything else they asked, and sends it.
Why a person still sends every reply: tenants often add a second question or a complaint to a simple one. An automatic reply would miss it.
Business case
| Line | Each month |
|---|---|
| Hours a month | 35 now, 13 after |
| Hours recovered | 22 |
| Costs removed | None |
| New revenue | None |
| Software added | US$10, for moving two personal AI assistant subscriptions to a business plan (section 5) |
| Checking and upkeep | Included in the 13 hours: checking each reply, and about an hour each quarter to update the templates |
3.Monthly landlord statements: your software's batch feature frees 14 hours a month
Ordinary automation, no AI
Your bookkeeper can set this up with no AI and no new software.
- Now
- 20 h
- After
- 6 h
Today
- Your bookkeeper builds 60 landlord statements a month by hand
- 20 minutes each: export to a spreadsheet, check the figures, write a note, save a PDF and email it
- 20 hours a month, all in the first week
What I recommend
- Your bookkeeper sets up the statement template in your property management software, following its help guide.
- The software builds every statement and emails them in one batch. Your bookkeeper sends you a test batch first.
- Each month, your bookkeeper checks the batch and writes notes only for landlords with something unusual, usually about 12.
Why no AI: your bookkeeper would still have to check every figure an AI wrote, and 12 short notes are quicker to write by hand.
Business case
| Line | Each month |
|---|---|
| Hours a month | 20 now, 6 after |
| Hours recovered | 14 |
| Costs removed | None |
| New revenue | None |
| Software added | None. Batch statements are part of your current subscription |
| Checking and upkeep | Included in the 6 hours: checking each batch, and updating the template when the layout changes |
| Other gain | Your bookkeeper gets most of the first week back for reconciling accounts |
4.Maintenance requests: a small AI intake frees 27 hours a month
Build a small AI intake
This is the one change that needs new software, so I recommend testing it in a pilot first (section 7).
- Now
- 50 h
- After
- 23 h
Today
- 250 repair requests a month, by email, through your website's form or by phone
- 12 minutes each for one of your two coordinators: read the request, pick the trade and urgency, check the landlord's approval limit, create the work order, book a contractor and reply to the tenant
- 50 hours a month in total
- 1 in 10 requests goes to the wrong trade first time, which means a second visit
What I recommend
- Requests from your inbox and your website's form go to an AI step.
- The AI step reads each request and any photos, suggests the trade and urgency, fills in the work order in your property management software, and drafts the reply to the tenant.
- A coordinator checks every work order and reply before anything is booked or sent.
- Emergencies skip the AI step. A fixed rule sends any request that mentions gas, flooding or no heat straight to the on-call phone.
- Phone requests stay as they are for now.
Why build instead of buying a maintenance app: the apps that sort requests this way would move your whole maintenance process to a new system. This intake feeds the software your coordinators already use, so nothing else changes for them.
Business case
| Line | Each month |
|---|---|
| Hours a month | 50 now, 23 after |
| Hours recovered | 27 |
| Costs removed | About 12 fewer second visits, if wrong trades fall from 1 in 10 to 1 in 20 |
| New revenue | Room for about 100 more rental homes before your coordinators need help |
| Software added | About US$80, for the service that runs the intake and the AI it uses |
| Checking and upkeep | Included in the 23 hours: about 5 minutes a request, plus about 2 hours a month to adjust the AI step |
| One-time cost | The pilot, at a fixed price (section 7) |
5.AI tools and subscriptions: upgrade one, cancel one
Your software costs rise by about US$36 a month in all, including the maintenance intake.
| Subscription | Who uses it | Recommendation | Change each month |
|---|---|---|---|
| Property management software | Everyone | Keep. It already builds the landlord statements | None |
| Email and office software | All 15 people | Keep. It holds the reply templates | None |
| AI assistant, 2 personal subscriptions at US$20 | 2 people, who pay and claim it back | Move both to the same assistant's business plan, at US$25 a person, so their work sits in your firm's account | +US$10 |
| AI note-taking app, 3 seats at US$18 | 3 people, who used it twice in the last six months | Cancel | −US$54 |
| Maintenance intake (section 4) | Your two coordinators | Add, after the pilot | +US$80 |
Before you switch, check the business plan's minimum number of seats. Some plans start above two.
6.Data rules: keep tenant details out of personal AI accounts
Share this page with your team before anyone uses these tools with tenant information.
| Information | AI assistant, business plan | The maintenance intake | Personal AI accounts |
|---|---|---|---|
| Office hours, rent dates, move-out steps and other public information | Yes | Yes | Yes |
| Tenant names, addresses, phone numbers and emails | Yes | Yes | No |
| Repair descriptions and the photos tenants send | Yes | Yes | No |
| Bank details, ID documents and credit checks | No | No | No |
| Landlord statements and payments | No | No | No |
- Training: neither tool uses your information to train AI. I checked the terms of your AI assistant's business plan, and I'll set up the intake's AI service so it doesn't keep requests after reading them.
- Who keeps the page: your office manager, who goes through it with new staff on their first day.
- When to review it: every six months, or whenever you add a tool.
- When in doubt: leave the information out and ask the office manager.
7.Pilot proposal: test the maintenance intake for 2 to 4 weeks
The pilot tests the intake on real requests before you commit to it.
- What it coversrequests that arrive by email or through your website's form. Phone requests stay as they are.
- Before setupfor two weeks, your two coordinators note how long each request takes and which ones go to the wrong trade. That's the baseline.
- Success measureunder 6 minutes a request, wrong trades down to 1 in 20 or fewer, and every emergency reaching the on-call phone without passing through the AI step. We agree it in writing before we start.
- SetupI connect your inbox and website form to the AI step and your property management software. I test it on 50 past requests before any real one goes through.
- Trainingan hour with both coordinators, working on their own requests, and a one-page guide on what to check and who to call when something goes wrong.
- Length2 to 4 weeks of use after setup.
- Pricefixed, and agreed with you before the pilot starts.
- At the endwe measure against the baseline, and you decide whether to add phone requests, change the setup, or stop.
8.Next steps: settle the three decisions on a 60-minute call
We'll go through this report together on a 60-minute call and settle the three decisions in section 1. Please bring your office manager and both coordinators.
