Skip to main content

​
Dimension C: Resident Advocate Intelligence Layer

This is Dimension C applied inside Resident, not a separate product surface: an intelligence layer that reads witnessed records, supports human decisions, and makes the Resident surface more useful.

​
What It Does

Converts the Trust Graph from a passive record into an active advantage for every Resident user. This is the most user-visible AI dimension and the most distinctive in the market. Every Resident user has advocate support that knows their full Trust Graph across all three tracks and acts in their interest — proactively, not reactively. It does not just surface matches. It advocates.

​
What the Agent Knows and Does

For apartment hunting (Track 3):
  • Proactively surfaces matching properties before the user searches
  • Flags compatibility mismatches (“this landlord raises rent aggressively — your budget trajectory suggests a problem”)
  • Flags property risks (“3 unresolved maintenance events in 18 months”)
  • Drafts application letters using verified Trust Graph as foundation
  • Pre-fills document packs from verified identity data
  • Alerts instantly when a previously-viewed property reopens
  • Sets expectations: “your Trust Graph tier makes you a strong candidate for this property but a weak candidate for that one — here is why”
For service requests (Track 2 consumer):
  • Surfaces compatible providers by trust tier, not just proximity
  • Flags provider patterns (“68% repeat hire rate from similar properties”)
  • Drafts job descriptions to reduce scope disputes before they start
For Leja Stay (Track 3):
  • Surfaces compatible hosts (trust tier matched)
  • Flags property risk signals from Property RIN
  • Handles booking initiation and monitors milestones

​
The Transparency Rule

The agent must communicate confidence levels, not just conclusions. If the agent is recommending based on thin data (new landlord with 1 previous tenancy), it must say so: “I have limited data on this landlord. Proceed with extra care.” If the recommendation is based on strong data: “This is based on 18 months of verified data across 4 tenancies. High confidence.” The agent never makes a final decision. It recommends. The user decides.

​
How It Generates Value

The Resident advocate layer is the core of Resident Pro subscription value. Users pay for the tier that gives them full support capability. The layer also drives engagement — users who interact with their advocate support process more transactions through Leja, deepening the witnessed history in their Trust Graph.

​
Dimension D: Professional Operator Intelligence Layer

This is Dimension D applied inside Leja Professional, not a separate product surface: an intelligence layer that reads witnessed operator records, supports human decisions, and makes the Professional surface more useful.

​
What It Does

Makes every agent, landlord, and organisation on Leja Professional more effective — not just better informed, but operationally capable of acting with integrity at scale.

​
For Agents (Solo and Org Staff)

  • Shortlists applicants by trust compatibility (behavioral, not just financial)
  • Drafts state-correct tenancy agreements pre-populated for that jurisdiction
  • Flags when a property’s maintenance history approaches Property RIN tier threshold
  • Manages viewing schedules and sends confirmations automatically
  • Alerts when listing status should change (UNDER_OFFER after deposit received)
  • Drafts client communication in the agent’s voice and style

​
For Landlords

  • Alerts when rent payment approaches late threshold
  • Flags maintenance patterns suggesting systemic issues
  • Generates deposit return documentation with witnessing trail
  • Surfaces compatible applicants from Resident pool without landlord having to review every application manually

​
For Organisations (BRN Level)

  • Monitors firm-level dispute rate, flags emerging patterns
  • Alerts when individual staff Trust Graphs are declining
  • Monitors LASRERA renewals and CAC filing deadlines
  • Assists with staff onboarding (linking new staff RINs to BRN)

​
The Discrimination Guardrail

The intelligence layer must not become a tool for discrimination. It must surface compatible matches at every Trust Graph tier — not just filter out lower tiers entirely. The Trust Graph is for trust assessment, not exclusion.

​
Dimension E: AI for Network-Level Fraud Detection

​
What It Does

Detects patterns that precede fraud at the network level — invisible to individual transaction review but visible when the whole network is watched.

​
What the Network-Level AI Detects

Velocity anomalies: unusual transaction speed for a new actor
Geographic clustering: suspicious concentration in one area rapidly
Cross-party patterns: same people repeatedly in disputed transactions
Document similarity: multiple properties with nearly identical ownership docs
Payment anomalies: deposits received, tenancy commencement not confirmed

​
The Non-Negotiable Rule

AI surfaces patterns to human reviewers. It NEVER auto-flags. The three-tier fraud flag system’s human review gate applies without exception to every network-level detection. A legitimate agent operating at high velocity in a hot market triggers the same velocity signals as a fraudulent one. The AI flags. The human decides.

​
Dimension F: AI as the Federated Computation Engine

​
What It Does

Runs institutional models on Leja’s three-track witnessed data without the raw data ever leaving Leja’s servers.

​
How It Works

1. Bank submits proprietary risk model parameters to Leja endpoint
2. Leja AI runs the bank's model against anonymized three-track data
3. Bank receives only the output score
4. Raw data never leaves Leja's infrastructure
5. User consent: named institution, named purpose, time-limited, revocable
The bank gets the benefit of Leja’s data. They never hold any of it. If the bank is breached, no Leja user data is exposed.

​
Revenue

This is the highest-priced institutional API tier. Per-computation fee reflecting the predictive value delivered. As Leja’s data depth increases, the accuracy improvement over internal models increases — justifying progressively higher fees.

​
Risk

Federated computation requires significant security architecture. Independent security audit required before any institutional computation. This is Milestone Wave 4 for good reason.

​
Dimension G: Service Participation Trust Matcher

​
What It Does

Matches service providers and requesters on trust compatibility at depth — not just availability and proximity.

​
Surface vs. Trust Matching

Surface matching (any platform can do this):
  • Electrician available in Surulere, within 5km, budget within range
Trust matching (only Leja can do this):
  • Trust tier compatibility (Gold-tier resident with Gold-tier provider)
  • Behavioral compatibility (requester’s service history vs. provider’s typical client profile)
  • Property context (provider’s experience with this property type and this specific maintenance category)
  • Payment reliability (requester’s payment timeliness surfaced before acceptance)
  • Repeat hire signal (if these two have worked together and it went well — surface prominently)

​
The Inclusion Rule

Trust tier matching must not exclude lower-tier users. Bronze-tier providers and requesters deserve service participation access. The algorithm must surface appropriate matches at every tier — not optimize only for top-tier transactions.

​
How It Generates Value

Higher match quality → higher completion rates → fewer disputes → higher repeat hire rates → stronger Trust Graph signals → higher platform engagement → higher transaction fee revenue. The compounding is the point.