Start here, briefly

• Tell me in one sentence what prompted you to consider replacing or modernizing your eligibility system now
• How soon does the federal corrective action or timeliness requirement constrain your decision window Options: Within 3 months, Within 6 months, Within 9 months, Within 12 months, More than 12 months
• Who is the single executive or office that will be held accountable by the governor for meeting the federal requirements
• Describe one recent incident where your current system or operations missed a federal timeliness target or produced significant accuracy exceptions

Where benefit delivery is most at risk

• If a single operational failure during transition would force you to stop the project immediately, what would that be
• Describe the legacy processes or interfaces that most frequently cause eligibility errors or delays Options: Batch data extracts with manual reconciliation, COBOL-era eligibility rules encoded in legacy files, Paper-first workflows requiring manual scanning, Multiple point-to-point interfaces, Other
• How often do those problematic legacy paths cause missed federal timeliness or corrective actions Options: Weekly, Monthly, Quarterly, Rarely
• Who on your team first notices a downstream benefit disruption when an interface or rule misfires Options: Eligibility managers, IT operations, Program policy office, Contact center supervisors, Other
• When those disruptions happen, what is the immediate impact on beneficiaries and on your reporting to federal partners

The rules no one wants to lose

• What legacy rule or edge case would you most worry a modern rules engine would fail to reproduce
• How well documented are those edge cases in machine-readable form, policy memos, or adjudicator notes Options: Fully documented and versioned, Partially documented, Mostly tribal knowledge, No documentation
• Tell me about a recent example where an exception required manual escalation to a supervisor or attorney
• Which program types contain the highest concentration of these exceptions Options: Medicaid, SNAP, TANF, Unemployment insurance, Child care subsidies, Multiple programs
• If the rules engine could not reproduce 1 percent of your current edge-case outcomes, what would that cost you in federal findings, and would you proceed Options: It would trigger a stop work decision, We could accept and mitigate, Depends on which cases, Unsure

Who signs, who stops, who escalates

• Who has final authority to approve a cutover or to pause a parallel run if issues arise
• Which internal stakeholders need to be consulted before any public-facing communication about change Options: Program leadership, Legal counsel, IT security, Communications office, Governor's office, Other
• How do your federal partners prefer to be notified and engaged when a state is transitioning eligibility systems Options: Formal briefing and weekly reports, Monthly status updates, On-demand if thresholds hit, Minimal unless asked
• Describe the approval path for temporary staffing increases in the contact center during a surge or parallel run
• Which decision or governance gap, if left unresolved at kickoff, would make you decline to start the engagement

Operational readiness and hard constraints

• Which source systems must the modern platform integrate with to process eligibility end-to-end Options: Legacy eligibility database, Benefit payment system, Case management records, Identity verification service, External data sources (employment, income), Other
• Is there an existing API or near-real-time feed for each required system, or will integration require batch extracts and screen scraping Options: APIs available for required systems, Partial API coverage, some batch extracts, No APIs, mostly batch or manual, Unknown
• Who owns access and authorization for those interfaces, and can they commit named contacts for integration work
• How complete and clean is the data you will need for initial rule validation and parallel testing Options: Ready for ingestion, Requires mapping and transformation, Significant cleanup needed, Data access blocked by policy or contracts
• Which regulatory approvals, legal reviews, or data-sharing agreements remain outstanding that could block the project start
• If integration or data access cannot be completed within your target transition window, what is your fallback Options: Extend transition, Limit scope, Delay cutover until resolved, Other

What options are you actively weighing

• If you step back, which alternatives are on the table right now Options: Retain incumbent contractor, Build and operate internally, Replace legacy stack with managed vendor, Short-term contractor augmentation, Other
• What specific capability or proof would have to exist for you to keep the current system and not change anything
• Has any internal team proposed solving this without an outside partner, and if so which team and what is their proposed timeline Options: Yes, internal IT, Yes, program office, No internal proposal, Unsure
• Which vendor or incumbent strengths do you find most attractive, and why would that matter at cutover
• What single outcome from a pilot or proof of concept would make you sign a contract within one week Options: Match legacy accuracy for all test cases, Demonstrate federal timeliness at scale, Show clear staffing and surge plan, Prove data and interface parity

