How to Cut Over

Cutting over to an orchestrated workflow is the plan most teams spend six months building. Decommissioning the manual chain it replaces is the plan most teams spend six hours on. That imbalance is where the cutover fails. The Slack DMs, the spreadsheet handoffs, and the manual CRM entry sequences are not passive infrastructure waiting to be switched off. Switch them off without a protocol and leads mid-path through the manual chain fall out of the process entirely, attribution baselines reset, and SDRs miss handoffs that look fine in the new system but never actually arrived. This post is a concrete 1-day playbook for the day you cut over to an orchestrated workflow: confirm readiness, handle in-flight leads, and retire the manual touchpoints without dropping pipeline. For the orchestration architecture this builds on, start with the workflow orchestration pillar.

Why does cutting over to an orchestrated workflow break down so often?

Most teams do not fail the orchestration build. They fail the decommission. NinjaCat's 2026 AI Maturity in Marketing report (n=500) found only 8% of marketing teams run multi-step AI workflow orchestration, and 89% operate with fragmented tooling and no integration layer. Salesforce's 2026 State of Marketing report (n=4,450) found 87% of teams use AI in at least one workflow, but only 13% have deployed agentic AI. That 74-point gap is the manual chain: teams with AI tools still routed by Slack DMs, spreadsheets, and manual CRM entry. Most of them are not anti-orchestration. They are transitioning, and the transition day, the cutover, is where most of them break.

The break is usually not a technical failure. The orchestrated workflow runs correctly. The manual chain also keeps running, in parallel, because no one formally stopped it. Leads get processed twice, routed incorrectly, or abandoned in the gap between systems.

The three layers of the manual chain

The manual chain is three layers: a routing layer (the Slack DM where a rep flags a new lead for manager assignment), a tracking layer (the spreadsheet row or the ad-hoc CRM note), and a handoff layer (the pull-list the SDR works from). Each layer must be formally retired. Abandoning the tool is not the same as retiring the process.

The invisible routing problem

The most dangerous component is unofficial routing: the Slack channel where reps paste lead names for Monday-morning assignment. This is invisible to your workflow platform and the first thing your team reverts to when the new system produces an unexpected result. Ask your team if they use a manual routing step and most will say no. Watch a Monday morning for 90 minutes and you will find it.

How do you confirm the orchestrated workflow is ready before decommissioning?

Readiness and performance are different questions. Performance asks whether the new workflow outperforms the manual chain. Readiness asks whether it can process leads correctly at all. Confirm readiness first; performance comes after. Four gates must all be true: the trigger map is documented (every entry point, branch condition, and terminal action including error-path handling), the field contract is signed off by every downstream system (CRM, SDR tool, reporting layer), an end-to-end test using 10-20 representative historical lead profiles has passed, and the error queue has been reviewed with at least one failed-validation lead confirming the failure path works as designed.

The field contract before cutover

The field contract is an explicit list of every CRM field the orchestrated workflow writes to, the expected value range, and which downstream system reads that field. Eight fields are required for most B2B SaaS lead routing workflows: lead_source, assigned_rep, next_step_type, next_step_due_date, score_band, first_touch_campaign, workflow_entry_date, and handoff_status. If any of these fields does not populate reliably during the end-to-end test, you have a contract gap. Close it before scheduling a cutover date.

The end-to-end test pass rate

Run 20 historical leads through the orchestrated workflow in a staging environment and compare output fields against known outcomes. A 95% field-population rate means the workflow is ready. Below 90%, investigate the failures before setting a cutover date. Re-run the test after any configuration change to the workflow, CRM field schema, or SDR tool integration.

What do you do with leads already in the manual chain on cutover day?

In-flight leads are the most operationally sensitive part of a cutover. On the day you cut over, some leads are inside the manual chain at various stages. Each category requires a different handling decision. The triage has three tiers based on how far along the manual chain each lead is. Tier A covers leads that entered within the last 24 hours and have not been contacted: route them directly into the orchestrated workflow after tagging them with a manual_transfer flag in the CRM to preserve attribution. Tier B covers leads that have been contacted at least once but are still mid-process: keep them in the manual chain until their current path completes, then enter the orchestrated workflow as re-engaged leads. Tier C covers leads in the manual chain for more than seven days with no recent activity: close the manual path, log their last known stage, and let them re-enter as new leads if they respond to a re-engagement trigger.

