Start Here — What's the single asset that keeps you up at night?

• Which asset or unit would you pick for a pilot if you had to choose one today? Options: Gas turbine / generator, Compressor train, Steam turbine, Heat exchanger / economizer, Reactor / distillation column, Pump skid, Condenser, Other (please specify)
• Why that asset? Describe the specific operating risk, cost, or safety concern that makes it a priority.
• How often does this asset experience meaningful deviation from expected performance (unplanned downtime, derates, trips, or near-misses)? Options: Daily, Weekly, Monthly, Quarterly, Less than quarterly
• Can you point to a recent incident, upset, or near-miss where having a reliable simulation would have changed the decision made on the floor?
• Who on your team is closest to the day-to-day performance questions for this asset? Options: Reliability Engineer, Process Engineer, Operations Supervisor, Maintenance Lead, Asset Manager, Other (please name)
• If we could cut the frequency of unplanned events on this asset by half, what would that free up for your team—time, budget, confidence, other?

Show Me Where It Hurts (and what you’ve tried so far)

• Have you been accepting recurring losses because your current tools make testing changes feel too risky, too slow, or too uncertain? Options: Yes — feels too risky to test, Yes — data and tools are too slow, Partly — some risks are accepted, No — we actively test changes
• What monitoring, analytics, or models are you already running against this asset today (select all that apply)? Options: Basic historian trends, Rule-based alarms, Proprietary vendor model, In-house statistical models, No model / manual only, Other (please describe)
• When an alert or anomaly appears today, who typically starts the investigation and how long before action is taken? Options: Operator (minutes), Shift engineer (hours), Reliability team (days), Maintenance plan next turnaround, Other — explain
• Tell us about a diagnostic or corrective action you tried that didn’t work—what did you learn and how did it change your expectations?
• Which outcome frustrates you most when current tools fail—false positives, missed degradation, long root-cause cycles, or lack of operator trust? Options: False positives / alert fatigue, Missed or late detection, Long root-cause / investigations, Operators don’t trust the outputs, All of the above
• How does that frustration show up in daily operations—pressure on shifts, overtime, delayed turnarounds, or conservative operating limits? Options: Increased overtime, Conservative derating, Delayed projects/turnarounds, Higher spare-part inventory, Safety/near-miss concerns, Other (please explain)

If you could simulate one change before touching a valve, what would it be?

• What specific operational question would you most want the twin to answer before you test it on the live asset?
• How often do you need to run these ‘what-if’ scenarios—daily, weekly, seasonally, or only for big changes? Options: Daily, Weekly, Monthly, Seasonally, Only for major changes
• Which types of scenarios matter most to you (select all that apply)? Options: Load ramps / transient response, Feedstock or fuel quality change, Ambient condition shifts (temperature/humidity), Equipment degradation / fouling, Control setpoint changes, Turnaround/maintenance scenarios, Other
• What trade-offs are you willing to accept between model speed, interpretability, and raw predictive accuracy (e.g., a slower but explainable physics model vs a faster black-box ML model)? Options: Prefer explainable physics-first, Hybrid physics + ML preferred, Willing to use black-box if accuracy is materially better, Undecided — want recommendations
• Have you used physics-based models before? If yes, describe what worked and what didn’t for adoption and accuracy.
• What would make a simulated scenario recommendation persuasive enough that an operator or engineer would act on it?

Where your data lives — and where it falls apart

• If an external team asked for historian and sensor data for the pilot today, how often would that data arrive complete, time-synced, and labeled without rework? Options: Almost always, Often with minor clean-up, Sometimes with heavy clean-up, Rarely — substantial gaps
• Which systems will we need to integrate with for this pilot (select all that apply)? Options: OSIsoft PI / AF, GE Historian, AVEVA/Schneider Historian, SCADA, DCS, PLC direct, CMMS / maintenance records, Other (specify)
• What typical sample rates, tag counts, and retention windows should we expect (e.g., 1s for control, 1min for trends, 5 years retention)?
• Describe common data quality issues you face (sensor drift, missing timestamps, unit mismatches, tag renaming, noisy signals, etc.).
• Does your organization have any data access or cyber requirements we need to plan for (e.g., DMZ, read-only accounts, on-prem agent, vendor security review)? Options: DMZ required, On-prem agent only, Read-only historian account, Security review & POA&M, No special requirements known
• How long does it typically take to get data access approvals from IT/OT in your org? Options: <1 week, 1–3 weeks, 3–6 weeks, >6 weeks, Depends on legal/compliance