Measuring success before you say yes

• How do you define acceptance for parallel operations, what metrics and thresholds will trigger formal sign-off
• Which accuracy, timeliness, and citizen satisfaction targets are nonnegotiable for federal compliance Options: Accuracy 98%+, Timeliness 95%+ on required windows, Citizen satisfaction measurable improvement, Other
• Who will own the acceptance tests, and do they have bandwidth to run case-level reconciliation during parallel processing Options: State acceptance team, Joint vendor-state team, Vendor-owned tests with state oversight, Other
• If the pilot demonstrates your targets except for a narrowly defined set of edge cases, what is your tolerance and remediation expectation Options: Accept with remediation plan, Require full parity before sign-off, Depends on case types, Unsure
• Which acceptance failure would push you to require an extended parallel run instead of cutover

Risks to beneficiaries and how you mitigate them

• What is the single citizen-facing risk you fear most during transition Options: Missed payments, Incorrect eligibility denials, Long contact center wait times, Privacy or data errors, Other
• Describe your current contingency playbook for missed payments or mass appeal events
• How quickly can you scale contact center capacity if case volume spikes during a parallel run Options: Within days, 1 to 2 weeks, More than 2 weeks, Not sure
• Who is authorized to enact emergency surrogate payments or manual processing when automated flows fail
• If beneficiary harm occurs during transition, who is the escalation path that reports to federal auditors

Timeline pressure and decision levers

• Given your federal deadline, which project milestone is the true make or break for staying on schedule Options: Named data feeds ready, Staffing and training complete, Rules validated in parallel, Acceptance tests passed
• How flexible is your procurement and contracting timeline if additional mitigation is required Options: Can accelerate, Standard timeline, Requires legislative or executive approval, Unknown
• Walk me through the minimum set of resources you must have in place before a parallel run can begin
• If a vendor can prove performance in a short pilot, what internal obstacle would still prevent you from moving forward immediately
• Realistically, how soon after a successful pilot would you be prepared to initiate transition activities Options: Immediately, Within 30 days, Within 60 to 90 days, Longer than 90 days

Next steps and commitment signals

• What evidence or deliverable from a seller would most reduce your perceived risk and accelerate your internal approval Options: Test case parity report, Named integration contacts and window, Staffing and surge plan with references, Formal data sharing agreement
• Which members of your team should be included in a technical kickoff to avoid rework Options: Program leadership, Integration engineers, Policy analysts, Contact center ops, Legal and compliance
• Who needs to sign final acceptance and what is their expected decision timeline once acceptance criteria are met
• If we could remove one of your top three concerns this week, which would you want it to be and why
• Would you be open to a short technical proof of concept focusing on your highest-risk rule families, and if yes, what is your preferred timing Options: Yes, within 2 weeks, Yes, within 1 month, Yes, within 2 to 3 months, Not at this time

• Operational Rules Walkthrough

• You confirm that the demonstrated rule outcomes eliminate the specific accuracy failures and rework you described.
• Confirm the current state and what it is costing you
• Provide a representative sample of legacy cases and the top 10 known edge cases for a tailored parity run.
• You confirm the demonstrated surge run meets the federal timeliness windows you must maintain.
• Orientation: end-to-end transition timeline and decision points
• Deliver a sample parallel-run plan and an accuracy and timeliness runbook based on the provided cases within five business days.
• List the federal timeliness benchmarks and any active corrective-action deadlines to be reflected in acceptance criteria.
• Proof 1 — Rules engine parity on representative cases
• You identify any remaining edge cases or acceptance criteria that must be addressed before a procurement decision.
• Proof 2 — Contact center handling and accuracy on policy questions
• Proof 3 — Surge scenario and timeliness validation
• Address legacy edge cases and remediation paths
• Validate the future state with you

• Operational Rules Walkthrough
• Operational Rules Walkthrough Deck
• Operational Rules Walkthrough Brief

• meeting
• slides
• document

• Convert and Migrate COBOL Case Records
• Deploy Eligibility Rules Engine
• Encode Legacy Policy Rules and Edge Cases
• Operate Automated Eligibility Determinations
• Run Parallel Processing with Legacy System
• Perform Quality Assurance and Case Reviews
• Operate BPO Casework and Backlog Processing
• Operate Citizen Contact Center with Policy-Trained Agents
• Generate Automated Notices and Correspondence
• Prepare and Submit Federal Timeliness Reporting
• Provide Audit Trail and Case Documentation
• Scale Infrastructure for Emergency Volume Surges
• Execute System Cutover and Go-Live

