ERP

What to Do in the First 90 Days After an ERP Go-Live (2026 Government Guide)

What to do in the first 90 days after ERP go-live in government: hypercare command center, payroll, vendor payments, bank recs, exit criteria and handoff.

Chris Rodney 20 min read
On this page
  1. Key takeaways
  2. What hypercare means after a government ERP go-live
  3. Set up the hypercare command center before day one
  4. First 30 days post go-live: protect pay, payments and cash
  5. Days 31 to 60 post go-live: close the books and fix what’s broken
  6. Days 61 to 90 post go-live: hold the vendor to the contract and plan the handoff
  7. How to communicate with council, staff and vendors during hypercare
  8. 90-day post go-live checklist
  9. FAQ
  10. Plan the first 90 days after ERP go-live before cutover

The first 90 days after ERP go-live are your stabilization window, usually called hypercare. Use them to run a daily command center, protect payroll, vendor payments and cash, close the first month, clean up security and reports, and hold the vendor to its punch list. Leave hypercare only when agreed stabilization metrics are met.

Governments carry extra weight in this window. A pay error can end up in front of a bargaining unit or the council within days, and a messy first close can push back the ACFR and the audit. Late vendor payments add another problem where state prompt-payment rules apply. The plan below is organized the way finance teams live it: first 30 days post go-live, days 31 to 60 post go-live, and days 61 to 90 post go-live.

Key takeaways

Treat the first 90 days as a planned phase with its own owner, budget and exit criteria, and end hypercare when the evidence supports it.

Protect the money first: payroll, vendor payments and bank reconciliation get daily attention in the first 30 days post go-live.

Separate defects from training gaps and enhancement requests. A published severity scheme keeps the fix team on what blocks work.

Expect some problems to surface late, when year-end, budget-year pay changes or other first-time processes run for real.

Keep a vendor punch list tied to acceptance and warranty terms, and don’t sign final acceptance with open critical items.

Staff, unions, vendors and council should hear about problems from you first.

What hypercare means after a government ERP go-live

Hypercare is the period of heightened, hands-on support right after go-live, when the project team, the vendor and your super users stay on call to fix issues fast and confirm that critical processes produce correct results. After hypercare, the system moves to normal steady-state support.

How long it lasts varies. Panorama Consulting’s 2026 hypercare checklist says most organizations run hypercare “for several weeks,” and recommends defining exit criteria before go-live using measures such as transaction accuracy, support volume and backlog reduction. Vendor contracts often fund a shorter window: the City of Arlington, Texas, noted in its FY2022 budget that Workday went live August 2, 2021, with post go-live support from Accenture through September.

A government finance calendar doesn’t fit neatly into a few weeks. Your first month-end close, several payroll cycles, a quarterly report and possibly a state filing all fall inside the first 90 days. In our experience, that is the real stabilization window, even if formal vendor hypercare ends earlier.

Some states already expect this planning. Washington’s technology agency, WaTech, lists “Define Post-Go-Live Support Plan” as a project step, asking projects to finalize a post-go-live support and stabilization plan and offering templates for an operational support plan, communications plan and change management plan. Requirements vary by state, so check what your own state IT office or oversight body expects.

Set up the hypercare command center before day one

A command center is the single place where issues are reported, triaged, assigned and tracked during hypercare. Stand it up before cutover so the first ticket on go-live morning has somewhere to go.

Who sits in the command center

Keep the core group small and named. Each person needs authority to make decisions in their area.

Hypercare lead (your side), who owns the daily standup, the issue log and escalation. Often the project manager or finance systems lead.

Functional leads for payroll, accounts payable, general ledger, purchasing, utility billing or HR, depending on what went live.

IT staff who own interfaces, batch jobs, bank files and single sign-on.

The vendor or integrator lead, with a written commitment on response times during hypercare.

Super users in each department, the first stop for “how do I” questions.

An executive sponsor who can be reached the same day for decisions that cross departments or affect pay.

The daily standup

Run a 15 to 30 minute standup every working day in the first 30 days post go-live, then step down to two or three times a week once volume drops (a KCT rule of thumb, not a standard). A simple agenda keeps it short:

1

New critical and high issues since yesterday, with owner and next step.

2

Status of today’s payroll, check run, ACH or bank file and interfaces.

