5 AI automations businesses pay $500 to $5k for

You are an automation engineer building an abandoned cart recovery agent for an ecommerce store.

Context:
Store platform: [SHOPIFY / WOOCOMMERCE / OTHER]
Products: [WHAT THE STORE SELLS]
Current abandoned cart handling: [NONE / BASIC EMAIL / OTHER]

Task:
1. Design the trigger logic for detecting an abandoned cart (time since last activity, cart value threshold).
2. Draft a 3 message recovery sequence (timing, channel, and copy for each) referencing the specific abandoned item.
3. Recommend which platform APIs or tools are needed to build this (e.g. Shopify Admin API, Klaviyo, Twilio).

Output format:
- A trigger logic summary
- The 3 message sequence with timing and copy
- A tools and APIs checklist

You are an automation engineer building an AI customer service agent with real order access for [BUSINESS NAME], a [TYPE OF BUSINESS].

Context:
Order or support system currently used: [SHOPIFY / ZENDESK / CUSTOM / OTHER]
Most common customer questions: [LIST TOP 3 TO 5]

Task:
1. Define the agent's scope, which questions it should answer directly and which should escalate to a human.
2. Design the order lookup flow (what info it needs from the customer, what it pulls from the backend).
3. Draft the agent's system prompt, including tone and escalation rules.

Output format:
- Scope definition (agent handles versus escalates)
- Order lookup flow as text steps
- Full system prompt for the agent

You are an automation engineer building an intake automation for [BUSINESS TYPE, e.g. bookkeeping firm or insurance agency].

Context:
Current intake method: [PAPER FORM / EMAIL / PHONE CALL / PDF]
Destination system: [QUICKBOOKS / AGENCY MANAGEMENT SYSTEM / SPREADSHEET / OTHER]
Fields that need to be captured: [LIST KEY FIELDS]

Task:
1. Design the extraction step, how raw intake becomes structured data.
2. Map each extracted field to its destination in [DESTINATION SYSTEM].
3. Flag which fields need human review before they are finalized.

Output format:
- Extraction step design
- Field mapping table
- Human review checklist

You are an automation engineer building a review and testimonial pipeline for [BUSINESS NAME].

Context:
Where reviews currently come from: [NONE / GOOGLE / PLATFORM NATIVE REVIEWS / OTHER]
Preferred review destination: [GOOGLE / TRUSTPILOT / ON SITE]

Task:
1. Design the request timing (how many days after purchase or service, how many follow ups).
2. Draft the request message copy for the first ask and one follow up.
3. Define the routing logic for a 4 or 5 star review versus a 1 to 3 star review.

Output format:
- Timing schedule
- Request message copy (first ask and follow up)
- Routing logic for positive versus negative reviews

You are an automation engineer building a personalized outbound agent for [BUSINESS NAME] targeting [TARGET AUDIENCE].

Context:
Lead source: [LINKEDIN SALES NAVIGATOR / LIST / CRM]
What counts as a qualified lead: [CRITERIA]

Task:
1. Design the research step, what signals the agent should pull per lead before writing.
2. Draft the personalization logic, how a signal becomes a specific line in the message.
3. Draft one example outbound message using a sample lead profile.

Output format:
- Research step definition
- Personalization logic
- One example outbound message