Who's in the room (and who needs convincing)

• Who would effectively block a pilot if they felt it created more work or risk than value (names or roles)?
• Which stakeholders need to sign off on the pilot to proceed (select all that apply)? Options: Reliability / RCM lead, Process engineering manager, Operations manager / shift lead, IT/OT manager, Plant manager / VP Operations, Asset owner / commercial, Legal / procurement
• Who on your side is likely to be the strongest internal champion for the twin, and why will they champion it?
• What are the executive priorities that will determine whether the pilot is seen as a success—cost reduction, safety, uptime, regulatory compliance, or innovation? Options: Cost reduction, Safety / incident reduction, Increased uptime, Regulatory compliance, Operational insight / visibility, Staff empowerment / training
• How does procurement typically prefer to structure pilots (PO for services, time-and-materials, fixed-scope pilot, master services agreement)? Options: PO for pilot services, Time-and-materials, Fixed-price pilot, Under existing MSA, Other (please describe)
• What internal objection have you seen most often when a new analytics or simulation tool is proposed? Options: Lack of trust in models, Fear of added workload, Security/data concerns, Budget constraints, No clear ROI

How we'll know it's working — real signals, not hunches

• If the twin produced highly accurate predictions but operators never used the outputs, would you consider that a successful pilot? Options: No — adoption matters as much as accuracy, Partly — technical validation is first step, Yes — proving accuracy is primary
• Which measurable success signals matter most for your team (select up to three)? Options: Reduction in unplanned downtime (hours), Improved prediction accuracy (MAE/RMSE/% within threshold), Fewer false alarms, Reduced start-up/shutdown excursions, Operator engagement / dashboard use, Maintenance cost savings, Faster root-cause resolution time
• What validation window would you prefer for pilot acceptance (we typically recommend 30–60 days): Options: 30 days, 45 days, 60 days, Longer than 60 days — explain
• What threshold of prediction performance would you need to see for engineering to sign off (e.g., X% within Y units, or specific alarm reduction)?
• Who will be the formal pilot acceptance signer(s) on your side? Options: Reliability lead, Process engineering manager, Operations manager, Plant manager, Other (please name)
• How do you want results delivered during the pilot—daily dashboards, weekly review calls, automated alerts, or integrated control-room displays? Options: Daily dashboards, Weekly executive summaries, Automated alerts to operators, Integrated DCS/SCADA displays, Ad-hoc deep-dive sessions

Obstacles, deadlines, and the small print

• What hidden constraints might quietly kill a pilot before validation—staff bandwidth, scheduled outages, upcoming upgrades, or contract/legal issues? Options: Staff bandwidth, Scheduled outage / turnaround, IT/OT upgrade, Contract/legal constraints, Budget reallocation, Other (please explain)
• How many engineering hours per week can you realistically commit to the pilot for data prep, calibration reviews, and validation feedback? Options: <4 hours/wk, 4–8 hours/wk, 8–16 hours/wk, >16 hours/wk
• Do you have any fixed deadlines or windows (regulatory inspections, peak seasons, maintenance windows) that will determine our start or validation timeline?
• What contractual or IP concerns should we be mindful of (data ownership, model IP, derivative insights)? Options: Customer owns data, Customer owns derived models, Vendor retains model IP (license to customer), Need custom IP terms, Undecided — want to discuss
• If we hit a technical roadblock, who should we escalate to on your side and what’s the fastest way to get an unblock?
• Assuming alignment, what is your desired pilot start month and an acceptable hard deadline for completion of validation? Options: Start within 2 weeks, Start in 1 month, Start in 2–3 months, Start later than 3 months
• What would make you comfortable moving from pilot to production (example: proven X% uptime improvement, 6-month ROI, operator adoption metric)?

