3 “Agent” Prompts + Examples You Can Copy

Stop chatting with AI. Start briefing it.

If you’ve been “having a conversation” with AI like it’s a wise friend over coffee, that’s charming. But it’s also why you keep getting answers that feel like a motivational poster wearing a blazer.

You know the vibe:

• “The market is growing rapidly.”

• “Competition is increasing.”

• “Focus on differentiation.”

Wow. Incredible. I feel inspired… and also still unable to choose a niche, price an offer, or explain what I sell without using the phrase “value-driven solutions.”

This article fixes that.

You’re going to get three ‘agent’ prompts that behave like a small research team. Each one has a job. Each one produces a specific output. And I’ll show you when to use each, how to tighten the scope so you don’t get generic fluff, plus common mistakes that turn AI into the world’s most polite nonsense machine.

Open these resources as you go:

• Homepage hub: https://aivillereview.com/

• Full workflow (how to get paid doing this): https://aivillereview.com/ai-market-research-how-to-get-paid/

• Workshop training (watch it done live): https://aivillereview.com/workshop/

Stop Chatting With AI. Start Briefing It.

Why “friendly chatting” creates generic fluff

AI is great at being agreeable. If you ask vague questions, it will respond with vague confidence. It’s like asking a stranger, “What should I do with my life?” and they say, “Follow your passion.”

Technically correct. Practically… we are now both unemployed.

Market research requires:

• a clear scope

• a clear decision goal

• clear constraints

• a required output format

If you don’t provide those, AI defaults to safe generalities. That’s not AI being “bad.” That’s you handing it an assignment that says: “Please guess what I mean and also don’t offend me.”

What a real brief includes (the 60-second version)

A real brief answers these:

1. What decision are we making?

2. What market and what buyer?

3. What geography and business model?

4. What constraints matter? (budget, time, skills, channels)

5. What output format do we want? (tables, scorecard, summary, next actions)

Once you provide that, AI stops being a chat buddy and starts behaving like an analyst who’s afraid you’ll forward its work to an investor.

The shortcut: 3 reusable “agent” prompts

Here’s the three-agent setup:

• Market Intel Agent: “Is this market worth entering?”

• Competitive Intel Agent: “Who am I up against and where is the gap?”

• Idea Validation Agent: “Go, no-go, or pivot and why?”

If you’re building offers or selling market research as a service, this trio is your “I can deliver a clean report by dinner” system.

Quick Navigation and Resources

• Start here: https://aivillereview.com/

• Step-by-step workflow: https://aivillereview.com/ai-market-research-how-to-get-paid/

• Workshop training: https://aivillereview.com/workshop/

The Briefing Framework (Use This Before Any Prompt)

This is the part that turns AI output from “nice” into “useful.”

Step 1: Define the decision

Write one sentence:

“This research will help decide: __________.”

Examples:

• Should I launch this offer in the next 30 days?

• Which niche should I start with first?

• What price range is realistic?

• What entry wedge is easiest?

Step 2: Define the scope

Fill in:

• Market/industry

• Buyer

• Geography

• Business model

• Price range

Step 3: Define constraints

Examples:

• I have 10 hours/week

• I have $500 budget

• I can’t run paid ads yet

• I need to sell in English-speaking countries

Constraints are where vague dreams become real plans. Without them, AI gives you advice meant for someone with unlimited time, money, and emotional stability. (So… not us.)

Step 4: Define output format

Tell AI what to produce:

• Executive summary

• tables

• competitor matrix

• scorecard

• assumptions + unknowns

• next actions

Step 5: Define “good”

Example:

“A good report includes 3–5 trends, 3 buyer segments, a competitor matrix, and 3 next actions I can do this week.”

When to Use Each “Agent” Prompt (Cheat Sheet)

Market Intel Agent

Use when you’re asking:

• Is there demand?

• Who pays?

• What are trends and risks?

• What’s the rough size of opportunity?

Competitive Intel Agent

Use when you’re asking:

• Who already wins here?

• What do buyers complain about?

• What gaps exist?

• What positioning could win?

Idea Validation Agent

Use when you’re asking:

• Should I do this at all?

• What assumptions must be tested first?

• What would make it fail?

• What’s the fastest next test?

If you’ve been using one mega prompt for everything, that’s like using one kitchen knife for cooking, shaving, and home repairs. You can, but your loved ones will stage an intervention.

Prompt #1: Market Intelligence Agent (Copy/Paste)

When to use it

Use this first when you’re evaluating a new niche, offer, or service idea.