Convert and Migrate COBOL Case Records

• How many active and inactive COBOL case records exist in your mainframe eligibility store (separate counts for active, pending, archived)? Options: Less than 100,000, 100,000-500,000, 500,000-2,000,000, More than 2,000,000
• Which export formats are available from your COBOL environment for a full case extract (select all that apply)? Options: Fixed-width mainframe tape/export, Delimited file (CSV/pipe), Relational database dump, Sequential EBCDIC dataset, Other
• Do you have existing field-level documentation for COBOL case layouts such as copybooks or record maps? Options: Yes, full copybooks available, Partial field maps available, No formal documentation
• Who will own decisions on disputed field mappings during migration (for example household composition, benefit issuance history, citizenship flags)?
• When do you plan to open a migration test window for initial extracts and reconciliation with the modern case store? Options: Within 2 weeks, 2-6 weeks, 6-12 weeks, Not yet scheduled
• How will you verify migration completeness against your COBOL master file (for example record counts, checksum reconciliation, and a sample of 500 high-risk cases)? Options: Record counts only, Record counts plus checksum sample, Full field-by-field reconciliation on sample, Other evidence (describe)

Deploy Eligibility Rules Engine

• Do you require the rules engine to support real-time eligibility checks, batch recalculations, or both for programs such as SNAP and Medicaid? Options: Real-time only, Batch only, Both real-time and batch
• Which integration endpoints must the rules engine call during determination (for example income verification feed, federal data hub, state wage/interchange APIs)? Options: State wage API, Federal data hub, Third-party income verifier, No external calls required, Other
• How many concurrent determinations per second must the engine support to meet peak caseworker and call center demand? Options: Less than 5/sec, 5-50/sec, 50-200/sec, 200+/sec
• Who on your team will own rule version approvals and sign-off for policy updates during the transition?
• Describe any federal business rules or program-specific regimes (for example categorical eligibility for SNAP, MAGI rules for Medicaid) that the engine must model from day one.
• Do you require a sandbox environment to validate rule changes with anonymized production data before promotion to parallel processing? Options: Yes, No, Prefer discussion

Encode Legacy Policy Rules and Edge Cases

• List the top three legacy edge-case scenarios (for example simultaneous SSI and SNAP eligibility, retroactive benefit adjustments, multi-jurisdiction household splits) that previously required manual overrides.
• Does your legacy rules inventory include a prioritized list of rules by failure risk (for example rules that triggered federal corrective action)? Options: Yes, prioritized list available, Partial inventory, No inventory available
• Which policy source documents must be encoded into the rule set (for example state SNAP handbook sections, state plan appendices, federal program memos)? Options: State policy handbook, State plan appendices, Federal program memos, Court guidance or settlements, All of the above
• Who will adjudicate rule encoding disputes where legacy behavior differs from written policy (for example legacy exceptions not found in the state handbook)?
• How many legacy rule variants or special-case handlers must be modeled initially (estimate count of unique conditional branches from your COBOL logic)? Options: Fewer than 50, 50-200, 200-1,000, More than 1,000
• What acceptance criteria will confirm parity between encoded rules and legacy determinations on a representative sample of claim histories (for example 98% first-pass parity on a 10,000-case sample)? Options: 98% parity on 5,000 sample, 95% parity on 10,000 sample, Custom threshold (describe)

Operate Automated Eligibility Determinations

• Do you expect automated determinations to clear a target percentage of incoming cases without manual review (specify percentage for SNAP, TANF, Medicaid)? Options: Less than 25%, 25-50%, 50-75%, 75%+
• Which case attributes should automatically escalate to manual review (examples: conflicting income feeds, missing SSN, multi-state residency)? Options: Conflicting income feeds, Missing identification, Retroactive benefit changes, Other
• Who will be the named owner for ongoing tuning of automated determination thresholds (title or role)?
• Describe the expected workflow for cases routed from automation to a human caseworker, including time-to-assignment and case notes requirements.
• Do you require daily metrics on automated-pass rates and false-positive manual referrals for program reporting? Options: Yes, No, Only weekly summaries
• Which enrollment channels must feed into automated determinations (for example intake web portal, phone intake from contact center, mailed applications)? Options: Web portal, Phone intake, Mail, In-person, Other

