Live in pilotDarwin Homes · Bullseye Property Management

Maintenance is the largest expenseno one actually manages in real estate.

Every other line item in a property P&L has comps, contracts, and process. Maintenance runs on phone calls and guesses. Baseline makes it a structured, tradeable decision — diagnosis, scope, and price surfaced before the truck rolls.

Active pilots — Darwin · Bullseye
12,845 cases processed
case.baseline-systems.com/c/4A19B3
Live
01Tenant
Capture
Problem intake
Photos, history, occupant context
02AI
Diagnose
Root cause
Confidence-scored, multimodal
03AI
Scope
Living scope
Materials, labor, alternatives
04AI
Price
Benchmark
Regional comps, cost range
05Vendor
Execute
Vendor work
Pre-scoped. No diagnostic lap.
06Trust
Verify
Validation
Photos · occupant confirms
From problem to payment· closed-loop
3–4 touchpoints~20m oversightDays to payment
Operators running tens of thousands of doors
Darwin HomesBullseye Property ManagementEvernest PartnersMynd OpsRoofstock EcosystemPillar PM Co.Homeriver GroupPoplar HomesDarwin HomesBullseye Property ManagementEvernest PartnersMynd OpsRoofstock EcosystemPillar PM Co.Homeriver GroupPoplar Homes
The Problem

The person with the problem can't describe it in terms a vendor can act on.

"The AC isn't working right" could be a $150 filter change or a $2,500 compressor replacement. The gap between "I have a problem" and a digital tradeable unit of work is where every market breaks down.

Property managers drown in coordination.

5–10 hours a week per PM — nearly a third of the week — disappears into vendor back-and-forth. A single repair fires through 10–15 touchpoints over 3–5 days. At 500 units, that's three full-time employees doing nothing but maintenance triage.

30%
of a PM's week lost to coordination
10–15
touchpoints per repair
20%+
of major invoices create disputes

3–5 day repair cycles

Vague tenant reports → vendor hunt → owner approval → schedule — every step a human waits on another human.

Overhead swallows margin

Maintenance coordination is overhead-heavy. The math breaks before you can scale.

Cannot scale profitably

Every hundred new doors demands another coordinator. Margins compress. Best PMs leave.

The Insight

Home repair is a market that runs on information no one has at the moment they need it most.

01
The vendor controls the diagnosis and scope.
02
The occupant can't describe the problem in actionable terms.
03
The owner approves decisions without comparables.

The problem was never that vendors were too expensive or owners too reluctant. The problem was that nobody had the information needed to trust the transaction.

The Solution

Baseline turns an unstructured problem into a tradeable unit of work.

Not a marketplace. Not a logistics middleman. A decision and transaction layer that removes the information gap before anyone gets in a truck.

01

Objective diagnosis

Tenant, owner, vendor, and PM all work off the same Baseline understanding of what is wrong. Confidence-scored. Multimodal. Surfaced before the truck rolls.

  • Photos + history
  • Confidence scoring
  • Repair library
02

Living scope

The scope is a document that updates as the root cause evolves. Materials, labor, alternatives — always current, always sharable, always priced against regional comparables.

  • Structured scope
  • Regional benchmarks
  • Scope alternatives
03

Trust brokered

Quality gates at every transition: problem → solution → price → completion. Payment only triggers on verified completion. Vendor reputation compounds across the network.

  • Quality gates
  • Verified completion
  • Reputation that travels
A real case

10:42 PM, Minneapolis. Robert's tenant texts: the furnace isn't lighting.

It's 9°F outside. A family with two kids is already down to 54° indoors and dropping. Without a diagnosis, Robert is about to authorize work between $4,800 and $9,200 — when the actual problem is a $180 igniter. This is the decision Baseline is built for.

Property
Address
4417 Yarrow St, Unit A · Minneapolis, MN
Details
Duplex · Unit A rented to the Chen family · Unit B owner-occupied
Equipment
Lennox ML180 furnace · installed 2011 · original to property
Managed by
Self-managed by Robert
Tenant report
Tue · 10:42 PM
Furnace is clicking but won't light. Down to 54° in here and dropping fast. Kids are sleeping in coats. Supposed to be 9° tonight — what do we do?
2 photos · thermostat reading · furnace control board with no flame
How this case plays out — toggle:

Robert approves a $312 igniter replacement in 34 minutes. Tenant has heat by 9 AM the next morning.

  1. Tue 10:42 PM
    Tenant submits report + photos through Robert's Baseline portal.
  2. Tue 10:46 PM
    Diagnostic: 94% hot surface igniter failure · 4% flame sensor · 2% gas valve.
  3. Tue 10:51 PM
    Scope + pricing drafted: $285–$340. Benchmarked vs. 38 Twin Cities jobs.
  4. Tue 11:16 PM
    Robert approves from phone — 34 min after the tenant's text.
  5. Wed 8:15 AM
    Vendor arrives with igniter on truck. Scope pre-defined, no re-diagnosis.
  6. Wed 9:05 AM
    Igniter replaced, system tested, tenant confirms heat. Payment released.
The difference
Decision time
34 min
vs. 11+ hrs of freezing
Final cost
$312
vs. $4,800 — and likely $9,200 for the 'safe' option
Tenant habitable
heat by 9 AM
vs. family in coats overnight
See what each agent produced
Product

Five agents. One case. What each one actually produces.