• Solution Experience Kickoff — Current State, Consequence & Future State
• Validation Plan Workshop — Metrics, Window & Acceptance Criteria
• What‑If Scenario Definition Workshop — Realistic Inputs & Expected Outputs
• Model Calibration & Data Quality Session
• Parallel Validation Review & Acceptance Decision

• Host to run first baseline calibration and deliver initial performance plots for the checkpoint meeting.
• Customer to submit one-sentence current state, one-sentence future state, and consequence summary document.
• Customer to provide a sample extract (7–30 days) from historian/SCADA/DCS and metadata mapping.
• Introductions & Meeting Objectives
• Anchor: Future State -> Problem Mapping
• Create 3–5 high‑value, realistic what‑if scenarios explicitly tied to customer operating questions.
• For each scenario, document inputs, expected outputs, and the KPI-based pass/fail rule.
• Agree the data slices and event markers necessary to run each scenario.
• Define the validation checkpoint scripts to force SME confirmation after runs.
• Customer SMEs to approve the final list of scenarios and provide any missing event annotations.
• Host to create scenario templates and baseline scripts that map inputs to expected outputs and KPIs.
• Data owner to extract and deliver the agreed representative data slices and event logs.
• Calibration Approach Overview
• Agree the calibration approach, timeline, and checkpoints.
• Document data quality issues with owners and a remediation plan.
• Schedule the first baseline run and define expected outputs for review.
• Agree pre-work deliverables and schedule for the validation activities.
• Customer to provide corrected metadata, sensor units, and any missing tags flagged in the data review.
• Host and customer to agree on temporary imputation rules or virtual sensor definitions for the validation window.
• Recap Accepted Criteria & Future State
• Demonstrate that the twin proves the defined future state for the prioritized scenarios.
• Reach an explicit acceptance decision tied to the numeric KPIs and the operational consequence.
• If not accepted, agree a concrete iteration plan with owners and timeline.
• Define immediate next steps for production integration or remediation.
• Host to publish the full validation report with scenario results, plots, and pass/fail summary.
• Acceptance authority to sign acceptance, conditional acceptance, or request iteration within the agreed timeframe.
• If accepted, schedule the first deployment/gateway meeting and assign integration owners; if iteration, create a remediation backlog.
• Obtain a crystal‑clear current state stated in one sentence by the customer.
• Quantify the consequence of the current state in operational/financial terms.
• Agree a one‑sentence future state outcome that the Solution Experience must prove.
• Confirm available data streams, identify gaps, and assign owners for access.
• Host to prepare a tailored validation plan template based on the inputs for the next workshop.
• Recap Objectives & Constraints
• Agree a measurable validation plan with explicit KPIs and numeric acceptance thresholds.
• Lock the validation window and representative operating cases to be used.
• Assign owners and an acceptance authority along with a decision timeline.
• Document key validation risks and immediate mitigations.
• Host to publish the validation plan document with KPIs, thresholds, window, and roles for signature.
• Customer to confirm acceptance authority and provide final dates for the validation window.
• IT/OT to provision read access to selected historian slices and confirm data delivery method.
• One‑Sentence Current State (Customer‑led)
• Results Presentation — Scenario by Scenario
• Prioritize Operating Questions
• Validation Approach Overview
• Live Data Quality Review
• Calibration Tasks & Checkpoints
• Diagnosis: Failure Modes & Residuals
• Define Success Metrics & Acceptance Thresholds
• Consequence Quantification
• Define Scenario Templates
• Gap Remediation Plan
• Validation Decision & Rationale
• Identify Representative Data Slices & Event Markers
• One‑Sentence Future State
• Select Validation Window & Sample Cases
• Next Steps: Production Handover or Iteration Plan
• Baseline Run Schedule & Outputs
• Validation Checkpoints & Acceptance Phrases
• Roles, Governance & Decision Rule
• Data Inventory & Access Readiness

