Real Estate AI agent · Energy filings

Building Energy & Emissions Filing AI Agent

Place each building on its regime and pathway, reconcile the consumption record and name what is missing from it, and hold the filing for the professional who certifies and the owner who files.

Typical delivery
4–6 weeks
Deployment
Your stack
Certifier signs
Pre-submission
After launch
Agent Care

What this agent does

Runs the filing year, not the repair ticket

In

  1. 01

    A meter year closes, and consumption is drawn together from utility accounts, fuel deliveries and submeters.

  2. 02

    A building is placed on an emissions and consumption reporting regime, not on a physical-condition inspection cycle.

Reason

  1. 03

    A pathway is set from housing evidence — in New York City, Article 320 or Article 321 — and the evidence is kept.

  2. 04

    A floor area is reconciled to the assessor record and the prior filing, because the emissions limit derives from it.

  3. 05

    A gap stays a gap: which meter, which tenant space, which delivery ticket did not report is written down.

Decide

  1. 06

    A deadline runs on each jurisdiction's own basis, and a New York City non-filer owes §28-320.6.2 under its limit.

  2. 07

    A pack goes to the professional that jurisdiction requires to certify it, then to the owner, who files it.

Out

  1. 08

    A source is kept with the figure it produced, because the audit that tests it arrives long after filing.

  2. 09

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

Product statement

The agent assembles the record and runs the filing clocks; a registered design professional certifies the New York City report, and the owner files it.

Example workflow

One building, meter year to filing pack

  1. 1Reporting year opensUtility accounts, fuel deliveries, submeter logs and the covered-buildings list
  2. 2Regime and pathway setRent-regulation, HDFC and federal-assistance evidence, floor area and space-use mix
  3. 3Record assembledConsumption totals, named gaps, computed exposure and confidence
  4. 4Controls appliedPathway checks, area reconciliation, certifier checks and confidence threshold

No human action required

Stages 1 to 4 run unaided, and nothing is certified or filed at any of them — the agent is assembling, and the portfolio director's lane opens at the confidence gate.

5DecisionBranches at the confidence threshold
High confidence

Goes to the portfolio director to approve.

Low confidence

Adds a design-professional read first.

Director approval

The record is held with its sources, its named gaps and the confidence.

Approve · Correct · Send to certifier review

Approved — released for certification

  1. 6Energy systems updatedOnly where write access and approval policy allow it
  2. 7Outcome evaluatedGap closure, corrections, on-time filings and post-filing audit outcomes

Corrections

Every director correction is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required

8 items

  • Certifying the annual report to the department.
    • Filing or submitting a report to a city or state.
    • Concluding that a building meets its emissions limit.
    • Estimating a missing meter into a reported total.
  • Automation boundary

    Agent acts unaided

    • ✓Place each building on its regime and pathway, with evidence.
    • ✓Reconcile floor area and space use against the source records held.
    • ✓Show which meters, spaces and deliveries did not report.
    • ✓Run the filing clock per jurisdiction and hold the pack.

    Any write happens inside the boundaries agreed at implementation, never ahead of certification.

    • Deciding to pay a penalty rather than retrofit.
    • Requesting or accepting a deadline extension.
    • Signing an attestation for a prescribed measure.
  • Changes to regime, pathway or filing rules.

Example output

One reported input, annotated

Everything the agent assembles is attached to the source it came from.

Filing-record output · single building

Illustrative example
Building
New York City tower
Named gap
Tenant submeter with no fourth-quarter read on file
Next report due
1 May 2027
Rule of record
§28-320.3.7
Confidence
88%
Certifier required
NYS-licensed PE or RA

As received

Taken from the utility account and the submeter log — nothing on this side is estimated by the agent.

Sources used

  • Utility account history
  • Fuel delivery tickets
  • Submeter read log

Why it is named

The certifier signs off on this total, and the owner is the one who files.

What the score decides

Below the configured threshold the record picks up a design-professional read before.

Action

ApproveCorrectSend to certifier review

Value

Where AI adds value

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

Where the value lands

Value 01 – 04

Every building

From the meter and fuel record

03Assembly

Work from the statute

Draw on the regime that binds that building — the filing clock, not the repair clock.

01Approved path

Filed is not compliant

Routine regimes, deadlines and packs arrive already assembled.

02Human review

Send review to the contested pathways

Unclear pathways and unnamed gaps are marked, so the certifier read starts where the exposure concentrates.

04

Build an evidence trail

The report, the pathway it sits on and the professional who certified stay on the filing.

