Cases, tasks, and smart queues
Work is modeled as cases and tasks, routed through smart queues with priority aging — so the oldest, riskiest work surfaces instead of sinking.
Servicing Control Plane · part of the LendEasy platform
Loan servicing software that decides what is allowed to happen next
The control plane is the brain of servicing: cases, tasks, and smart queues for humans and AI agents alike — with a compliance gate in front of every action, contact budgets computed in real time, and an evidence record behind everything that happens.
Servicing mission control
Humans and AI agents · one set of queues
Smart queues · priority aging on
Collections48Hardship12Disputes7Bankruptcy5
SLA health · business-hours calendars
On track
164
Due soon
6
Breached
1
Aged to the top of the queue
Case #C-2417 — Collections
Next action: outbound call · waiting 3 days
The oldest, riskiest work surfaces instead of sinking — and every action still clears the gate
The control plane
The layer that turns servicing policy into machinery
Cases, tasks, smart queues, and omnichannel interactions — governed by a compliance engine that runs before actions, not after. This is where servicing policy stops being a PDF and becomes executable.
Omnichannel interactions
Voice, email, chat, and voicemail are first-class interactions on the case — with consent, timing, and contact budgets enforced per channel.
Regulation-traceable rules
Every rule carries a named regulation, jurisdiction, and effective date, and is version-pinned — "which rule version was in force when this message was sent" is a query, not a project.
Execution-time re-check
Eligibility is re-evaluated at the moment of send or dial, not when the work was queued. Stale permission never reaches a borrower — and missing facts never permit.
Real-time contact budgets
Contact-frequency limits like Reg F's 7-in-7 are computed per debt in real time, with voicemails and limited-content messages counted.
Product-class-aware rules
Consumer rules and merchant-advance rules never silently cross. The engine knows which class an obligation belongs to and applies the right book.
Cases, tasks & queues
Work that surfaces instead of sinking
Everything in servicing is a case with tasks, and every task lives in a smart queue. Priority aging means the work that has waited longest — and risks the most — rises to the top, instead of being buried under whatever arrived this morning.
Smart queues with priority aging
Queues route by skill, capacity, and policy — and re-rank continuously as work ages. SLAs run on business-hours calendars, so a case opened Friday afternoon is not already breached on Monday morning. Humans and AI agents pull from the same queues, under the same priorities, visible in the same command center.
Omnichannel, on the case
Voice, email, chat, and voicemail are first-class interactions attached to the case — not entries in a separate contact tool. Consent, timing windows, and contact budgets are enforced per channel, and every interaction lands in the case timeline, recorded and attributable.
The compliance gate
Compliance runs in the path of work, not behind it
Every action the control plane releases — a message, a dial, a payment — clears the compliance engine first. Four properties make that more than a checkbox.
Checked before execution, not reported after
Every send and every dial passes the compliance engine at the moment it happens — for humans and AI agents alike. There is no path around the gate.
Every rule answers for itself
Rules carry a named regulation, jurisdiction, and effective date, and are version-pinned — so any past action can be traced to the exact rule version in force.
Missing facts never permit
If the engine cannot establish that an action is allowed — because a fact is absent or stale — the action is held, not waved through.
Budgets computed in real time
Contact-frequency limits like Reg F's 7-in-7 are live, per-debt computations your workers can see before they dial — not a nightly reconciliation report.
Events & reconciliation
Desired state, observed state — and work for the gap
The control plane holds a model of what should be true and reconciles it against what the core says is true. When the two diverge, the divergence does not sit in a report — it becomes work.
- Ledger facts arrive with explicit freshness and as-of semantics, so every decision knows how current its inputs were
- Core events stream back continuously, keeping the control plane's view and the system of record converging instead of drifting until month-end
- When desired and observed state diverge — a payment that should exist, a status that should have changed — the divergence becomes a case, with an owner and an SLA
Divergence becomes work
A promise to pay with no matching payment, a protection flag that should have stopped outreach, a status the core never confirmed — each opens a case in a queue with an owner and an SLA. Reconciliation is continuous and operational, not a month-end forensic exercise.
The evidence trail
Every action leaves a record built for the examiner
Each decision the control plane makes — allowed, held, blocked, approved — lands in a hash-chained record linked into the evidence graph: the facts it saw and how fresh they were, the rule versions in force, the approvals, and the outcome.
Tamper-evident by construction
Records are hash-chained — each commits to the hash of its predecessor, so any after-the-fact edit breaks the chain visibly. The graph is queryable, and litigation or exam evidence exports in one click rather than being reconstructed from logs.
One trail for humans and AI
AI agent actions write to the same evidence graph as human work — with the model and prompt versions that produced them pinned to the record. When an examiner asks why something happened, the answer looks the same regardless of who did the work.
Part of the LendEasy platform
The control plane, and what it governs
One of four products on the platform: the workspace your team works in, the AI agents working beside them, and the lending core beneath — all governed by this layer.
FAQ
Questions servicing leaders ask
It is the layer that decides, routes, and records every unit of servicing work. The control plane manages cases, tasks, smart queues, and omnichannel interactions, and runs a compliance engine in front of every action — so whether a human or an AI agent does the work, the same rules apply at execution time and the same evidence is captured. Think of it as the operating layer between your workforce and your system of record.
Put your hardest queue on the control plane
Bring a Reg F-sensitive portfolio or a reconciliation headache. We will trace a real action from queue to compliance gate to evidence record — against our core, or yours.