In this article

The analyst has a good reason to upgrade. She needs to schedule a report her team already uses. But she cannot approve the subscription change, and the person who can approve it has never seen the workflow. A button that says “Contact your administrator” leaves the analyst to reconstruct the case somewhere else.

Build the expansion request around that handoff. Preserve the work, package the decision for the authorized buyer and return the user to the right place once the decision takes effect. The request is successful when it produces an understandable resolution, including a useful refusal when the account should not upgrade.

Check whether the account needs to buy anything

Before sending a budget request, determine what blocks the user. The account may already own the capability but need to assign an unused seat. The user may have the right subscription and lack permission on a particular object. Or the account may genuinely need a new commercial commitment.

These routes deserve different actions. Asking a budget owner to approve spending when an available seat would solve the problem creates unnecessary work. Offering a subscription upgrade when the issue is a restricted file creates a misleading promise.

Figma’s seat approval settings illustrate this distinction. Administrators can require manual approval, automatically assign an available prepaid seat, or allow automatic seat purchases. The choice depends on the organization’s policy. Your product should make its own authority and availability rules explicit before deciding where to send a request.

Capture the reason while the work is visible

For an illustrative reporting product, the analyst reaches the limit while configuring a recurring report. Preserve its filters, recipients and intended schedule. Show what is missing and let her request the specific capability from someone authorized to decide.

The request should contain enough context that the approver does not need to ask the analyst to start again. A useful handoff includes:

  • The requesting user, workspace and affected team.
  • The capability or capacity requested, with the current limit.
  • The existing work the upgrade would support.
  • The commercial scope: one seat, several seats or the whole account.
  • The applicable price or a clear indication that a quote is required.
  • The request’s owner, current status and validity period.

Let the user add a short explanation, but populate what the product already knows. “Schedule the weekly inventory report for the operations team” carries more decision value than an empty message box. Do not invent a business return from usage data. The request can show the work without promising how much money it will save.

Give the approver the decision they actually control

The budget owner needs to understand the scope and consequence of approval. Show the existing commitment, the requested change and when the change would take effect. If a price depends on contract terms or a negotiated quote, route the request into that process rather than showing a deceptively precise self-serve amount.

Distinguish approval of the need from authorization to purchase. A manager may endorse the workflow while procurement must approve the commitment. Your system should represent that extra step instead of treating a manager’s click as permission to charge the account.

Use the account’s current authority rules at decision time. The person who could approve a request when it was created may no longer hold that role. Forwarded links should not confer approval power by themselves. A recipient can be allowed to inspect context and route the request without being permitted to accept the commercial change.

Keep sensitive workspace content out of broadly forwarded notifications. The approver usually needs the report’s purpose and ownership, not every row of its underlying data. A signed-in view can reveal details according to existing permissions.

Make request state visible to both people

A proposed state sequence is requested, awaiting authorized decision, approved, provisioned and used. Declined, withdrawn and expired are explicit exits. Treat these as separate states because each answers a different question.

StateWhat the product can truthfully tell the user
RequestedThe request has been recorded and routed
ApprovedAn authorized decision has been made
ProvisionedThe requested access or capacity is available
UsedThe user completed the intended product action
Declined or expiredThe request is closed, with a reason or next step

Sending an email proves neither receipt by the right person nor a purchase. Approval can precede provisioning, especially when contracts or administrative steps remain. Keep those gaps visible rather than showing a cheerful success screen while the user is still blocked.

Prevent duplicate requests from turning one need into several purchases. Link repeated attempts to the open request when the scope is unchanged. If the user requests a different quantity or capability, make the difference explicit and require the appropriate review.

Return the user to the unfinished task

When access becomes available, notify the requester and provide a route back to the preserved work. Confirm that its settings are still valid. A report saved weeks earlier may reference a removed destination or a schedule that no longer makes sense.

The final permission check still matters. Figma’s seat-request documentation distinguishes product access from permission to edit an individual file; a seat upgrade does not grant editing rights everywhere. In the reporting example, an approved premium capability should similarly leave ordinary data and recipient permissions intact.

If the request is declined, preserve whatever work remains permitted and show the reason the approver chose to share. A smaller scope, an existing seat or a manual workflow may be a sensible alternative. Avoid repeatedly prompting the user to submit the same rejected request without new information.

Review the handoff, not just the button

Measure requests at the account and request level, with timestamps for each transition. Look for requests with no authorized recipient, approvals waiting on provisioning and newly provisioned users who never return to their work. Each group points to a different failure.

An approval rate alone can reward poor targeting: a product might generate many weak requests and create more administrative burden. Read it beside duplicate requests, time awaiting a decision and successful use after access is granted. Do not attribute later expansion to the request flow without an appropriate comparison.

For the earlier question of when to present the opportunity, see evaluating upgrade prompts. For larger deals that need a human commercial process, see product-led sales handoffs. The product’s immediate responsibility is simpler: carry a specific need to the right decision-maker, preserve the user’s work and make the outcome clear.

Sources

Something we should clarify?

Suggest a correction