SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsthree-value-drivers

three-value-drivers

Office & productivity

View repository

Updated 10-15 23:44

About this skill

Source Metadata

Original cangjie-skill frontmatter from the distillation run:

id: f23
title: "AI价值驱动三途径 — 降本/增收/新业务"
book: "AI for Everyone"
author: "Andrew Ng（吴恩达）"
source: p15
tags: [value-assessment, business-model, project-selection, roi]
related_skills:
  - slug: triple-due-diligence, relation: depends-on
  - slug: cross-functional-brainstorming, relation: composes-with

R — Reading (original quote)

"AI projects can create value in three ways: reducing costs, increasing revenue, creating new businesses... ask 'What cost does it reduce? What revenue does it increase? What new business does it create?'" — p15

I — Interpretation (restate in your own words)

When evaluating the business value of an AI project, Andrew Ng provides a concise three-question framework:

  1. Reduce costs: Can AI automate certain tasks, thereby reducing labor costs, time costs, and error costs?
    • Example: customer service automation reduces human agents, quality inspection automation reduces inspectors, predictive maintenance reduces downtime losses
  2. Increase revenue: Can AI drive more conversions, higher average order value, and a better user experience?
    • Example: recommendation systems boost cross-selling, dynamic pricing improves profit margins, personalization improves user retention
  3. Create new business: Can AI enable the company to enter an entirely new business domain or launch an entirely new product?
    • Example: the data itself as a new product, AI capabilities opened up as a platform, entirely new intelligent services

Why is this framework useful?

  • Avoid "AI for AI's sake" — every project must be able to answer one of these three questions
  • Helps structure the quantification of project value during business due diligence
  • The three pathways have different risk and return characteristics: cost reduction is the most predictable, revenue increase requires changes in user behavior, and new business has the highest risk but also the highest ceiling

A1 — Past Application (past application cases)

  • Cost reduction case: A manufacturing company used AI visual inspection to replace manual quality inspection. Each inspector has an annual salary of 80,000, and the AI system replaced 5 inspectors, saving 400,000 in annual labor costs. The project investment was 600,000, with a payback period of 1.5 years.
  • Revenue increase case: An e-commerce platform introduced a recommendation system. The recommendation system increased the user conversion rate from 2% to 3.5%. Based on annual GMV of 1 billion, the incremental revenue was 15 million.
  • New business case: A bank opened up its internal risk control model as a SaaS service, providing risk control capabilities to other small banks. This is an entirely new business line, unrelated to the original banking business.

A2 — Future Trigger (future trigger conditions)

Trigger scenarios:

  • Evaluating whether an AI project is "worth doing"
  • Explaining the business value of an AI project to management/investors
  • Comparing the priorities of multiple AI projects
  • Needing to quantify project value during the business due diligence stage

Language signals:

  • "How does this project make money"
  • "How is the ROI of an AI project calculated"
  • "What business value does this project have"
  • "Is it worth investing in"
  • "How much revenue can it bring"

Key distinction points:

  • Does the user needquick classification(choose one of three) ordetailed quantification(financial model)? This skill provides a classification framework; detailed quantification requires triple-due-diligence.
  • Is the project in theidea stageor theevaluation stage? Use cross-functional-brainstorming in the ideation phase, and this skill in the evaluation phase.

E — Execution (execution steps)

  1. Identify the value driver type: Does this project mainly create value through cost reduction, revenue increase, or new business? (multiple selections allowed)

  2. Quantify value：

    • Cost reduction: calculate the replaceable labor/time/error costs
    • Revenue increase: estimate the improvement potential for conversion rate/average order value/retention
    • New business: assess market size and barriers to entry

  3. Assess investment: estimate development cost, data cost, deployment cost, and maintenance cost

  4. Calculate ROI: (quantified value - investment cost) / investment cost

  5. Risk assessment：

    • Cost reduction: can the technology meet the automation requirements?
    • Revenue increase: will user behavior change as expected?
    • New business: does the market exist? What is the competitive landscape?

  6. Comprehensive judgment: if you cannot answer all three questions, the project value is questionable.

Completion criteria: clarify the project's value driver type, and provide a quantified value estimate and input-output analysis.

B — Boundary (usage boundaries)

Not applicable scenarios:

  • Pure research projects (no commercial application goal)
  • Infrastructure-type projects (indirect value, difficult to quantify directly)
  • Compliance/ethics-driven projects (value is not at the economic level)

Failure modes:

  • Saying all projects have "all three" — no focus
  • Overly optimistic estimates of the value of revenue increase/new business
  • Ignoring implementation costs and change management costs
  • Looking only at economic value and ignoring strategic value (such as data accumulation and capability building)

Author blind spots:

  • Andrew Ng's three-question framework leans toward short-term economic value, but some AI projects' value isstrategic— such as accumulating data assets, cultivating AI capabilities, and building organizational confidence. These long-term values are difficult to measure with the three-question framework.
  • The three pathways are not mutually exclusive — a project may simultaneously reduce costs and increase revenue, and it is necessary to determine which is the main driver.

Related skills

  • triple-due-diligence — depends-on: business due diligence needs to quantify project value, and the three value drivers framework provides a structured method for the "value assessment" part of business due diligence.
  • cross-functional-brainstorming — composes-with: after brainstorming produces candidate projects, use the three value drivers framework to assess the value of the candidate projects and prioritize them.