Decide Build or Buy for AI

Executive and Founder, GTM Engineering, Revenue Operations, Revenue Finance.

• Any AI capability decision with a price tag over a pilot.
• When a vendor renewal and an engineering estimate land in the same quarter.

• The capability you need and the workflow it serves
• Vendor quotes for buying
• Your engineering capacity and what else it would not do

Best for one task. Copy it, add your information, and run it in your AI assistant.

You are an advisor helping me decide build versus buy for one AI capability: [paste]. Vendor quotes: [paste]. Our engineering capacity and competing priorities: [paste]. Walk me through the decision: the true three-year cost of each side including maintenance, model changes, and the opportunity cost of engineering time; what we would need to believe for build to win; what we would need to believe for buy to win; and your recommendation with the assumption it rests on. If the capability is close to our core product advantage, say so early, because that changes the answer.

Best for repeatable work. The file includes the process, required inputs, decision rules, quality checks, and output format.

---
name: decide-build-or-buy-ai
description: A structured build-versus-buy call for one AI capability, with the hidden costs on both sides priced in.
license: MIT
metadata:
  author: The Revenue AI Report
  version: 1.0.0
  last-reviewed: 2026-09-04
  source: https://www.therevenueaireport.com/skills/decide-build-or-buy-ai
---

# Decide Build or Buy for AI

A structured build-versus-buy call for one AI capability, with the hidden costs on both sides priced in.

## When to use this skill

- Any AI capability decision with a price tag over a pilot.
- When a vendor renewal and an engineering estimate land in the same quarter.

## Inputs to collect

- The capability you need and the workflow it serves
- Vendor quotes for buying
- Your engineering capacity and what else it would not do

## Process

1. Define the capability as a workflow outcome, not a feature list.
2. Price both sides over three years, including maintenance and model churn on the build side.
3. Price the opportunity cost of engineering time on the build side.
4. Ask whether the capability is core to your product advantage. Core tilts to build; commodity tilts to buy.
5. Make the call and record the assumption it rests on.

## Decision rules

- Build only wins when the capability is core or no vendor fits the workflow. Save money is not a build reason.
- Buy only wins when the vendor's roadmap matches yours and exit is affordable.
- Whichever you pick, write the kill criteria for the decision itself.

## Output requirements

- Three-year true cost on both sides.
- The belief required for build and for buy.
- Recommendation with its load-bearing assumption.

## Quality checks

- Maintenance and model churn are priced into build.
- Opportunity cost of engineering is stated.
- The core-versus-commodity question is answered explicitly.

## Limitations

- Three-year forecasts of AI pricing are soft. Treat totals as ranges.
- The build estimate is only as honest as your engineering capacity number.

## Example input

Capability: forecasting assistant over our CRM. Buy quote: $90k per year. Build estimate: 2 engineers for 4 months, then ongoing maintenance.

## Example output

Three-year buy: roughly $270k plus integration time. Three-year build: roughly 14 engineer-months including maintenance and model updates, plus those engineers not shipping two roadmap items. The capability is not core to your product. Recommendation: buy, with a contract out at renewal and kill criteria if forecast accuracy does not improve within two cycles. Assumption to record: vendor roadmap keeps pace with your CRM changes.

## Review checklist

- Three-year costs on both sides?
- Core-versus-commodity answered?
- Kill criteria for the decision itself?

## Works with

- Playbook: The agent control plane (L4) (revops, L4) https://www.therevenueaireport.com/playbooks/agent-control-plane-l4
- Playbook: Rebuild coverage around agents, not headcount (L6) (org, L6) https://www.therevenueaireport.com/playbooks/rebuild-coverage-around-agents-not-headcount-l6
- Playbook: Point retrieval at the revenue corpus (L4) (enablement, L4) https://www.therevenueaireport.com/playbooks/point-retrieval-at-revenue-corpus-l4
- Tool: Structured (Data & Analytics) https://www.therevenueaireport.com/tools/structured
- Tool: LiteLLM (Foundation Models & Infrastructure) https://www.therevenueaireport.com/tools/litellm
- Tool: StartStak (Foundation Models & Infrastructure) https://www.therevenueaireport.com/tools/startstak
- Tool: WhyHow.AI (Foundation Models & Infrastructure) https://www.therevenueaireport.com/tools/whyhow-ai
- Tool: Helicone (Foundation Models & Infrastructure) https://www.therevenueaireport.com/tools/helicone