3

Reconciliation status: cash, payroll clearing, AP clearing.

4

Issues waiting on the vendor, with promised dates.

5

Decisions needed from the sponsor.

6

What staff and departments need to hear today.

Severity definitions

Agree on severity levels before go-live and publish them, so a department head can’t escalate a report formatting request to “critical.” The targets below are KCT rules of thumb for hypercare. Your vendor contract may define its own levels, and those govern what the vendor owes you.

Swipe to see all columns →

SeverityDefinitionExamplesResponse target (rule of thumb)Resolution or workaround target (rule of thumb)
1 CriticalStops pay, payments, cash or a legal deadline, with no workaroundPayroll won’t calculate, bank file rejected, positive pay file missingWithin 1 hour, any timeWorkaround same day
2 HighMajor process impaired, workaround exists but is costly or riskyWrong overtime rule for one bargaining unit, approvals stuck in workflowWithin 4 business hours2 to 3 business days
3 MediumProcess works with a reasonable workaroundReport missing a field, slow screen, minor posting errorNext business dayWithin the hypercare window
4 Low or enhancementUsability request or new requirementNew dashboard, extra report, layout changeLogged weeklyGoes to the post-hypercare backlog

Every ticket should also be tagged by type: defect, configuration, data, security, training or enhancement. The type tells you who fixes it. When one department keeps logging “training” tickets, schedule a refresher for that department before anyone touches configuration.

First 30 days post go-live: protect pay, payments and cash

The first 30 days post go-live are about money leaving and entering the organization correctly. Payroll, vendor payments and cash reconciliation get checked every cycle until they run clean on their own.

Run every early payroll like a test

Payroll is where post-go-live problems become public fastest. In Nashville, Metro moved to a new Oracle Cloud payroll and HR system on January 1, 2026. By late March, WSMV reported that about 2% of Metro employees had problems, Metro had issued more than 4,200 interim paychecks, and the council held a special joint committee meeting. Metro’s finance director said some issues only appeared once live data was in use, and warned that more could surface when the system handled tasks it hadn’t run yet, such as end-of-budget-year salary increases.

Testing covers what you can simulate. Live payroll finds the rest. For at least the first two or three pay cycles (a KCT rule of thumb), build in checks you would normally skip:

Before release, compare gross pay, hours, deductions and net pay by department against the last legacy payroll. Investigate any employee whose net pay moved more than a set tolerance.

Spot-check overtime, shift differentials, step increases and leave accruals for each unit, not just the largest.

Verify retirement, union dues, garnishments and tax withholding against the legacy system and the receiving parties.

Confirm file totals match the register before release, and confirm routing changes were captured.

Have a tested way to cut a manual or off-cycle check the same day. You will need it.

Keep a list of “first-time events” coming up in the next six months: retro pay, a cost-of-living adjustment, a new fiscal year, W-2 processing, calendar-year tax changes. Schedule a dry run for each one before it happens live.

Keep vendors paid

Accounts payable often slows down after go-live while staff learn new approval workflows. Clear communication with vendors, plus a daily look at invoice aging, keeps the slowdown from turning into late-payment complaints or penalties under your state’s prompt-payment rules.

In the first 30 days post go-live:

Run an aging report daily and track invoices stuck in approval queues by approver.

Confirm converted open purchase orders and encumbrances match the legacy balances before staff start receiving against them.

Verify vendor master data, especially remit-to addresses and bank details for ACH. Lock down changes to vendor banking information with a call-back verification step.

Reconcile each check run and ACH file to the bank before the next one.

Tell your largest and most sensitive vendors (utilities, nonprofits on reimbursement grants, small local businesses) what to expect and whom to call.

Reconcile cash daily

Bank reconciliation is where conversion errors hide. Reconcile cash daily in the first 30 days post go-live, even if your normal practice is monthly, and reconcile payroll and AP clearing accounts at least weekly.

Accomack County, Virginia, shows how long a reconciliation tail can run during a transition. At its December 12, 2024 audit committee meeting, staff reported that most bank accounts were reconciled through November except accounts payable and schools payroll, with entries still needed back to September. The schools planned to reconcile prior months in the old system, then carry the balance into the new system and reconcile from there, which staff said could take a couple of months.

