AI-Generated Pricebook Descriptions

A home services company needed to build a pricebook for their new ServiceTitan implementation. 400+ items. Every one needed a customer-facing description that would appear on invoices and quotes. The descriptions had to be professional, explain value clearly, and—critically—not reveal manufacturer names or model numbers that would let customers comparison shop on Amazon.

Writing those descriptions manually? Two months of someone’s time, minimum. Probably closer to three once you factor in reviews, revisions, and the creeping realization that description #247 sounds exactly like description #52.

Now those descriptions generate automatically. A team member adds a raw item name to a spreadsheet. Within seconds, AI generates a customer-ready description that highlights function, capacity, and value—without revealing sourcing details. The description writes back to the record, ready to import into ServiceTitan.

400 items done in a day instead of a quarter.

The Problem

Pricebook building is a universal pain point for service businesses:

• Descriptions are the bottleneck — Setting up a new pricebook or migrating to a new FSM platform requires customer-facing descriptions for every item. Most companies either skip descriptions (making invoices look unprofessional) or assign someone to write them manually (slow and inconsistent).

• Bad descriptions enable price shopping — Include a manufacturer name or model number in your description, and customers can Google it before you’ve left the driveway. “Oh, that’s a $200 part on Amazon—why are you charging $450?” The description should sell value, not enable reverse engineering.

• Inconsistent tone damages professionalism — When descriptions are written by different people over different years, the result is a pricebook that reads like a ransom note. Some items get three paragraphs. Some get three words. Tone varies wildly. Invoices look amateurish.

The result: companies either tolerate bad descriptions, pay for expensive professional copywriting, or spend months of internal time on a task that never quite gets done right.

What This System Does

The automation monitors for new pricebook items and generates professional, value-focused descriptions that are ready for import into ServiceTitan, Housecall Pro, or any FSM platform.

Industry Context Injection

The AI knows your industry. For a water treatment company, it understands that a “Grundfos 10SQ05-160” is a submersible well pump—and can describe its function without naming the manufacturer. For HVAC, it knows the difference between condensing and non-condensing furnaces. Context shapes output.

Value-Focused Language

Descriptions emphasize what the equipment does for the customer, not what it is. “Reduces iron content to protect fixtures and eliminate staining” hits different than “Iron removal filter.”

Strategic Omissions

The prompt explicitly instructs the AI to exclude:

• Manufacturer names
• Model numbers
• Specific part numbers
• Technical details that enable comparison shopping

This isn’t about being deceptive. It’s about focusing the customer on value delivered rather than commodity pricing.

Invoice-Ready Formatting

Descriptions are constrained to 250 words with bullet-point formatting. They’re designed to look good on invoices and quotes—not to be encyclopedia entries.

The Prompt Architecture

The key to useful output is a prompt that encodes your pricing philosophy:

You are an expert equipment analyst for a solutions-driven company.

Generate customer-ready descriptions that highlight:
1. Equipment type or category (general, not specific)
2. Primary function and operational problems it solves
3. Sizing or capacity in general terms
4. Strategic advantages and reliability benefits

CRITICAL CONSTRAINTS:
- NO manufacturer names
- NO model numbers
- NO specific part numbers
- Focus on functional details and practical benefits
- Descriptions should NOT enable reverse engineering or comparison shopping

FORMAT REQUIREMENTS:
- 250 words or less
- Use bullets for scannability
- Invoice-ready simplicity

INDUSTRY CONTEXT:
We are in the [water well pump / HVAC / electrical / plumbing] space.
Use this context when interpreting items.

Why These Constraints Matter

No manufacturer names protects margins. When a customer sees “Carrier” or “Rheem” on their invoice, they price shop. When they see “high-efficiency condensing furnace rated for homes up to 2,500 sq ft,” they evaluate value.

General sizing language prevents spec matching. “Suitable for medium-scale operations” or “handles up to 3-4 bathroom homes” gives context without enabling exact product identification.

Bullet formatting respects the invoice context. Nobody reads paragraphs on a bill. Bullets scan quickly and look professional.

Industry context shapes interpretation. The same part number means different things in different trades. The AI needs to know whether it’s writing for plumbers, electricians, or HVAC techs.