Integrations

Typical integrations

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

  • Utility and meter data

    • Utility APIs
    • Green Button
    • Submeter and BMS feeds
  • Benchmarking

    • ENERGY STAR Portfolio Manager
    • Web Services exchange
  • Property and asset data

    • Yardi
    • MRI
    • RealPage
    • asset registers

Agent

Energy-filing assembly

  • Reads the meters
  • Sets the pathway
  • Holds for certification
  • Filing portals

    • DOB NOW
    • BERDO portal
    • State benchmarking portals
  • Observability & evaluation

    • OpenTelemetry
    • 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 report

The six tighten as you go inward. Whatever gets through them all is listed in the map below.

L6 · Outermost — last line of defence

Inward → L1 · closest to the model

L6Rollback / safe modePull assembly back to record-only when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt, regime-rule and jurisdiction-configuration changes.Track
L4TraceabilityRecord the sources, the named gaps, the pathway evidence, the corrections and the approval time.Record
L3Director approvalHold the record for the named portfolio director; it governs release, not whether the total is right.Gate
L2Policy guardrailsTest each figure against the rule set configured for that jurisdiction; a failure returns the record.Restrict
L1Confidence thresholdsRoute low-confidence records to a design-professional read before the director sees them.Require review
Model coreRecord produced — regime, pathway, totals, named gaps, exposure and confidence
L1 – L2
Test whether a figure may stand
L3
Puts the release in the director's hands
L4 – L5
Keep the report and the pathway behind it
L6
Holds the annual report unsent when signals degrade

How Nestack evaluates it

Evaluate the assembled record — not only the total at the top.

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

Surface — the report the department receives

  1. E1Final-output evaluationDid each reported figure match the source it was drawn from?
  2. E2Step-level evaluationDid the agent use the right regime, pathway and jurisdiction rules?
  3. E3Tool evaluationDid it read the correct building and the correct meter year?
  4. E4Confidence calibrationDo low-confidence records actually attract more corrections?
  5. E5Slice evaluationHow does performance change across specific regimes and building types?
  6. E6Business outcomeHow many records were corrected, re-filed or queried after the report went in?

Floor — the penalty the owner carries

Failure modes

Where each failure originates in the agent

Seven failure modes, placed where the annual record first breaks.

Agent lifecycle

Direction of processing →

  1. 01 · Retrieval

    1 mode

    • FM-03

      Submeter gap unrecorded

      A tenant meter fails to report and nobody writes it down.

    Stage gathers

    Meter reads, fuel tickets, area and pathway evidence

  2. 02 · Reasoning

    2 modes

    • FM-04

      Pathway read off the meters

      Article 320 is assumed from size while the housing evidence says Article 321.

    • FM-06

      Limit built on stale area

      Floor area is carried from last year rather than reconciled.

    Stage proposes

    Regime, pathway, totals, gaps and confidence

  3. 03 · Tool / write

    2 modes

    • FM-02

      Gap filled, not flagged

      A missing read is estimated into the total without being named.

    • FM-05

      Cure window mis-routed

      A Notice of Deficiency is read as a bill, not a deadline.

    Stage writes

    Only where write access and approval policy allow it

  4. 04 · Output

    1 mode

    • FM-01

      Non-filing exposure missed

      A building under its limit is left unfiled and a penalty accrues anyway.

    Stage returns

    The report the owner files and the certifier signs

  5. 05 · Change / Version

    1 mode

    • FM-07

      Silent calendar drift

      A rule change moves a deadline and the calendar keeps the old one.

    Stage tracks

    Model, prompt, regime rules and jurisdiction config

  • Sev-1 · a submission leaves the boundary
  • Sev-2 · a wrong figure reaches the certifier
  • Sev-3 · source degrades, record to review
See the complete Real Estate failure-mode catalogue

Affected slices

A portfolio figure can hide one regime

A portfolio-level filing-readiness figure can look sound while a few regimes carry nearly all of the incomplete records. Nestack reports the incomplete-record rate by slice, not only in total.

Slice performance — reported separately, not only in aggregate

Illustrative example
SliceFailure rateLiftLift vs. thresholdStatus
Article 321 rent-regulated stock7.1%3.6×Review
Tenant-submetered office5.2%2.7×Review
Buildings new to the covered list3.7%1.9×Watch
Single-regime Article 3202.4%1.2×Normal

Bar: incomplete-record-rate lift vs. single-regime Article 320 baseline · scale 0–4.0× · tick marks the 2.0× review threshold

2 of 4 slices over threshold

Evidence-linked improvement

The year closes into the regression suite

A cycle closes when the non-filing is a regression case. That suite is what the next year reported is measured against.

