How to Use the AI Decision Maker in Sparkforge Workflows

By Sparkforge Support Team ·Last reviewed May 3, 2026

The AI Decision Maker action routes contacts down workflow branches based on plain-English instructions you write. It replaces brittle if/then trees with AI evaluation of contact data.

The AI Decision Maker is a workflow action that routes contacts down branches based on plain-English instructions you write — no nested if/then chains, no manual filter stacks. The AI reads the contact's data, your instructions, and any context you provide, then picks the matching branch.

Note: AI Decision Maker is a premium action. Each execution incurs an additional charge.

When to reach for it

  • Routing logic that depends on multiple fields, tags, or fuzzy criteria ("large company OR high engagement").
  • Workflows where the right branch depends on context the AI can interpret — like classifying spam form submissions.
  • Decision trees you keep editing because the rules keep changing.

Set it up

Step 1 — Open or create a workflow

  1. Log in to your sub-account.
  2. Click Automation in the left menu.
  3. Edit an existing workflow or click + Create Workflow.

Step 2 — Add the AI Decision Maker action

  1. Confirm the workflow has a trigger (Contact Changed, Form Submitted, etc.).
  2. Click the + icon in the builder.
  3. Scroll to AI Actions and select AI Decision Maker.
  4. Set a clear Action Name.

Step 3 — Write the instructions

In the Instructions field, describe how contacts should be routed. Use the variable picker (tag icon) to insert dynamic fields:

  • {{contact.engagement_score}}
  • {{contact.company_size}}
  • {{contact.tags}}

Example:

Look at the contact's engagement score and company size.
If engagement score is over 60 OR they work at a large company, send them to sales.
If engagement score is between 30-60 and the company is small or medium, send them to email nurturing.
If engagement score is under 30, send them to a re-engagement campaign.

Step 4 — Add context

Use the Additional Context field to give the AI business rules and definitions. Example:

We sell software to companies with 10-500 employees.
Small = 1-50, Medium = 51-200, Large = 200+.
Engagement scores range 0-100.
Sales can handle ~20 new leads per week.

Step 5 — Default branch

The Default Branch is locked in and cannot be deleted or renamed. It's the fallback for contacts that don't match any custom branch, so workflows never get stuck.

Step 6 — Custom branches

Click + New Branch for each route you want. Each branch needs:

  • A clear Branch Name (for example, High Priority – Sales).
  • A short description of when contacts should land there.

Step 7 — Save

Click Save Action. The workflow now evaluates contacts and routes them based on AI-driven decision logic.

Use case: spam form filter

Send form submissions through the AI Decision Maker. Instruct the AI to classify each submission as spam or authentic, and provide a few example spam patterns. Only legitimate submissions get routed to the "Create Opportunity" branch — the rest go to a quarantine tag for later review.

Tips for accuracy

  • Always include the contact data the AI needs to decide — fields, tags, prior step outputs.
  • Be explicit about thresholds ("over 60", "between 30-60") rather than vague terms.
  • Use Execution Logs to confirm where each contact landed and refine instructions if routing is off.

Frequently Asked Questions

How does the AI choose a branch?
It reads your instructions and the contact data you reference, then picks the branch that best matches. Be explicit about thresholds and field names.
What if a contact doesn't match any rule?
They go to the Default Branch. It's permanent and ensures no contact gets stuck mid-workflow.
Can I edit the Default Branch?
No. The Default Branch can't be deleted or renamed. You can change what actions follow it inside the workflow.
How do I see where a contact was routed?
Open the Execution Logs tab inside the workflow. It shows the chosen branch for each contact run.
Can I use form or survey data?
Yes. Insert variables from prior workflow steps using the variable picker, and the AI will factor them into the decision.

Was this article helpful?