• Historian Data Ingestion and Cleaning
• Equipment Design Data Import
• Physics-Based Asset Model Implementation
• Data-Driven ML Model Training
• Model Calibration to Historical Operations
• Parallel Live-Data Twin Deployment
• Real-Time Monitoring Dashboard Deployment
• Automated Alert and Anomaly Rules Deployment
• Virtual Sensor Creation and Deployment
• Degradation Prediction Engine Deployment
• What-If Scenario Simulation Workspace
• DCS/SCADA Integration via OPC/REST
• CMMS Integration for Maintenance Workflows
• Engineer and Operator Training on Twin Use
• Continuous Model Retraining Pipeline Deployment

Historian Data Ingestion and Cleaning

• Which historian(s) or time-series systems hold the asset data we should ingest? Options: OSIsoft PI, AVEVA Historian, Honeywell PHD, Ignition, SQL/Time-series DB, Other / Custom
• What variables/tags are required for the pilot twin (e.g., temperatures, flows, pressures, setpoints)? List key tags or upload a sample mapping.
• What is the typical sampling frequency and expected data volume for the selected tags? Options: Sub-second (<=1s), 1–60s, 1–5min, 5–15min, Event-driven/Irregular
• What common data quality issues should we expect (select all that apply)? Options: Missing timestamps/gaps, Outliers/spikes, Duplicate points, Incorrect units/scale, Clock drift/timezone issues, None/Unknown
• Are there any access, security, or privacy requirements for historian ingestion (VPN, jumpbox, read-only accounts, anonymization)? Options: Read-only accounts, VPN / private network, Jumpbox/DMZ, Data anonymization required, No special requirements, Other

Equipment Design Data Import

• Do you have equipment design documents available for the pilot asset (P&IDs, datasheets, nameplates, OEM curves)? Options: All available, Partial available, Only datasheets/nameplates, None available
• In what formats are design documents currently stored? Options: PDF drawings, Excel/spec sheets, CAD/CAE files, Proprietary OEM format, Other
• Are there existing thermodynamic or process models (e.g., Aspen, HYSYS, vendor models) we can reference? Options: Yes — shareable, Yes — restricted, No
• Are OEM performance curves or calibration constants required from vendors to build the physics model? Options: Yes — available, Yes — need assistance to acquire, No
• Please list any critical mechanical or control limitations (e.g., max RPM, bypass valves, safety interlocks) that must be represented in the model.

Physics-Based Asset Model Implementation

• Which physical phenomena must the physics model capture for the pilot (e.g., thermodynamics, fluid dynamics, heat transfer, rotor dynamics)? Options: Thermodynamics/energy balance, Fluid dynamics/flow networks, Heat transfer, Rotordynamics/vibration, Control loop behavior, Other
• Do you require a first-principles model from scratch, adaptation of an existing model, or a hybrid physics+data approach? Options: New first-principles model, Adapt existing model, Hybrid physics + ML (recommended), Unsure — need consult
• What level of fidelity is required for the pilot (steady-state only, dynamic transient response, sub-second dynamics)? Options: Steady-state, Slow dynamics (minutes-hours), Fast dynamics (seconds-minutes), Sub-second/high-frequency
• Are there proprietary equations, vendor IP, or safety constraints that limit the model implementation decisions? Options: Yes — limited sharing, Yes — fully shareable, No
• Who will be the technical owner for physics model decisions on the customer side (role/title)?