The practical point: decide before go-live which system owns the reconciliation for the cutover month, and finish the legacy reconciliations first. Running two half-finished reconciliations in two systems is how differences become permanent.

Triage tickets by pattern

In our experience, ticket counts peak early, and most of the early tickets are “how do I” questions that super users can answer on the spot. Review the log every day for patterns: the same error in three departments is usually one configuration or security problem with three symptoms.

Track an enhancement “parking lot” separately. Requests for new reports and screen changes are legitimate, but working on them during the first 30 days pulls people away from defects that affect pay and payments.

Days 31 to 60 post go-live: close the books and fix what’s broken

Days 31 to 60 post go-live shift from survival to accuracy. The first month-end close happens, security and reporting gaps show up, and staff start building workarounds if nobody addresses those gaps.

Get through the first month-end close

In our experience, the first close on a new ERP takes longer than closes did on the old system. Plan for it: set a realistic close calendar, reconcile subledgers to the general ledger, check that fund balances and interfund activity carried over correctly, and review encumbrances and budget-to-actual reports with department heads before you publish them.

Write down every manual step, adjusting entry and workaround used during that first close. That list becomes your punch list for month two. For more depth on running the close itself, a companion KCT article covers the first financial close on a new ERP. You can also estimate where close time goes with KCT’s free close time calculator.

Clean up security roles

Security roles are usually configured generously at go-live so nobody gets locked out. By days 31 to 60 post go-live, tighten them before your auditors look.

Pull a list of users with administrator or broad finance access and confirm each one is still justified.

Remove temporary access granted during cutover, including vendor and consultant accounts that no longer need it.

Review segregation of duties: nobody should be able to create a vendor, enter an invoice and release payment alone.

Confirm workflow approvers match your purchasing policy and signature authority thresholds.

Document the role design and approval process. Your external auditor will ask for it.

Close the reporting gaps

Report gaps become visible once departments need their first monthly numbers. Inventory every report the legacy system produced and every report people are now asking for. Sort them into three groups: delivered and correct, delivered but wrong, and missing.

Prioritize reports with external audiences first: grant reimbursements, state reporting, debt disclosures, council budget reports and anything feeding the ACFR. Internal convenience reports can wait for the backlog.

Find the shadow spreadsheets

Shadow spreadsheets are side files staff build to track what they don’t trust or can’t find in the new system. A few are harmless. Many of them mean the system isn’t adopted, even if ticket counts look fine.

Prosci’s 2026 post-go-live guidance makes this point directly: fewer support tickets may simply mean employees have stopped asking for help and built their own workarounds. Prosci suggests reviewing each workaround and asking whether the system needs fixing or the habit needs changing. Ask each department to list its side spreadsheets in days 31 to 60 post go-live, then turn each one into a ticket, a training topic or an accepted local practice.

Run training refreshers

Training before go-live is abstract. By month two, staff know which screens confuse them. Use the ticket log to target refreshers: short sessions on the handful of processes generating the most “how do I” tickets, run by super users, recorded for new hires. Our article on managing staff resistance to new ERP software covers the change management side in more detail.

Minnesota State, the state’s college and university system, illustrates why data and process training matter after go-live. The Office of the Legislative Auditor’s February 2026 faculty payroll audit found that the system inaccurately compensated some faculty and did not always pay them on time. The Register reported that auditors found 19 of 202 sampled faculty paid incorrectly, plus 38 more in additional testing, and that the issues predated Workday but increased after it was introduced. Minnesota State’s stated response included standardizing practices and expanding training and validation.

Many agencies don’t have the staff to run a command center, a first close and a security review at the same time as their regular jobs. KCT’s ERP implementation team stays through the first fiscal close and payroll cycles, and for as long as you want after that. In Oxnard, California, KCT stayed on as the city’s production support partner and stood up a tiered L0 to L2 support desk within weeks.

Days 61 to 90 post go-live: hold the vendor to the contract and plan the handoff

Days 61 to 90 post go-live are about finishing. Close out the vendor punch list, measure stability against agreed exit criteria, and move the system to steady-state support with clear ownership.

Work the vendor punch list

The punch list is every open defect, missing deliverable and unmet requirement the vendor owes you under the contract. Build it from the ticket log, the first-close issue list and the report gap inventory.