Real-World Example

Input

Item: Franklin Electric 10FA05S4-PE 1/2 HP 10GPM Submersible Pump

AI-Generated Description

Deep Well Submersible Pump System

• Designed for residential well applications with depths up to 250 feet • Delivers consistent water flow for homes with 2-3 bathrooms • Stainless steel construction resists corrosion in mineral-rich water • Quiet operation—no above-ground pump noise • Energy-efficient motor reduces operating costs vs. jet pump alternatives

Ideal for:

• Primary home water supply from drilled wells
• Replacement of aging or failing well pumps
• Upgrading from less efficient pump styles

Professional installation ensures optimal placement, electrical connections, and system longevity. Includes pressure tank coordination for consistent household water pressure.

What the AI Did

1. Classified the equipment — Recognized this as a submersible well pump without using the manufacturer name (Franklin Electric) or model number
2. Translated specs to benefits — “1/2 HP 10GPM” became “homes with 2-3 bathrooms” and “consistent water flow”
3. Added context the customer cares about — Quiet operation, energy efficiency, corrosion resistance
4. Formatted for invoices — Clean bullets, scannable structure, under 250 words
5. Protected sourcing — Nothing in this description enables price shopping

What Didn’t Appear

• ❌ “Franklin Electric” (manufacturer)
• ❌ “10FA05S4-PE” (model number)
• ❌ “1/2 HP” or “10GPM” (specific specs that enable matching)

The Data Flow

Batch Processing Considerations

The workflow processes items one at a time in a loop. This is intentional:

• Prevents rate limiting on the AI API
• Ensures each description is written fresh (not influenced by prior items)
• Allows partial failures without losing all progress

For bulk pricebook builds, you can paste hundreds of items at once. The workflow processes them sequentially, typically completing 100+ items per hour.

What You Get

Metric	Before	After
Time per description	5-10 minutes	Seconds
500-item pricebook	2-3 months	1 day
Consistency	Variable	100% consistent tone
Price shopping risk	High (if specs visible)	Minimized

Quality Benefits:

• Every description follows the same structure
• All items have professional, customer-ready language
• New items get descriptions immediately
• No more “placeholder” or “TBD” descriptions on invoices

Margin Protection:

• Customers can’t easily comparison shop
• Focus stays on value delivered
• Professional presentation justifies premium pricing

Operational Benefits:

• New team members can add items without copywriting skills
• Pricebook maintenance becomes trivial
• FSM platform migrations don’t require manual description writing

Implementation Patterns

Pattern 1: New Pricebook Build

For a new ServiceTitan implementation:

1. Export supplier catalog to spreadsheet
2. Paste item names into Airtable (one column)
3. Let workflow generate all descriptions overnight
4. Review, adjust as needed
5. Export with descriptions for ServiceTitan import

Pattern 2: Ongoing Maintenance

For day-to-day operations:

1. When adding new items to supplier orders, add to Airtable
2. Description generates automatically
3. Copy to ServiceTitan when creating the pricebook entry

Pattern 3: Pricebook Cleanup

For fixing an existing pricebook with bad/missing descriptions:

1. Export current pricebook to CSV
2. Import items without good descriptions into Airtable (leave Description column empty)
3. Workflow generates fresh descriptions
4. Merge back into pricebook

Going Further

This foundation enables more sophisticated pricebook management:

• Category-specific prompts — Different prompt templates for equipment vs. materials vs. labor. Equipment gets feature descriptions; materials get application context; labor gets scope clarity.

• Good-better-best bundling — Generate descriptions for equipment bundles that upsell premium options without revealing the underlying components.

• Multi-language support — Generate Spanish or other language versions for multilingual customer bases.

• Competitor intelligence — When you know what competitors are charging, adjust value language to justify premium positioning.

• Dynamic pricing integration — Connect descriptions to margin targets. Items with higher margins get more value-focused language; commodity items get more matter-of-fact descriptions.

The hard part isn’t generating descriptions. It’s knowing what to say—and what not to say. That’s where pricing strategy meets AI capability.

The pricebook is where margin gets made or lost. Descriptions that sell value instead of enabling comparison shopping compound over thousands of invoices per year.

That’s not a copywriting project. That’s a margin protection strategy.