Data-Driven ML Model Training

• Which ML objectives are prioritized for this pilot (e.g., residual correction, anomaly detection, virtual sensors, predictive degradation)? Options: Residual correction / bias compensation, Anomaly detection, Virtual sensors, Predictive degradation, Optimization surrogate models, Other
• What historical data window is available for training (e.g., 3 months, 1 year, multiple years)? Options: <1 month, 1–3 months, 3–12 months, 1–3 years, >3 years
• Are labeled events available for supervised training (e.g., failures, maintenance logs, performance test labels)? Options: Yes — event timestamps available, Partial labeling, No labeled events
• What are acceptable model development constraints (training on-prem vs cloud, data retention, compute limits)? Options: On-prem training required, Cloud acceptable, Hybrid/edge inference, Unsure — discuss
• Do you require explainability/feature importance reporting for ML models to satisfy engineering reviews? Options: Yes — required, Optional, No

Model Calibration to Historical Operations

• Which historical period should be used for calibration to represent typical and extreme operating conditions? Options: Recent steady months (30–90 days), Full year (seasonal), Period including known excursions, Customer selected window
• What calibration accuracy targets or KPIs will determine success (e.g., RMSE, MAE, % error bounds)? Options: RMSE target, MAE target, % deviation band, Other / custom
• Who will sign off on calibration results (roles: reliability engineer, process manager, operations lead)? Options: Reliability Engineer, Process Engineering Manager, Operations Manager, IT/OT Lead, Other
• Are there maintenance or operational records available to correlate model errors with equipment state? Options: Yes — full maintenance logs, Partial records, No
• Are we allowed to run calibration iterations against live historical data or only offline sandboxes? Options: Offline sandbox only, Allowed on historical production data, Need approval

Parallel Live-Data Twin Deployment

• Do you want the twin to run in parallel with the live asset during a validation window of 30–60 days? Options: Yes — 30 days, Yes — 60 days, Prefer custom window, No / not needed
• What target latency is required for parallel predictions (real-time, 1–5 min, 5–30 min, hourly)? Options: Real-time (<1s), Near-real-time (1–60s), Short-window (1–5min), Batch (5+ min)
• Who will monitor the parallel run on your side and how should discrepancies be reported? Options: Operations lead, Reliability engineer, Process engineer, Automated alerts only
• Are there constraints for connecting to live systems (maintenance windows, read/write restrictions)? Options: Read-only allowed, Read/write allowed, Maintenance window required, No live connections allowed
• What acceptance criteria must be met at end of parallel deployment to consider pilot successful? Options: Pre-agreed accuracy threshold, Operational validation by engineers, No critical false alerts, Other — specify

Real-Time Monitoring Dashboard Deployment

• Which users/groups need dashboard access (operators, engineers, managers, executives)? Options: Operators, Reliability Engineers, Process Engineers, Operations Managers, Executives
• What KPIs and visualizations are required (e.g., predicted vs actual, residuals, trends, alarms)? Options: Predicted vs Actual, Residual/Error trends, Health score / degradation, What-if scenario results, Custom KPI
• Do dashboards need to be embedded in existing control-room displays or served via web/mobile portals? Options: Embedded in control-room HMI, Web portal, Mobile app, Integrated into MES/portal
• Are there authentication/SSO requirements or role-based views to implement? Options: SAML/SSO, LDAP/AD, Role-based views required, No special auth required
• What refresh frequency and historical window are required for dashboard charts? Options: Sub-second/real-time, 1–60s, 1–5min, Hourly, Historical up to 1 year

Automated Alert and Anomaly Rules Deployment

• What types of alerts are needed (threshold breaches, anomaly detection, model divergence, maintenance triggers)? Options: Threshold breaches, Anomaly detection, Model divergence/residuals, Maintenance/workflow triggers
• What delivery channels should be used for alerts (email, SMS, SCADA alarm, CMMS ticket)? Options: Email, SMS/Push, SCADA/HMI alarm, CMMS ticket, Slack/Teams
• What severity levels and escalation paths should alerts follow? Options: Info/Informational, Warning, Critical/Escalate, Custom escalation matrix
• Are there existing anomaly rules or thresholds we should import, or should we define new ones? Options: Import existing rules, Define new rules with customer, Hybrid
• Do alerts require enrichment with process context or suggested actions (e.g., checklist for operators)? Options: Yes — include suggested actions, Optional, No

