Nestack Agent Care
Industries / Accounting / Revenue-recognition agent

Accounting AI agent · Revenue recognition

Revenue-Recognition AI Agent (ASC 606 / IFRS 15)

Read the contract, propose the performance obligations and the allocation, build the deferred-revenue schedule and assemble the disclosure inputs — for the technical-accounting lead to accept, with the trail an auditor will test.

4–6 weeksTypical delivery
Your stackDeployment
Pre-bookingLead approval
Agent CareAfter launch

What this agent does

Reads the contract, not the conclusion

In
01

A contract is signed, amended or renewed, and the agent ingests it from the configured CLM, CRM or repository feed.

02

Clauses are extracted and normalised, and each one stays tied to the page and section it came from.

Reason
03

Distinct promises are proposed as performance obligations, each anchored to the clause that supports it.

04

Transaction price is allocated across the proposed obligations using the configured estimation method.

05

Variable consideration is estimated and the configured constraint factors are applied to the proposal.

Decide
06

Modifications, renewal options and principal-versus-agent indicators are flagged, each with the clause behind it.

07

Every proposed obligation, allocation and constraint is routed to the technical-accounting lead for acceptance.

Out
08

The clause, the proposal, the reviewer's edits and the acceptance are retained against the contract.

09

Execute write actions only inside the approval boundaries agreed during implementation.

Product statement

The agent proposes the obligations, the allocation and the constraint; the technical-accounting lead decides what is booked, and the company remains the preparer of record.

Example workflow

One contract, clause to schedule

AgentHuman
1Contract receivedCLM export, CRM entry, order form or contract-repository API
2Clauses extractedDeliverables, pricing, standalone-price data and prior modifications, each with its clause
3Obligations proposedProposed obligations, allocation, constraint and confidence
4Controls appliedClause-support check, modification-treatment check, dual-framework flag and confidence threshold
No human action required

Stages 1 to 4 run unaided, and nothing is booked at any of them — the technical-accounting lead's lane opens at the confidence gate.

5DecisionBranches at the confidence threshold
High confidence

Goes to the technical-accounting lead to approve.

Low confidence

Adds a controller read first.

Lead approval

The proposal is held with its clauses, its flags and the confidence.

Approve · Adjust · Send to controller review
Approved — released to book
6Revenue systems updatedOnly where write access and approval policy allow it
7Outcome evaluatedAllocation accuracy, override rate, audit findings and post-close corrections
Overrides

Every lead override is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required8 items
Booking the deferred-revenue entry to the ledger.
Concluding that a promise is a distinct performance obligation.
Setting a standalone selling price with no observable input.
Deciding principal versus agent for a promise.
Automation boundaryAgent acts unaided
Propose performance obligations and the allocation, each attached to its clause.
Estimate variable consideration and propose the constrained amount for review.
Build the deferred-revenue schedule from the accepted.
Flag modifications, material rights.
Any write happens inside the boundaries agreed at implementation, never ahead of approval.
Releasing a variable-consideration constraint.
Classifying a contract modification's treatment.
Determining functional versus symbolic IP.
Signing the revenue policy memo.

Example output

One obligation, annotated

Everything the agent proposes is attached to the clause it was drawn from.

Allocation output · single obligationIllustrative example
Contract
Proposed obligation
Allocated price
SSP method
Confidence
Basis
Three-year platform contract
Implementation services distinct from the hosted platform
$186,400
Adjusted market assessment
89%
Comparable market-rate SaaS deals
As receivedTaken from the contract and the order form — nothing on this side is inferred.
Evidence used Statement of Work, clause 4.2 Standalone price list Prior modification history
Why this allocationThe clause supports a distinct promise, but the price isn't observable.
ActionApproveAdjustSend to controller review
What the score decidesBelow the configured threshold the proposal picks up a controller read before the lead sees it.

Value

Where AI adds value

The same four claims, placed at the point in the workflow where each one applies.

Where the value landsValue 01 – 04
Every contractFrom the CLM or contract repository
03Extraction

Read against the clause library

Draw on deliverables, pricing and prior modifications already on file.

01Approved path

Judgement is not extraction

Routine clause extraction and drafting run without a person touching them.