For each item, record the contract reference (statement of work deliverable, requirement number or acceptance criterion), the date reported, the vendor’s committed fix date and the status. Then use your contract:

Don’t sign final system acceptance while severity 1 or 2 items remain open. If your contract ties payment milestones to acceptance, retainage or holdback gives you a reason for the vendor to finish. Use it calmly and in writing.

Check when the warranty period starts (often at go-live or acceptance) and what it covers. Log defects inside that period with dates so they can’t be reclassified later as paid change orders.

If the vendor says a fix is out of scope, ask for the contract language that supports that position before you approve any change order.

Use the contract’s escalation path. Weekly punch list reviews with the vendor’s project lead, and a monthly review with their account executive if items slip, keep pressure steady without turning every issue into a dispute.

Measure stabilization against exit criteria

End hypercare when the numbers show the system is stable. The vendor’s support hours running out is a contract question, and a separate one. Agree on the measures before go-live, report them weekly, and exit when they hold for several consecutive weeks.

The targets below are KCT rules of thumb for a mid-sized government, not published benchmarks. Adjust them to your size, modules and risk tolerance.

Swipe to see all columns →

Stabilization metricWhat it tells youExit target (rule of thumb)
Open severity 1 issuesWhether anything still blocks pay, payments or cashZero for at least 2 consecutive weeks
Open severity 2 issuesWhether major processes run without costly workaroundsZero, or each has an agreed fix date and workaround
Payroll accuracyShare of paychecks needing correction or off-cycle checksTwo or three consecutive cycles with only isolated, explained corrections
AP processing timeDays from invoice receipt to paymentBack to pre-go-live level or your prompt-payment target
Bank reconciliation statusWhether cash is under controlAll accounts reconciled through the last month-end, no unexplained differences
First month-end closeWhether the books can be closed in the new systemCompleted, with manual adjustments documented
Interface and batch job successWhether integrations run without interventionFailures are rare and each has a documented fix
New ticket volumeWhether users are settling inTrending down week over week, mostly training or enhancement
Shadow spreadsheetsWhether staff trust the systemInventoried, each with a decision
Critical reportsWhether you can meet external reporting dutiesGrant, state and council reports validated

If you can’t meet these by day 90, extend hypercare for the specific areas that missed. Extending payroll hypercare while finance moves to steady state is often a sensible split.

Transition to steady-state support

Steady-state support is the operating model you live with after hypercare: who answers tickets, who owns configuration, how vendor releases are tested and how issues reach the vendor. Write it down before the command center closes.

The handoff document should cover:

Tiered support: who handles first-line questions (super users or a help desk), who handles configuration (your ERP team or a support partner), and when tickets go to the vendor.

Service levels by severity, carried over from hypercare and adjusted.

Release and patch management: how vendor updates are tested and scheduled around payroll and close.

Monitoring of interfaces, batch jobs and month-end tasks.

Ownership of the enhancement backlog and how requests are prioritized.

Runbooks for payroll, close, year-end and other recurring processes.

Size the steady-state support team from your actual ticket data. KCT’s support queue cost calculator can help you estimate what an unmanaged queue costs in staff time.

How to communicate with council, staff and vendors during hypercare

Communicate on a schedule, before people ask. In our experience, staff, unions, vendors and elected officials are far more patient with problems they were warned about.

San Francisco Unified is a recent example. After the district launched a new payroll system on July 1, 2025, KQED reported that some employees were unpaid, paid at the wrong rate or missing union dues deductions, and the teachers union filed a state labor complaint. The district responded with a website where employees could find information and report payroll concerns. Have that kind of channel live before go-live, with named staff answering it.

In Nashville, a council member’s sharpest complaint was that affected employees couldn’t reach anyone by phone or email, according to WSMV’s coverage. Staff the phone line and inbox along with the fix team.