Baseline isn't a human workflow with AI features bolted on. It's an AI workflow with human oversight — each agent specialized, each handoff structured. Pick an agent to see what it produced for Robert's furnace case.

Multimodal · structured outputs
ViewingCase · 4417 Yarrow St · Furnace won't light
Baseline Agent
Diagnostic Agent

Multimodal reasoning over the occupant report, photos, the property's history, and similar cases in the network. Returns a root cause with a confidence score and alternate hypotheses when uncertainty is high.

Baseline · artifact

Diagnostic output

4417 Yarrow St · Furnace won't light

94% confidence
Root cause
Hot surface igniter failure — classic Lennox ML180 fault pattern
Alternate
Flame sensor fouling (4%) · gas valve solenoid (2%)
Severity
Urgent — 9°F overnight low, tenant at 54°F with children
Signals
Clicking without ignition + photo shows no flame at start sequence · 14-yr-old HSI well past 7–10 yr typical lifespan
Next step
Replace HSI · verify flame sensor clean · gas pressure check on-site
Evidence: 2 tenant photos · install record 2011 · 17 comparable Lennox HSI failures in Twin Cities network
85%+
correct diagnosis, vendor-validated
Outcomes

Both sides of the transaction win. Simultaneously.

01 / 04
0×
faster repair resolution
4 days → 1 day for standard cases
†Based on 87 closed Baseline cases vs. PM baseline
02 / 04
0%
of PM week freed
coordination overhead eliminated
†Darwin Homes pilot, Q3 time-study
03 / 04
0%
fewer invoice disputes
scope clarity closes the gap
†Pilot data, Austin + Denver portfolios
04 / 04
0%
pricing within range
actual cost tracks estimate
†Measured on completed cases, n=412
Vendors

Certain jobs. Quick payment. Zero marketing cost.

  • Pre-scoped work — no unpaid diagnostic lap
  • Days-to-payment, not weeks-to-invoice
  • Reputation that compounds across PMs
  • Vendors good at showing up get more volume
Occupants & Owners

Fast resolution. Predictable cost. Trust built in.

  • Approve with full context, not guesswork
  • See what similar repairs cost in your market
  • Quality assurance built into every gate
  • One inflated invoice no longer breaks the PM relationship
The PM calculus

30% of a PM's week is coordination Baseline eliminates.

Hours per month spent on maintenance coordination at a 500-unit portfolio.

Reduction
81%
Before Baseline≈ 3 FTE equivalent · $11.8K/mo staff cost vs. $7.5K revenue
150hrs / month
After BaselinePM oversight only · scope + approval + payment automated
28hrs / month
†Benchmarked against 4 mid-market PM companies, 2025 pilots
The shift

The market isn't missing vendors.

It's missing the information to trust them.

Why Now

The AI to solve this didn't exist three years ago. It does now.

For two decades, property maintenance has tried to become a marketplace. It never could — because nothing upstream was structured. Standardizing the decision was the hard problem. AI just solved it.

Multimodal reasoning

Photos, text, and audio interpreted together with the accuracy a trained technician used to provide.

Structured scope generation

Free-text diagnoses converted into materials lists, labor breakdowns, and vendor-ready work orders.

Confidence-scored diagnostics

Not just an answer — a calibrated answer, with alternates surfaced when uncertainty is high.

Production-viable cost

What was a research-lab demo in 2023 now runs inside a transaction margin.

Roadmap

Property managers today. Every homeowner eventually.

We start with the operators who coordinate the most maintenance — professional property managers. The infrastructure we build for them extends naturally to owner-managed rentals and, eventually, every homeowner facing a repair decision.

Phase 1
Now
Active
PM-managed rental
14M units

Aggregators of doors and vendors. Cold-start solved by design.

Phase 2
Next
Owner-managed rental
23M units

Self-managed landlords inherit the Baseline vendor network.

Phase 3
Later
Owner-occupied residential
82M units

Consumer-facing on infrastructure the industry paid to build.

Team

Operators who have lived this problem.

Institutional real estate, AI systems architecture, vendor operations, product at scale. Every founder hired their own shadow.

01 / 04
Cameron Wu
Co-Founder · CEO
Strategy · product thesis · customer partnerships
02 / 04
Alexander Murmann
Co-Founder · CTO
Engineering · AI agent architecture · infrastructure
03 / 04
Will Reid
Co-Founder · CPO
Product vision · PM workflow · case model
04 / 04
Aaron Noe
Co-Founder · Head of Marketplace
Vendor network · supply-side economics
Pilot → Land → Expand

See Baseline on 10–20 of your doors.

Six to eight weeks. Zero cost. We instrument your highest-volume repair categories and prove a validated end-to-end case before anyone signs anything.

The pilot
Scope
10–20 properties
Duration
6–8 weeks
Cost
No subscription during pilot
Outcome
One validated end-to-end case