Skip to content
The Workflow Library
Get the full library

Empty output fallback: branch on a null AI result to a manual review task instead of a blank property

One of 565 written HubSpot workflow specs, by a collective of independent HubSpot consultants.

Professional

Seat pricing: Professional edition. To build it, a GTM Seat or an Ops Seat (an Ops Seat cannot create marketing emails or sequences). Per-hub plans: Sales Hub Professional.

Contact15 stepsUpdated for HubSpot seat pricing
OutcomeWhen an AI action returns nothing, the record gets a visible status and a review task for a named person. No property is left blank without anyone knowing, and nothing downstream runs on a missing value.
Workflow mapai-agents-015
Set status pending leads to Summarize record. Summarize record leads to Output is known. Output is known leads to Yes: Write property, status filled, stamp date; No: Delay 60 minutes, retry once. Delay 60 minutes, retry once leads to Second output is known. Second output is known leads to Yes: Write property, status filled, stamp date; No: Status empty review, create review task. Yes No Yes No Set status pending Summarize record Output is known Delay 60 minutes, retry once Second output is known Status empty review, create review task Write property, status filled, stamp date
Set status pending leads to Summarize record. Summarize record leads to Output is known. Output is known leads to Yes: Write property, status filled, stamp date; No: Delay 60 minutes, retry once. Delay 60 minutes, retry once leads to Second output is known. Second output is known leads to Yes: Write property, status filled, stamp date; No: Status empty review, create review task. Yes No Yes No Set status pending Summarize record Output is known Delay 60 minutes, retry once Second output is known Status empty review, create review task Write property, status filled, stamp date

One of 565 written HubSpot workflow specs, by a collective of independent HubSpot consultants.

When to use it, when not to

Use it

  • Use it after every AI action in a workflow. The other specs in this category reuse this branch. It is shown here with "Summarize record", but the pattern is the same for "Data Agent: Custom prompt", "Data Agent: Research", "Run agent" and "Use a custom LLM".
  • Use it because an empty output is a normal case, not an edge case. When HubSpot Credits run out, the AI action fails and its outputs are null. A thin record can also produce nothing useful.
Read more on when to use itShow less

Skip it

  • Do not create one task per record on high-volume workflows. A review queue of 2,000 tasks is not reviewed. Use the default value variant and review the list in bulk.
  • Do not use the fallback to hide a broken prompt. If more than about one record in ten lands in review, fix the prompt or the enrollment criteria.
Read more on when to skip itShow less

The rest of this spec

Prerequisites · 10 items

Every property with its label, internal name, type and options, plus the assets the workflow needs.

Enrollment · 4 parts

The trigger, the criteria as ticked in the UI, the re-enrollment rule and the suppression list.

Steps · 15 numbered actions

Each action in order with its parameters, branches and delays.

Known pitfalls · 6

Traps seen in production, each with what breaks and what to do instead.

Test plan and UAT script · 5 cases, 3 variants

How to prove it works in a test portal, one case per branch, and a sign-off sheet.

Files · PDF, Markdown, JSON, CSV

The properties JSON creates the fields in one call. The workflow JSON documents the build.

Install · 1 click, switched off

The installer creates the properties and the workflow, switched off, then opens a task for what it cannot create. Nothing is deleted, enabled or enrolled.

Unlock all 565 specs, 149 EUR

One-time, excl. VAT. 30-day refund. Installs in 3 portals. Founding price until 30 November.

Read the free AI agents spec

Not the workflow you needed?

More in AI agents

25 specs in this section

See all 25 This category, 39 EUR
FreeCategorize inbound form submissions with Data Agent: Custom prompt, then route by categoryContact17 steps Log every AI output with a timestamp and model version to properties for QA and rollbackContact13 steps Credit guard: a daily counter property that pauses AI actions once the budget is spentContact8 steps Summarize the record into the SDR-to-AE handoff task so the AE opens the call with contextContact8 steps

Read 12 complete specs before you pay.

One per category, same format as the paid ones. Email only, no card.

We email the link, then a short note when specs change. Nothing else.

149 EUR199

Full library, 565 specs
Get the full library