Run Parallel Processing with Legacy System

• When do you plan to begin a parallel-run period with the legacy COBOL system (start month or milestone)? Options: Immediately after migration, Within 1 month, 1-3 months, Undecided
• Which reconciliations must run daily during parallel processing (for example eligibility parity, benefit issuance totals, caseload counts)? Options: Eligibility parity, Benefit issuance totals, Caseload counts, All of the above
• Do you have named owners for daily parallel reconciliations and exception resolution for each program (provide role or team)?
• How will you surface mismatches between systems for root-cause analysis (for example automated flagged case list, severity coding, required documentation)? Options: Automated mismatch reports, Manual audit list, Both
• What SLA target will you set for resolving parity exceptions during the parallel run (for example 5 business days for critical mismatches)? Options: 48 hours, 5 business days, 10 business days, Custom
• Do you require a rolling reconciliation sample size (for example 1,000 cases per week) to validate ongoing parity with the legacy system? Options: Yes, specify sample size, No, ad hoc samples only, Prefer recommendation

Perform Quality Assurance and Case Reviews

• Do you require a QA program that includes both automated checks (rule audit trails) and manual case reviews for eligibility decisions? Options: Both automated and manual, Automated only, Manual only
• Which QA sampling cadence do you prefer for high-risk programs (for example weekly sample of 200 SNAP cases, monthly 100 Medicaid cases)? Options: Weekly samples, Biweekly, Monthly, Custom cadence
• Who will approve QA severity classifications that trigger federal corrective action reporting?
• Describe the corrective feedback loop you expect from QA to the rules engineering team (for example ticket, priority SLA, source code change request).
• Do you require QA dashboards with drill-down to case-level evidence for auditors and program managers? Options: Yes, No, Limited summary only
• Are there program-specific error types that must be tracked separately (for example client identity errors, income underreporting, missed retro adjustments)? Options: Identity errors, Income errors, Retro adjustments, Other

Operate BPO Casework and Backlog Processing

• How large is the current manual backlog by program that you expect the BPO team to address (provide counts per program: SNAP, TANF, Medicaid)?
• Do you require time-to-clear SLAs for backlog cases (for example clear 80% of backlog within 90 days)? Options: Yes, No, Need to define
• Who will be the state's single point of contact for intake and prioritization of backlog work?
• What training artifacts must BPO staff complete before handling program cases (for example policy modules, certification on eligibility rules, shadowing hours)? Options: Policy modules, Certification required, Shadow hours, All of the above
• Describe any security clearance or background checks required for BPO agents to access PII in case records.
• Do you require performance reporting on BPO productivity by case type and error rate for monthly governance? Options: Yes, weekly and monthly, Monthly only, No

Operate Citizen Contact Center with Policy-Trained Agents

• What contact center channels must be supported on day one (for example voice, web chat, SMS, email) for benefit inquiries? Options: Voice, Web chat, SMS, Email, All listed channels
• How many policy-trained agents do you expect to staff initially for peak hours to meet federal timeliness standards for phone-based applications? Options: Less than 50, 50-200, 200-500, 500+
• Do you require custom contact scripts that include program-specific interview prompts (for example income probing for SNAP, household composition for Medicaid)? Options: Yes, full script bundles, Yes, outline only, No
• Who will own the agent certification standard and re-certification cadence for policy proficiency?
• Specify your target average speed-to-answer and abandonment rate for voice channels during peak intake windows. Options: Answer within 30 seconds / <5% abandonment, Answer within 60 seconds / <10% abandonment, Custom targets
• Do you require call recording retention and transcript indexing for audit and QA purposes, and if so for how long? Options: Yes, 1 year, Yes, 3 years, Yes, custom duration, No

Generate Automated Notices and Correspondence

• Which notice types must be templated and generated automatically from determinations (for example approval, denial, request for information, change in benefit)? Options: Approval, Denial, Request for information, Change notices, All of the above
• What language coverage do your notices require at go-live (for example English plus Spanish and other priority languages)? Options: English only, English + Spanish, English + Spanish + other
• Do you require system-driven deadlines embedded in notices to satisfy federal timeliness notification rules? Options: Yes, No, Need guidance
• How will you test notice accuracy and regulatory compliance prior to cutover (for example legal review cycles, sample batch proofs)? Options: Legal review + sample proofs, Automated template checks, Ad hoc reviews
• Who will approve final notice language and effective dates for state-specific program communications?
• Do you require integration with a print-and-mail vendor or preference for electronic delivery only for benefit notices? Options: Print-and-mail integration, Electronic delivery only, Hybrid

