Skip to content
Tallyfy

Set up automation rules in plain language

Describe the rule - your AI builds the automation

Tallyfy automations use if-then logic to route tasks, skip steps, change assignments, and trigger actions based on data in your processes. Instead of configuring these rules through the automation builder, just describe what you want in plain language. Your AI translates the business logic and creates the rules directly in your template.

No menu hunting. No figuring out which condition type maps to which field. Just say what should happen and when.

What you’ll accomplish

  • Create automation rules by describing business logic in plain language
  • Set up conditional step visibility, assignment changes, and deadline adjustments
  • Skip the automation builder entirely

Try it now

Setup: If you haven’t connected Claude to Tallyfy yet, follow the connect your AI to Tallyfy guide first - it takes about two minutes.

Prompt to try:

On the Sales Pipeline template: if the deal value field is under $10,000, skip the Legal Review step. If the deal value is over $50,000, add the VP of Sales as an approver on the Final Approval step.

What happens: Claude calls get_template and get_template_steps to read the template structure and find the exact step IDs and form field IDs it needs. Then it calls analyze_template_automations to check whether any existing rules might conflict with what you’ve described. Finally, it calls create_automation_rule twice - once for each condition you described - translating “skip the Legal Review step” into a hide-step automation and “add the VP of Sales as an approver” into an assignment automation. Claude confirms what was created so you can verify before testing.

How it works behind the scenes

Your AI works through a predictable sequence every time you describe an automation rule. Understanding the steps helps you write better prompts.

Step 1: Read the template structure

Before creating anything, the AI calls get_template and get_template_steps to read the full template. It’s looking for step IDs, form field IDs, existing assignments, and the field types that matter for your conditions. A field called “deal value” might be a number field, a currency field, or a text field - that distinction changes how the condition is written.

Step 2: Check for conflicts

The AI calls analyze_template_automations to scan existing rules. If a rule already hides the Legal Review step under a different condition, the AI flags it before creating a potentially contradictory rule. This is where describing your intent matters - “skip Legal Review for small deals” tells the AI something different than “always show Legal Review for enterprise customers,” even if both involve the same step.

Step 3: Create each rule

The AI calls create_automation_rule for each condition you described. Each rule has two parts:

  • Trigger condition: which field value or event starts the rule (a dropdown selection, a number threshold, a date comparison, a checkbox state)
  • Action: what happens when the condition is true (hide a step, show a step, change an assignee, adjust a deadline)

The AI maps your natural language to Tallyfy’s automation types:

  • “skip step” / “hide step” → hide step automation
  • “show step” / “add step” → show step automation
  • “assign to” / “add approver” → assignment automation
  • “set deadline” → deadline automation

Step 4: Confirm what was created

Once the rules are created, the AI summarizes them in plain language so you can verify the logic before testing. Something like: “I created two rules. Rule 1 hides the Legal Review step when deal value is less than $10,000. Rule 2 adds VP of Sales as an approver on Final Approval when deal value is greater than $50,000.”

That confirmation is worth reading carefully - it’s the fastest way to catch a misread condition before it affects live processes.

Common automation patterns

Different business scenarios map to different automation types in Tallyfy. Here’s what each looks like in plain language.

Conditional step visibility

“If [field value matches condition], show/hide [step name]”

This is the most common pattern. Steps that don’t apply get hidden; steps needed in special cases get revealed. Real examples:

  • “If the department is HR, show the HRIS Access Setup step”
  • “If the project type is Internal, skip the Client Sign-off step”
  • “If the applicant answered No to the background check consent question, hide the Background Check step”

The condition can be a dropdown selection, a checkbox, a number comparison, a date, or even whether a particular field was filled in at all.

Dynamic assignment

“If [condition], assign [step] to [person or group]”

Route tasks to different people based on the data in your process. Useful when the same step should go to different approvers depending on deal size, department, geography, or any other field.

  • “If the region is EMEA, assign the Compliance Review step to the EU Legal team”
  • “If the budget exceeds $100,000, assign Final Approval to the CFO instead of the department head”
  • “If the ticket type is Security, add the Security team as an additional assignee”

Note that assignment automations add or change assignees - they don’t remove existing ones unless you specify that explicitly.

Deadline adjustments

“If [condition], set deadline to [time period]”

Tighten or extend deadlines based on priority, urgency, or other criteria. The time period can be hours, business days, or calendar days.

  • “If priority is Critical, set all step deadlines to 4 hours”
  • “If the renewal date is within 30 days, set the Contract Review deadline to 1 business day”
  • “If the deal size is over $500,000, extend the Due Diligence deadline to 10 business days”

Multi-condition rules

“If [condition A] AND [condition B], then [action]”

Combine conditions when a single field value isn’t enough to trigger the right behavior. Both conditions must be true for the action to fire.

  • “If department is Sales AND deal size is over $50,000, add the Revenue Operations team to the Forecast Review step”
  • “If vendor type is New AND contract value is over $10,000, show the Extended Vetting step”

The AI handles AND logic naturally when you describe it that way. If you need OR logic - where either condition alone should trigger the action - say “or” explicitly and the AI will create separate rules or ask how you want it structured.

Tips for better results

Reference exact step names and field names from your template. The AI reads your template to find matches, but “the approval step” is harder to resolve than “the Final Approval step.” Field names matter too - “the amount field” and “the deal value field” might be different fields in your template.

Be explicit about thresholds. “Greater than $50,000” and “at least $50,000” are different conditions. “Less than” and “up to” have edge cases at the boundary. State the exact threshold and whether the boundary value is included.

Ask the AI to check for conflicts. Before creating new rules on a template that already has automations, say: “Check for any existing rules that might contradict this before creating it.” The AI calls analyze_template_automations and flags anything worth reviewing.

Describe one logical group at a time. If you have five unrelated rules, describe them in one prompt and let the AI work through them. But if you’re setting up a complex routing system with overlapping conditions, take it one rule at a time and verify each one before adding the next.

Test the logic before going live. After creating rules, ask: “What would happen if someone selected Engineering in the department field and filled in a start date 3 days from now?” The AI walks through which rules would fire and what the resulting step visibility and assignments would look like - without actually launching a process.

Specify the field value format your template uses. If your priority dropdown uses “High / Medium / Low” but you say “urgent” in your prompt, the AI will either ask for clarification or try to map it. Save time by matching the exact option text from your form field.

Handle edge cases explicitly. “If the deal value is over $50,000, add the VP of Sales as an approver” - but what happens for deals under $50,000? If you want the step assignment to stay as-is, that’s fine. If you want a different assignee for smaller deals, say so. The AI won’t assume what the else condition should be.

Take it further

Automations

Set up if-this-then-that rules that trigger actions automatically in your Tallyfy processes.

Automation examples

Real-world examples of automation rules you can use in your Tallyfy templates.

BYO AI

Learn about the bring-your-own-AI integration and what it enables across your Tallyfy workflows.
Get in touch
About Tallyfy