The lead migration log

Every lead that moves from the manual chain to the orchestrated workflow on cutover day needs a CRM record entry documenting three things: date of transfer, manual chain stage at transfer, and workflow entry point assigned. When a transferred lead closes two months later, this log determines whether first-touch credit belongs to the pre-cutover manual chain or the post-cutover orchestrated workflow. Without it, your quarterly attribution report contains an unexplained credit gap.

Which leads never transfer

Some leads should not transfer: leads in a legal or compliance hold, leads flagged do-not-contact, and leads tied to an active opportunity managed outside the workflow. Create a no-transfer list before cutover day, confirm it with both marketing operations and the sales development manager, and document the reason in the CRM for each no-transfer lead.

How do you validate handoff fidelity in the first four hours?

The first four hours after a cutover are the highest-risk window. Assign one person to run a 30-minute monitoring cadence: check the field-population rate on every lead that has completed the first workflow step. Field-population rate is the percentage of those leads with all eight required CRM fields populated. Target 95% or higher. If the rate drops below 90% on any 30-minute check, stop new lead routing to the orchestrated workflow and investigate before continuing. A failure caught at the one-hour mark costs one hour. A failure discovered at the end of the day costs a full pipeline-building cycle. See the workflow versioning playbook for the same principle applied to v1/v2 parallel runs.

The five-field spot check

In addition to the full field-population rate, the validator runs a spot check on five fields most commonly affected by cutover-day integration errors: assigned_rep (a real rep name, not a placeholder), score_band (populated, not null), next_step_type (a valid value from the allowed list, not a raw API tag), first_touch_campaign (the correct campaign, not the orchestration platform's internal tracking label), and handoff_status (advanced past pending_assignment). These five fields cover the most common breakpoints between a new orchestrated workflow and the CRM and SDR sequence tools it depends on.

What triggers an immediate pause

Three conditions trigger an immediate pause of new lead routing. First, assigned_rep is null or contains an inactive rep name on more than 5% of leads. Second, first_touch_campaign is being overwritten with the orchestration platform's internal tracking tag on any lead. Third, handoff_status is not advancing past pending_assignment within two hours of entry. Each condition means an integration error that compounds with every new lead that enters. Pause, fix the root cause, then resume.

How do you structure the parallel period before full decommission?

The parallel period is the two-to-five business day window when the orchestrated workflow handles all new leads and the manual chain handles only Tier B in-flight leads completing their existing paths. This is not the same as a versioning parallel run, where both versions receive new leads simultaneously. In a decommission parallel period, only the orchestrated workflow receives new leads. The manual chain exists only to close out paths already underway. Entry requires four conditions: end-to-end test passed, field contract verified, in-flight lead triage complete, and the four-hour validation window showed a field-population rate above 95%. Exit requires two conditions: 48 consecutive hours above 95% field-population rate, and the Tier B in-flight lead count reaching zero.

Monitoring the Tier B count

Track the Tier B count daily. It should decline by natural attrition as those leads complete their manual-chain paths. If the count is not declining after three business days, investigate: reps may be holding leads in the manual chain as a hedge against the new system, or the manual-chain completion criteria may be unclear. The orchestration owner should drive the daily Tier B review and clarify completion criteria with each rep who has outstanding Tier B leads.

The parallel-period daily log

Every day of the parallel period, the orchestration owner writes one log entry: date, new-lead count processed by orchestrated workflow, Tier B count remaining, field-population rate (24-hour rolling), and any pause events with root cause and resolution time. This log is what you present to the sales development manager to confirm the manual chain can be fully decommissioned. Without it, the decommission decision is a trust exercise.

What are the two cutover-day failures that destroy attribution baselines?