Virtual Sensor Creation and Deployment

• Which physical measurements are unavailable or unreliable and should be replaced with virtual sensors?
• What accuracy and latency requirements do virtual sensors need to meet? Options: High accuracy (<5% error), Moderate accuracy (5–15%), Low accuracy acceptable (>15%), Low latency/real-time required
• Are virtual sensors for control loops (closed-loop) or for monitoring/analysis only? Options: For control loops, Monitoring/analysis only, Both
• Where should virtual sensor outputs be published (historian tags, dashboards, control system)? Options: Historian tags, Dashboards, Control system (DCS/SCADA), CMMS/maintenance workflows
• Do virtual sensors require certification or validation by OEM or plant engineering before use? Options: Yes — OEM/engineering sign-off, No — internal validation acceptable, Unsure

Degradation Prediction Engine Deployment

• Which degradation modes are most important for the pilot (e.g., fouling, erosion, bearing wear, efficiency loss)? Options: Fouling, Erosion, Bearing wear, Efficiency/performance loss, Corrosion, Other
• What lead time do you require for actionable degradation predictions (days, weeks, months)? Options: Hours, Days, Weeks, Months
• Do you have failure/performance thresholds that should trigger maintenance workflows? Options: Yes — defined thresholds, Partial thresholds, No
• How should predicted degradation translate into actions (automated CMMS ticket, investigation alert, schedule maintenance)? Options: Auto-create CMMS ticket, Send investigation alert, Schedule preventive maintenance, Advisory only
• Are there historical failure/inspection records to validate degradation models? Options: Comprehensive records, Partial records, None

What-If Scenario Simulation Workspace

• Which scenario types should be supported (operational setpoint changes, equipment failures, feedstock changes, ambient variations)? Options: Setpoint changes, Equipment failures, Feedstock/quality changes, Ambient/environmental changes, Maintenance strategies
• Who will be allowed to run scenarios (operators, engineers, managers) and what guardrails are required? Options: Operators (limited), Engineers (full access), Managers (view only), Cross-functional teams
• Do scenarios need to run real-time interactive simulations or offline longer-horizon analyses? Options: Interactive real-time, Offline batch (longer horizon), Both
• Should scenario results be stored and auditable for regulatory/process reviews? Options: Yes — store with audit trail, Optional, No
• Are there pre-defined scenario templates or use cases you want pre-built for the pilot? Options: Yes — provide templates, No — create new templates, Unsure

DCS/SCADA Integration via OPC/REST

• Which protocols and endpoints are available for integration (OPC DA/UA, Modbus, REST APIs, MQTT)? Options: OPC UA, OPC DA, Modbus, REST API, MQTT, Other
• Are integration connections allowed directly to DCS/SCADA or must they go through an intermediary (e.g., historian, DMZ)? Options: Direct connection allowed, Use historian intermediary, DMZ/jumpbox required, Not allowed
• Do you require write-back capabilities (setpoints, advisory commands) or read-only integration? Options: Read-only, Write-back allowed (restricted), Write-back not allowed, Unsure
• What change management approvals are required to integrate with operational control systems? Options: Engineering change approval, IT/OT security approval, Operations sign-off, No formal approvals
• Are there latency or determinism requirements for data exchanged with DCS/SCADA? Options: Hard real-time required, Near-real-time acceptable, Batch updates acceptable

• Non-Disclosure Agreement (NDA)
• Master Services Agreement (MSA)
• Statement of Work (SOW)
• Commercial Terms & Order Form
• Payment Schedule & Invoicing
• Pilot Acceptance Criteria & Validation Plan
• Data Access, Integration & Security Approval
• Data Processing Agreement (DPA)
• Service Level Agreement (SLA) - Pilot Support
• Intellectual Property & Licensing Agreement
• Roles, Responsibilities & Governance Plan
• Change Control & Scope Amendment
• Production Transition & Expansion Agreement
• Regulatory & Compliance Confirmation
• Insurance, Liability & Indemnity Confirmation