Prepare and Submit Federal Timeliness Reporting

• Which federal timeliness metrics must the system produce automatically (for example percent applications processed within required days for SNAP, Medicaid queue times)? Options: Percent processed within days, Average processing time, Missed-timeliness counts, All of the above
• Do you require automated feeds or formatted exports to federal reporting endpoints on a weekly or monthly cadence? Options: Weekly, Monthly, Ad hoc, No automated feeds
• Who will own certification of the federal reporting extracts and submission schedule?
• Describe any historical timeliness penalties or corrective action plans that should inform report thresholds and exception handling.
• Are there program-specific reporting templates mandated by your federal partner that must be replicated exactly in format and fields? Options: Yes, Partial overlap, No
• What SLA do you require for generating and validating federal timeliness extracts prior to submission (for example 48 hours for monthly extract validation)? Options: 24 hours, 48 hours, 5 business days, Custom

• Non-Disclosure Agreement (NDA)
• Master Services Agreement (MSA)
• Statement of Work (SOW)
• Service Level Agreement (SLA)
• Pricing & Payment Schedule
• Data Processing Agreement (DPA)
• Data Sharing Authorization
• Acceptance Test Plan & Criteria
• Transition, Parallel Run & Cutover Governance Agreement
• Change Order Agreement
• Termination & Exit Agreement
• Surge Capacity & Emergency Operations Addendum
• Public Sector Procurement & Security Addendum

• Go-live Health Check
• First Measurement Review
• Acceptance Gate Decision
• Monthly Remediation Backlog and Federal Reporting Review
• Quarterly Outcome Realization Review

• Publish the prioritized backlog for the next 30 days with target completion dates for high-priority items.
• Produce a timeliness remediation plan showing process changes and expected compliance improvement by the acceptance gate.
• Deliver QA sampling evidence and representative case examples for the acceptance packet.
• Restate numeric acceptance criteria from Solution Scope
• Produce a documented pass/fail record for each acceptance criterion and capture the designated buyer signatory's decision.
• Confirm the plan and timeline for incumbent system decommissioning or read-only retention and data archiving.
• Publish the acceptance decision record with pass/fail results and any conditional remediation timelines.
• Deliver incumbent decommissioning status report showing archive completion or read-only configuration and scheduled cutover of access.
• If criteria failed, publish the retest plan with target dates and required evidence for re-evaluation.
• Open remediation backlog review
• Reduce the number of high-priority remediation items month-over-month and shorten MTTR for severity-1 incidents.
• Ensure federal reporting completeness and on-time filing rate remain within acceptable thresholds and remediation items are assigned.
• Re-confirm acceptance criteria and owners
• Provide a federal reporting exceptions log with corrective actions and expected filing dates.
• Deliver a short CSAT remediation plan for any downward trends identified during the review.
• Executive summary of outcomes vs Solution Scope targets
• Confirm whether the primary outcome metrics meet or are trending toward Solution Scope targets and document any strategic remediation needs.
• Agree the prioritized backlog for the coming quarter with clear acceptance criteria for each item.
• Publish the quarterly outcome report with metric trends, sample evidence, and a prioritized backlog for the next quarter.
• Deliver a federal reporting audit-prep checklist showing current readiness and outstanding actions.
• Provide a QA improvement plan for any recurring accuracy failure modes identified in the detailed review.
• Confirm deployment checklist items are complete or have defined remediation actions with target dates.
• Agree a documented 7-day remediation plan for all high-severity items discovered during early operations.
• Publish the 7-day remediation plan with owners and resolution dates.
• Deliver a short findings note listing integration checks that passed and failed for the next operational sync.
• Present first measurement data vs Solution Scope targets
• Determine whether first-pass accuracy rate and federal timeliness compliance are on trajectory to meet Solution Scope targets or require remediation.
• Agree a prioritized corrective action plan with completion dates ahead of the acceptance gate.
• Produce a root-cause summary for each accuracy gap with proposed fixes and target completion dates.
• Deployment and integration validation
• Root cause analysis for accuracy shortfalls
• Detailed review on accuracy and QA program
• Incident SLA and MTTR dashboard
• Present outcome data against each acceptance criterion
• Diagnose timeliness exceptions
• Document pass/fail per criterion and capture acceptance decision
• Federal reporting status and exceptions
• Parallel-run health and early signals
• Citizen experience and contact center performance
• Incumbent system wind-down verification
• Citizen satisfaction and QA trends
• Quality sampling and contact center performance
• Federal reporting and audit readiness
• Adoption and staffing readiness
• Agree corrective actions and timeline to acceptance gate
• Prioritization and next-month commitments
• Agree remediation plan for any failed criteria
• Open blockers and immediate remediation plan
• Quarterly backlog prioritization and roadmap for remediation