Improvement cycle · five stages

Switchback — the path turns at Improve and returns at Learn

  1. 01

    Detect

    Incomplete records rise in one regime.

  2. 02

    Diagnose

    The building under its limit that owed a penalty anyway is read back until the cause narrows to one.

  3. 03

    Improve

    The change goes out with a number, and the reports behind it travel with it.

  4. 04

    Verify

    Nothing files while one touched report case is still red.

  5. 05

    Learn

    It is kept for good, and the pathway rules are amended in the same commit.

  6. Learn → Detect

    The return edge. Next year's detection meets a suite one case longer.

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, filing workflow, evaluation, integration, then production validation and handover.

Workstream

Week 1Week 2Week 3Week 4Week 5Week 6

  1. 01Portfolio, regime and pathway discovery workshops.Week 1
  2. 02Utility and property source assessment.Week 1, Week 2
  3. 03Pathway, certifier and filing-calendar mapping for each regime.Week 1, Week 2
  4. 04Meter and consumption data ingestion.Week 2
  5. 05Pathway determination and area checks.Week 2, Week 3
  6. 06Confidence scoring and exception routing.Week 3
  7. 07Director approval workflow.Week 3, Week 4
  8. 08Utility-data and portal integration.Week 5
  9. 09Pathway and data cases.Week 4, Week 5
  10. 10Guardrails and submission controls.Week 4
  11. 11Report-trail instrumentation.Week 5, Week 6
  12. 12Deployment, documentation and Agent Care handover.Week 6

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 tierPilotOne regime, one portfolioProductionProduction energy systemsAdvancedMultiple regimes / states
Introduced at Pilot
Assembly to your rule set✓✓✓
Director approval✓✓✓
Data-completeness baseline✓✓✓
Introduced at Production
Reporting by regime—✓✓
Approval workflow in your systems—✓✓
Approved write-back—✓✓
Utility-data integration—✓✓
Introduced at Advanced
Multi-regime and multi-state rules——✓
Multi-stage owner approvals——✓
Large portfolios——✓
Multi-regime reporting controls——✓
Build priceFrom $5,000From $8,000Custom quote
Final build price
Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
Separate from build
Build 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.

  1. 01Your building list and the regimes they report under

    Meter and property ingestion and regime binding

    Week 1
  2. 02The report filed last year for each building

    Pathway-determination baseline and the data-gap register

    Week 2
  3. 03Your certifier roster and the licences they hold

    Pathway, certifier and filing-calendar mapping

    Week 1
  4. 04Access to relevant APIs, feeds or exports

    Utility, portal and property assessment, then integration setup

    Week 2
  5. 05Reports you would not want re-filed

    Filing cases and failure-mode testing

    Week 4
  6. 06What no certified report may estimate

    Confidence scoring, exception routing, guardrails and approval controls

    Week 3
  7. 07Named portfolio directors to review records

    Director approval workflow, then pilot and production validation

    Weeks 5–6

Nothing else is required

Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Widths are read off the weeks the work actually fills, which is why two bands must share the fifth.

Phase

W1

W2

W3

W4

W5

W6

  1. DiscoveryW1
  2. BuildW2 – W3
  3. EvaluateW4 – W5
  4. Pilot & LaunchW5 – W6

Week focus

W1Portfolio and regime discovery, pathway mapping and the automation boundary

W2Utility integration and the data-completeness baseline

W3Filing workflow, confidence logic and approval controls

W4Evaluation suite, certifier checks and failure-mode testing

W5Portal integration, pilot buildings and targeted corrections

W6One reporting year run under the portfolio director, then Agent Care handover

Reading the band
Each bar spans only the weeks its work is named in. The week 5 overlap is real, not padding.
At the end of W6
Validation closes on live portfolios, and Agent Care picks up monitoring.
Duration
Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Real Estate AI agent

Build an energy-filing agent around the regimes your buildings report under.

Show us the buildings, the regimes they report under and who certifies each report. A filing has to survive a New York City audit that is still running, a false-statement misdemeanour under Administrative Code §28-320.6.3, and a §28-320.6.2 square-foot penalty that a building under its limit still owes — so that is what we build the record to carry.

Typical delivery
4–6 weeks
Deployment
Your buildings and meters
Certifier signs
Pre-submission
After launch
Agent Care available

Nestack Agents · Energy filings

AGT-RE-19 · Agent Care available after launch

Build an energy-filing agent

Four questions so the call starts from your actual setup — then pick a time that suits you.

About you
Your portfolio

No obligation. We’ll map the workflow, name the automation boundary and tell you if an agent is the wrong answer.