02Human review

Send review to the judgement calls

Modifications, material rights and constraint estimates are flagged, so the lead's read starts where judgement concentrates.

04Build an evidence trail

The obligation, the clause it was read from and the controller who accepted it stay on the contract.

Integrations

Typical integrations

Five system groups connect to the same agent. Which of them are in scope is decided in discovery.

Contract and CLM systemsDocuSign CLM · Ironclad
Salesforce CPQ · Conga
Billing and ERP systemsNetSuite · SAP · Oracle
Zuora · Chargebee · Stripe Billing
CRM and order dataSalesforce · HubSpot
Order forms · amendments

Agent

Revenue recognition

Reads the contract
Proposes the allocation
Holds for approval

Disclosure and reportingWorkiva · Trintech
Approval systems
Observability & evaluationOpenTelemetry · Langfuse
Supported monitoring/evaluation sources

Integration availability depends on the client's existing systems and API access.

Agent controls

Six layers between the model and the schedule

Each layer contains the next. What escapes all of them is named in the map below.

L6 · Outermost — last line of defenceInward → L1 · closest to the model
L6Rollback / safe modeRestrict the agent to proposal-only when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt, clause-library and estimation-method changes.Track
L4TraceabilityRecord the clause, the proposed obligation, the flags and every override.Record
L3Lead approvalHold proposals for the named technical-accounting lead; it governs release, not whether the allocation is right.Gate
L2Policy guardrailsTest proposals against clause-support, modification-treatment and dual-framework rules; a failure holds the item.Restrict
L1Confidence thresholdsRoute low-confidence proposals to a controller read before the lead sees them.Require review
Model coreObligations proposed — the allocation, constraint, flags and confidence
L1 – L2Test whether a proposal may stand
L3Puts the release in the lead's hands
L4 – L5Keep the obligation and the clause behind it
L6Falls back to contract summarisation when signals degrade

How Nestack evaluates it

Evaluate the allocation workflow — not only the finished schedule.

Coverage runs the whole depth of the workflow, and every layer is cut by slice.

Surface — the schedule the auditor sees
Depth of coverage ▼
E1Final-output evaluationDid every proposed obligation tie to a supporting clause?
E2Step-level evaluationDid the agent use the right clause library, estimation method and framework?
E3Tool evaluationDid it read and write the correct contract and obligation?
E4Confidence calibrationDo low-confidence proposals actually attract more overrides?
E5Slice evaluationHow does performance change across specific contract families?
E6Business outcomeHow many proposals needed an override or a later restatement?
Floor — the outcome the auditor tests

Failure modes

Where each failure originates in the agent

Seven failure modes, plotted against the five stages the agent moves through.

Agent lifecycleDirection of processing →
01 · Retrieval1 mode
RU-03

Superseded contract version

Clause terms are read from a version an amendment already replaced.

Stage gathersContract terms, pricing history and prior allocations
02 · Reasoning2 modes
RU-04

Modification treatment reversed

Prospective treatment is applied to a change that needs a cumulative catch-up.

RU-06

Residual method defaulted

Every non-observable price falls to residual, not just the highly variable ones.

Stage proposesThe proposal and its confidence
03 · Tool / write2 modes
RU-02

Commission period undersized

Sales-commission amortisation stops at the initial term though renewals are anticipated.

RU-05

Constraint left stale

A variable-consideration constraint carries forward without a period refresh.

Stage writesOnly where write access and approval policy allow it
04 · Output1 mode
RU-01

Gross presentation error

Revenue books gross though the agent never held control before transfer.

Stage returnsThe schedule the lead approves and the auditor sees
05 · Change / Version1 mode
RU-07

Silent estimation-method drift

A model or rule change widens which allocations default to residual.

Stage tracksModel, prompt, clause library and method rules
Sev-1 · books outside the boundary Sev-2 · wrong allocation reaches the schedule Sev-3 · source degrades, proposal goes to review

Affected slices

A healthy portfolio can hide one contract family

Portfolio-level accuracy can look strong while a handful of contract families absorb most of the rework. Nestack reports the allocation-rework rate by slice, not only in total.