• Pilot Validation & Acceptance Review
• Operations Handover & Runbook Workshop
• Lessons Learned & Continuous Improvement Retrospective
• Asset Expansion Prioritization & Scale Roadmap
• Issue Triage & Enhancement Governance Setup

• Produce a prioritized expansion roadmap document with owners and high-level estimates.
• Deliver and get sign-off on operational runbooks, dashboards, and alerting behavior.
• Align on operational owners, SLAs, and escalation paths.
• Schedule and commit to operator and engineering training with clear acceptance criteria for proficiency.
• Share final runbooks, playbooks and access credentials to operations and archive in the agreed repository.
• Configure and test alert delivery path and confirm a test alert with operations.
• Schedule operator training sessions and assign trainees for completion tracking.
• Purpose, Scope and Ground Rules
• Document a clear, actionable lessons-learned report capturing root causes and mitigations.
• Create and prioritize an enhancement backlog with owners and estimated effort.
• Agree updates to standards and processes to reduce recurrence of identified issues.
• Draft and circulate the lessons learned report including RCA and recommended mitigations.
• Populate the enhancement backlog in the issue tracker and assign owners with target dates.
• Update modeling and data quality standards documents and circulate for approval.
• Recap of Pilot Outcomes and Business Case
• Create a prioritized, time-bound expansion roadmap for additional assets.
• Align on resource commitments, estimated effort and preliminary commercial approach.
• Identify key risks and mitigation plans for scaling the program.
• Introductions & Meeting Objective
• Prepare detailed SOW and commercial proposal for the first 1-2 expansion assets.
• Schedule a follow-up governance meeting to approve funding and start dates.
• Purpose and Cadence
• Establish a documented triage and governance process with SLAs and owners.
• Assign and schedule resolution for outstanding issues from the pilot.
• Agree on a safe, repeatable release process for model and dashboard updates.
• Create the triage board in the issue tracker with severity fields and SLA automation.
• Assign owners and target dates to all outstanding validation issues.
• Publish the release management checklist and schedule the first maintenance release window.
• Verify the twin meets the predefined success signals with evidence and representative examples.
• Secure a formal acceptance decision or a prioritized remediation plan with owners and timelines.
• Ensure all stakeholders confirm that demonstrated outcomes tie directly to their stated operational consequences.
• Produce and circulate a validation evidence package (metrics, plots, datasets, and test cases).
• If remediation required, create a remediation plan with tasks, owners, acceptance criteria and target completion dates.
• If accepted, schedule the Operations Handover & Runbook Workshop and assign initial owners for handover artifacts.
• Handover Status & Acceptance Recap
• What Worked Well
• Roles, Responsibilities and SLA Overview
• Severity Matrix, SLAs and Escalation Paths
• Selection Criteria & Candidate Review
• Current State, Consequence, Future State (Preconditions)
• Estimate Effort, Data Readiness and Risk per Asset
• Review Outstanding Validation Issues & Assign Owners
• Validation Methodology & Dataset
• Runbooks, SOPs and Playbooks
• What Did Not Meet Expectations
• Results: Metrics, Time-series Overlays, and What-if Cases
• Root Cause Analysis
• Phased Roadmap & Resource Plan
• Dashboards, Alerts and KPI Definitions
• Release Management and Test Requirements
• Improvement Backlog & Prioritization
• Data Access & Integration Handover
• Reporting, Metrics and Continuous Feedback Loop
• Commercial Model and Governance for Scale
• Forced Validation Checkpoints
• Decision: Accept, Accept-with-Conditions, or Remediate
• Update Standards & Next Steps
• Decision, Risks and Next Steps
• Training & Knowledge Transfer Plan
• Confirm Recurring Meeting Invite and Charter