A simple communication plan for the first 90 days:

Swipe to see all columns →

AudienceWhat they needChannelFrequency (suggested)
Council or boardStatus, risks, what you are doing, when it will be fixedShort written update, plus a verbal update at a regular meetingEvery 2 to 4 weeks, immediately for pay or payment failures
Executive teamMetrics, decisions needed, vendor performanceDashboard and standup summaryWeekly
All employeesHow to check their pay, how to report a problem, who to callEmail, intranet page, union stewardsBefore each early payroll and when issues arise
Bargaining unitsKnown issues affecting their members and correction timingMeeting with union leadershipEvery 2 weeks in the first 60 days
Department heads and super usersKnown issues, workarounds, training datesWeekly bulletinWeekly
VendorsAny payment timing changes and an AP contactLetter or email, AP portal noticeAt go-live and if delays occur

When something goes wrong, tell council what happened, how many people are affected, what you have done for them, and when the next update will come. Keep the numbers precise and don’t promise a fix date the vendor hasn’t committed to in writing.

90-day post go-live checklist

Use this checklist as a starting template. Owners will differ by agency size.

Swipe to see all columns →

TaskOwnerTiming
Command center staffed, severity levels publishedHypercare leadBefore go-live
Daily standups and issue log runningHypercare leadFirst 30 days post go-live
Pre-release payroll register review each cyclePayroll leadFirst 2 to 3 cycles, then as needed
Daily AP aging and check run reconciliationAP leadFirst 30 days post go-live
Daily bank reconciliation, legacy recs finishedTreasury or accountingFirst 30 days post go-live
Vendor and employee communication sentCommunications or HRBefore go-live and ongoing
First month-end close completed and documentedControllerDays 31 to 60 post go-live
Security role and segregation of duties reviewIT security and controllerDays 31 to 60 post go-live
Report gap inventory and prioritizationFinance systems leadDays 31 to 60 post go-live
Shadow spreadsheet inventoryDepartment headsDays 31 to 60 post go-live
Targeted training refreshersSuper users and training leadDays 31 to 60 post go-live
Vendor punch list tied to contract termsProject manager and procurementDays 61 to 90 post go-live
Stabilization metrics reviewed against exit criteriaExecutive sponsorWeekly, decision by day 90
Steady-state support model and runbooks signed offIT and finance leadershipDays 61 to 90 post go-live

FAQ

What is hypercare in an ERP implementation?

Hypercare is the period right after go-live when the project team, vendor and super users provide intensive, fast-response support. Its goal is to fix defects quickly and confirm that payroll, payments, reconciliations and reporting work correctly before the system moves to normal support. Exit should depend on agreed metrics such as open critical issues and payroll accuracy.

How long should hypercare last after an ERP go-live?

Hypercare length depends on scope and complexity. Panorama Consulting says most organizations run it for several weeks. For governments, KCT plans support around the first 90 days because the first month-end close, several payroll cycles and early reporting deadlines all fall inside that window. End hypercare when exit criteria are met.

What are hypercare exit criteria?

Hypercare exit criteria are the measurable conditions that show the system is stable enough for steady-state support. Typical criteria include no open critical issues, accurate consecutive payrolls, current bank reconciliations, a completed month-end close, reliable interfaces, declining ticket volume and validated external reports. Agree on them before go-live.

Why do ERP problems appear after go-live if testing passed?

ERP problems appear after go-live because testing can’t reproduce every live condition. Real employee data, full transaction volumes and first-time events such as retro pay, budget-year salary changes or year-end processing expose gaps. Nashville’s finance director said some issues only surfaced with live data, which is why hypercare checks every early cycle.

Who should own post go-live support in a city or county?

Post go-live support should be owned by your organization, with clear tiers. Super users or a help desk answer first-line questions, an internal ERP team or support partner handles configuration and releases, and the vendor fixes product defects under the contract. Name one accountable leader, usually in finance or IT, before hypercare ends.

When should we hold back vendor payments after go-live?

Vendor payment holdbacks should follow your contract’s acceptance and retainage terms. If critical or high-severity defects remain open, don’t sign final acceptance or release payments tied to it. Document each open item against the specific deliverable or requirement, and involve procurement and legal counsel before withholding anything.

Plan the first 90 days after ERP go-live before cutover

Most of the decisions that make hypercare go smoothly can be made months before cutover: who runs the command center, which severity levels apply, what counts as stable, and which punch list items the vendor must close before final acceptance. Write them into the cutover plan, and brief council on them too. If you are planning a go-live, or you are already in the middle of a rough first quarter, KCT can help through our implementation team or ongoing managed services and support, with senior people who stay engaged through year one and beyond. Book a conversation with KCT to map out your first 90 days.

Ready when you are

Let's talk before it slips.