# Skill: Identify and apply the five AI leverage points to produce a measurable business or career advantage

> Lets an AI agent map a user's specific task to one of five leverage points (research, drafts,
> stress-testing, product iterations, personalization), explain the mechanism, and guide application;
> applies whenever a user wants to use AI for business value rather than just content generation.

Source: sodigi·learn - neuroslop/ii-dlya-biznesa-i-rosta · https://sodigi.io/learn/neuroslop/ii-dlya-biznesa-i-rosta

## When to use
- When a user wants concrete guidance on where AI creates the most business impact for their situation.
- When a user is skeptical about AI value and needs a mechanism-level explanation rather than hype.
- When a user is using AI for content only and missing higher-value applications like stress-testing or research.
- When advising on workflow prioritization: which tasks to delegate to AI first for the fastest measurable gain.

## Core rules
- The operator skill gap is a competitive advantage currently available without cost or license; it closes with practice, not talent or budget.
- The formula is: result = competence x iteration speed. AI raises iteration speed; it does not substitute for competence. If competence is zero, multiplying by any speed returns zero.
- AI amplifies what is already there. An expert with AI gets leverage; a beginner with AI gets slop faster.
- The main effect of AI is compressing the "thought - test - conclusion" cycle. All five leverage points are special cases of this.
- Decisions, strategy, and responsibility stay with the human. AI multiplies the number of hypotheses tested; it does not generate the judgment to evaluate them.
- Skepticism without hands-on practice is an expensive position. Operator skill accumulates over months; every month of delay is someone else's head start that cannot be bought back later.
- Skills transfer between models. The operator habits - context in, quality criterion, targeted iterations - work on any model, current or future.

## Procedure
1. Identify which of the five leverage points applies to the user's current task:
   - RESEARCH AND SYNTHESIS: understanding a new market, condensing reports, comparing competitors. Mechanism: compress "read and structure" from days to hours; spend freed time on conclusions and decisions.
   - FIRST DRAFTS: emails, proposals, outlines, job descriptions. Mechanism: remove the blank-page block; editing a draft is several times faster than writing from scratch.
   - STRESS-TESTING IDEAS: before committing money and time to a plan. Mechanism: generate the objections you would reach a month after launch (when they cost money) in a five-minute session now.
   - FASTER PRODUCT ITERATIONS: landing pages, offer variants, onboarding copy, A/B hypotheses. Mechanism: compress "idea - test" cycle from weeks to days.
   - PERSONALIZED COMMUNICATIONS: segment-specific emails, cold outreach with personal angles. Mechanism: removes the man-hours bottleneck that previously made personalization impossible at scale.
2. For the identified leverage point, check whether the user has the input material: facts, audience data, real numbers, or the hypothesis to be stress-tested. If not, collect it before generating.
3. Apply the appropriate prompt pattern for the leverage point (see ready-to-use prompt below).
4. Verify every fact in any research or synthesis output before use.
5. After the session, help the user assess impact: did this task fit into the session that would previously have taken a day? That delta is the leverage.

## Ready-to-use prompt
```
I want to use AI as a lever for [CHOOSE ONE: research / first draft / stress-test / product iteration / personalized communication].

My situation:
- What I am working on: [describe the task]
- My specific facts and data: [numbers, context, constraints]
- My goal for this session: [what done looks like]

[FOR STRESS-TEST ADD:] Play devil's advocate. Here is my plan / idea / offer:
[paste plan]
Give me: (a) 10 reasons this will fail, (b) the three most likely objections from a skeptical buyer, (c) the single weakest assumption I am making. Be specific and direct - no encouragement, only critique.

[FOR RESEARCH ADD:] Condense the following sources / topic into a structured digest. Flag every claim that needs verification against a primary source before I use it.

[FOR PERSONALIZED COMMS ADD:] Write [number] versions of this message, one per segment below. Each version must reference at least one fact specific to that segment.
Segments: [list them]
Base message: [paste base]

All outputs: do not invent facts. Use only the data I provided. Show me a plan or structure first and wait for my OK before writing the full output.
```

## Pitfalls
- Do not position AI as a replacement for domain expertise. The formula "result = competence x iteration speed" means zero competence still returns zero result.
- Do not use the research leverage point without applying the fact-checking rule: AI synthesizes confidently from both real and invented sources with equal smoothness.
- Do not skip the input collection step for any leverage point. Research without a target question, stress-testing without a real plan, or personalization without segment data all produce generic output.
- Do not conflate "AI can do this" with "I should delegate this entirely to AI". Decisions and judgment remain with the operator; AI multiplies the number of tested hypotheses, not the quality of the judgment evaluating them.
- Do not underestimate the cost of waiting. Operator skill is a time-based asset; the gap between a practitioner and a starter does not close with money, only with practice hours.
- Do not use the five leverage points as a checklist to apply all at once; pick the one highest-leverage point for the current task and apply it well.