## Rules of conduct

- Write for a Director, VP, or operator. Short sentences. Explain uncommon terms.
- Separate facts from assumptions. Never hide uncertainty.
- Do not invent numbers, benchmarks, quotes, or customer names.
- Do not send messages, change CRM records, or publish anything unless the user explicitly asks.
- Flag when a decision needs human review.

## Evidence

This skill is grounded in The Revenue AI Report research:
- https://www.therevenueaireport.com/research/proof-gap

- Related framework: https://www.therevenueaireport.com/frameworks/proof-gap



Source and updates: https://www.therevenueaireport.com/skills/decide-build-or-buy-ai

1. 1.Define the capability as a workflow outcome, not a feature list.
2. 2.Price both sides over three years, including maintenance and model churn on the build side.
3. 3.Price the opportunity cost of engineering time on the build side.
4. 4.Ask whether the capability is core to your product advantage. Core tilts to build; commodity tilts to buy.
5. 5.Make the call and record the assumption it rests on.

• Build only wins when the capability is core or no vendor fits the workflow. Save money is not a build reason.
• Buy only wins when the vendor's roadmap matches yours and exit is affordable.
• Whichever you pick, write the kill criteria for the decision itself.

• Three-year true cost on both sides.
• The belief required for build and for buy.
• Recommendation with its load-bearing assumption.

Capability: forecasting assistant over our CRM. Buy quote: $90k per year. Build estimate: 2 engineers for 4 months, then ongoing maintenance.

Three-year buy: roughly $270k plus integration time. Three-year build: roughly 14 engineer-months including maintenance and model updates, plus those engineers not shipping two roadmap items. The capability is not core to your product. Recommendation: buy, with a contract out at renewal and kill criteria if forecast accuracy does not improve within two cycles. Assumption to record: vendor roadmap keeps pace with your CRM changes.

• Three-year costs on both sides?
• Core-versus-commodity answered?
• Kill criteria for the decision itself?

What does the Decide Build or Buy for AI skill do?

A structured build-versus-buy call for one AI capability, with the hidden costs on both sides priced in.

Who is the Decide Build or Buy for AI skill for?

Executive and Founder, GTM Engineering, Revenue Operations, Revenue Finance. It sits at the advanced level and takes about 1 hour.

What do I need before I start?

Collect these first: The capability you need and the workflow it serves; Vendor quotes for buying; Your engineering capacity and what else it would not do.

What is the difference between the quick prompt and the SKILL.md file?

The quick prompt is for one task. Copy it, add your information, run it. The SKILL.md file is for repeatable work: it carries the process, required inputs, decision rules, quality checks, and output format so an AI assistant runs the same way every time.

What should I check before trusting the output?

Three-year costs on both sides? Core-versus-commodity answered? Kill criteria for the decision itself?

Is it free to use?

Yes. Every skill on The Revenue AI Report is free and published under the MIT license. Attribution is welcome, not required.

• Three-year forecasts of AI pricing are soft. Treat totals as ranges.
• The build estimate is only as honest as your engineering capacity number.

Run the skill, then roll it out with a playbook. Vendor links are supporting context, not a recommendation.

• Playbook: The agent control plane (L4) (revops, L4 L4 Orchestrated)
• Playbook: Rebuild coverage around agents, not headcount (L6) (org, L6 L6 Rebuilt)
• Playbook: Point retrieval at the revenue corpus (L4) (enablement, L4 L4 Orchestrated)
• Tool: Structured (Data & Analytics)
• Tool: LiteLLM (Foundation Models & Infrastructure)
• Tool: StartStak (Foundation Models & Infrastructure)
• Tool: WhyHow.AI (Foundation Models & Infrastructure)
• Tool: Helicone (Foundation Models & Infrastructure)

• Research: proof gap
• Related framework