Slice performance — reported separately, not only in aggregateIllustrative example
SliceFailure rateLift Lift vs. thresholdStatus
Contracts modified mid-term7.2%3.8× Review
Bundles with no observable price5.4%2.8× Review
Usage and variable consideration3.3%1.7× Watch
Single-obligation ratable contracts1.9%0.7× Normal
Bar: allocation-rework rate lift vs. the single-obligation baseline · scale 0–4.0× · tick marks the 2.0× review threshold 2 of 4 slices over threshold

Evidence-linked improvement

Every cycle ends inside the regression suite

A cycle is closed when the misallocation is a case the next release must survive. That suite is what the next contract read is measured against.

Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn
01Detect

Allocation-rework rate rises in a contract family.

02Diagnose

The renewal option nobody priced is traced back through the clause and the model that misread it.

03Improve

The change ships against a version, with the contracts that exposed it attached.

04Verify

Release is blocked until the affected allocation cases pass again.

05Learn

The case joins the standing suite and the policy memo moves with it.

Learn → DetectThe return edge. The next contract is read against a suite one case larger.

Typical build scope

Twelve workstreams across six weeks

The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, allocation workflow, evaluation, integration, then production validation and handover.

Workstream Week 1Week 2Week 3Week 4Week 5Week 6
01Contract workflow discovery and boundary definition.
02CLM, CRM and billing-system assessment.
03Clause-library and estimation mapping and rule mapping.
04Contract ingestion and normalisation.
05Obligation and allocation logic.
06Confidence scoring and exception routing.
07Technical-accounting approval workflow.
08Billing and ledger integration.
09Allocation and constraint cases.
10Guardrails and acceptance controls.
11Contract-trail instrumentation.
12Deployment, documentation and Agent Care handover.
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel Final scope and sequence confirmed in discovery

Engagement tiers

What each tier includes

Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.

Capability✓ in scope · — not at this tier PilotOne contract source, one entity ProductionProduction revenue systems AdvancedMultiple entities / standards
Introduced at Pilot
Allocation and schedule recommendations
Lead approval
Allocation-accuracy baseline
Introduced at Production
Reporting by contract type
Approval workflow in your systems
Approved schedule posting
Billing and ledger integration
Introduced at Advanced
Multi-standard allocation rules
Multi-stage technical review
High contract volume
Multi-standard reporting controls
Build price From $5,000 From $8,000 Custom quote
Final build priceConfirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
Separate from buildBuild pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.

What we need from you

What you bring, and what we build with it

Each input maps to a piece of build scope and a week in the delivery timeline.

You bringWe build with it
01Your contract templates and clause library Contract ingestion and clause mappingWeek 1
02Representative signed contracts Allocation baseline, obligation drafting and confidence scoringWeek 2
03Your revenue-recognition policy and estimation methods Clause-library and estimation-method mappingWeek 1
04Access to relevant APIs, feeds or exports CLM, CRM and billing-system assessment, then integration setupWeek 2
05Schedules you would not want booked Modification cases and failure-mode testingWeek 4
06What no schedule may assume Confidence scoring, flag routing, guardrails and acceptance controlsWeek 3
07Named technical-accounting leads to review proposals Technical-accounting approval workflow, then pilot and production validationWeeks 5–6
Nothing else is required Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Each phase sits over the weeks it really occupies, and week 5 carries evaluation alongside the pilot.

Phase W1W2W3W4W5W6
Discovery W1
Build W2 – W3
Evaluate W4 – W5
Pilot & Launch W5 – W6
Week focus W1Revenue workflow discovery, clause mapping and the boundary W2Source integration and the allocation baseline W3Allocation workflow, confidence logic and approval controls W4Evaluation suite, guardrails and failure-mode testing W5Billing integration, pilot contracts and targeted corrections W6One reporting cycle booked under the technical-accounting lead, then handover
Reading the bandA bar covers the weeks its work is named in, and nothing else. The week 5 overlap is real, not padding.
At the end of W6The last checks clear on live contracts and Agent Care picks up monitoring.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Accounting AI agent

Build a revenue-recognition agent around your technical-accounting team's approval chain.

Show us your contract templates, your policy memos and who signs off on judgement calls. Not conclusions. Proposals, with the clause behind each one, for your technical-accounting lead to accept.

Nestack Agents · Revenue recognitionAGT-ACC-18 · Agent Care available after launch