Brynjolfsson, Rock, and Syverson's AEA research on AI and the productivity J-curve established that costs of an AI investment appear on the income statement immediately while pipeline gains accumulate in later-closing deals. A clean attribution baseline is what lets you track where on that J-curve you actually are. The Deloitte CMO Survey Spring 2026 (n=300) found 64% of CMOs cite demonstrating financial impact as their primary challenge. A bad cutover corrupts the baseline and makes that challenge impossible. Two cutover failures cause this: attribution-field reset and in-flight lead orphaning.

The attribution-field reset failure

An attribution-field reset happens when the orchestrated workflow uses different field names, stage labels, or a different campaign-tagging taxonomy than the manual chain. Leads that entered through Paid Social Campaign A get re-attributed to the orchestration platform's internal tracking tag. Your quarterly report shows a gap in Campaign A leads, and Campaign A looks underperforming. The campaign is not underperforming. Your field mapping is wrong. The fix is to define the field mapping before cutover day, confirm it uses existing CRM field values (not platform-internal labels), and verify it during the end-to-end test. An attribution-field reset caught before cutover is a configuration fix. One caught after cutover is a retroactive data-repair project.

The in-flight lead orphan failure

An orphaned lead is a Tier B lead abandoned before completing its manual-chain path. The manual chain was shut down, the lead was never transferred to the orchestrated workflow, and the rep who owned the manual-chain step had no fallback. The lead's CRM record shows the last manual-chain stage, then nothing. If this lead later closes, it will appear in your close report as a no-source deal. Prevent it by completing Tier B triage before the parallel period ends and running a no-activity plan on all CRM records created in the 14 days before cutover before declaring the decommission complete.

How do you formally retire the manual touchpoints?

Formal retirement creates a record that the component has been decommissioned, gives the team a redirect, and preserves historical data. Three components require specific actions. For Slack channels used for manual lead routing: archive the channel (not delete it), and post a final pinned message stating that lead routing has moved to the orchestrated workflow with a link to the process documentation. For manual tracking spreadsheets: lock the file for editing, add a dated header row stating what replaced it, and retain the file for 90 days. For manual CRM entry steps in your process runbook: remove the step, add a note explaining the replacement, and date the change. Run an AI system plan after the decommission to confirm the new workflow covers every use case the manual chain handled.

The decommission checklist

Before declaring the cutover complete, confirm six items: Tier B count is zero; 48 hours of parallel-period monitoring showed field-population rate above 95%; all Slack routing channels have been archived with a final pinned message; all manual tracking spreadsheets have been locked and dated; the process runbook has been updated to remove manual steps; the lead migration log is complete and stored where both marketing operations and sales development can access it.

The 14-day post-cutover plan

Two weeks after the cutover, run a CRM query for records with a last-modified date before the cutover and a current stage of new or assigned with no workflow activity. These are potential orphans that were missed during triage. Confirm the current status of each with the assigned rep and either route them into the orchestrated workflow manually or close the record as stale. Check your AI marketing maturity benchmark scores before and after the decommission to confirm that handoff fidelity has improved. A decommission that reduces Handoff Quality benchmark scores is a sign the orchestrated workflow is not yet replacing what the manual chain provided.

Methodology

This post draws on four verified sources. Salesforce State of Marketing 2026 (n=4,450) provided the 87% AI adoption and 13% agentic AI deployment figures. NinjaCat 2026 AI Maturity in Marketing (n=500) provided the 8% multi-step orchestration and 89% fragmented tooling figures. The Deloitte CMO Survey Spring 2026 (n=300) provided the 64% financial impact challenge figure. Brynjolfsson, Rock, and Syverson's AEA paper provided the J-curve framework. The three-tier in-flight lead triage, the eight-field contract, and the parallel-period log format are operational frameworks representing reproducible processes, not client results. Your specific field names, lead volumes, and parallel-period durations will differ based on your CRM configuration and orchestration platform. Run the cut over to an orchestrated workflow against each step in this playbook before treating any threshold as a fixed target.