What it’s best at

It gives you:

• market overview

• buyer segments

• trend + risk map

• TAM/SAM/SOM style logic (with assumptions)

• opportunities and entry wedge

Copy/paste prompt

MARKET INTEL AGENT PROMPT
You are a world-class market research analyst who has provided strategic intelligence for Fortune 500 companies and venture capital firms. Deliver a comprehensive market research report that supports a real business decision.

Decision this research supports: [write the decision in one sentence].

Scope:

• Market/industry: [ ]

• Geographic focus: [ ]

• Target buyer: [ ]

• Business model: [ ]

• Price range: [ ]

Deliver a polished report with:

1. Executive summary (1 page)

2. Industry overview: definition, current state, 3–5 key trends

3. Market size: TAM/SAM/SOM with assumptions and ranges (conservative/base/aggressive)

4. Customer analysis: segments, pain points, buying triggers, willingness-to-pay signals

5. Competitive landscape overview: main categories of competitors

6. Barriers and risks

7. Opportunities and recommended entry wedge

8. Next 3 actions to validate quickly

9. Methodology notes and key assumptions (explicit)

Be concise, avoid generic statements, and flag unknowns.

Example scope (safe niche)

Decision: “Should I launch AI market research reports as a service?”
Scope: US/Canada, founders + creators, freelance packages $99–$399.

Sample output preview (what “good” looks like)

You want to see sections like:

• Buyer segments (founders, marketers, creators, ecommerce)

• Triggers (“need proof before ads,” “investor deck,” “niche selection”)

• Where they already spend (Upwork, Fiverr, agencies, tools)

• Risks (AI hallucinated numbers, credibility gap, commoditization)

• Entry wedge (specialize: local businesses, ecommerce, SaaS, etc.)

If your output looks like a Wikipedia summary, tighten scope and force format.

Prompt #2: Competitive Intelligence Agent (Copy/Paste)

When to use it

Use this after you have a market direction and want to see:

• how people buy today

• who dominates

• how to position differently

Copy/paste prompt

COMPETITIVE INTEL AGENT PROMPT
You are an elite competitive intelligence specialist. Build a battle-ready competitor analysis that shows exactly who we’re up against and how to win.

Decision this research supports: [decision].

Scope:

• Market/industry: [ ]

• Target buyer: [ ]

• Geography: [ ]

• Offer: [what we sell in one sentence]

• Price range: [ ]

Deliver:

1. Competitive set (direct, indirect, future entrants, substitutes)

2. Top competitor profiles (at least 7): positioning, pricing, channels, strengths, weaknesses

3. A competitor matrix (table): Competitor | Buyer | Promise | Price | Weakness (from complaints)

4. Customer perception patterns: top praises, top complaints, switching triggers

5. “White space” opportunities: 3 entry angles with pros/cons

6. Recommended positioning statement (one paragraph)

7. Next 3 moves we should make

Avoid vague claims. Use evidence-style reasoning and list assumptions.

Sample output preview

A good output includes:

• “Most competitors sell templates, not decisions”

• “Most complaints: too generic, slow delivery, unclear sources”

• “Positioning wedge: ‘decision-grade report in 24 hours’ for X niche”

If it doesn’t give you a wedge, ask:
“Give 3 different wedges and rank them.”

Prompt #3: Business Idea Validation Agent (Copy/Paste)

When to use it

Use this last, when you want the “judge” to say:

• go/no-go

• what must be tested first

• what risks matter most

Copy/paste prompt

IDEA VALIDATION AGENT PROMPT
You are a senior business analyst specializing in validating ideas before time and money are wasted. Deliver a comprehensive validation report that ends with a go/no-go recommendation and next steps.

Business idea (one sentence): [ ]
Target buyer: [ ]
Geography: [ ]
Business model: [ ]
Target price: [ ]
Constraints: [time, budget, skills]
Decision: [what we must decide today]

Deliver:

1. Problem validation: urgency, frequency, current alternatives, willingness-to-pay signals

2. Solution validation: differentiation, feasibility, expected outcomes

3. Market validation: size, growth, competition intensity, barriers

4. Customer and reach validation: where buyers are, reachability, CAC risk

5. Business model validation: pricing realism, delivery cost/time, margin risk

6. Risk assessment: top 5 risks + mitigations

7. Go/No-Go scorecard (1–5): demand, budget, competition, differentiation, reachability, feasibility, profit potential

8. Overall recommendation: Go / No-Go / Go with constraints

9. Next 7-day validation sprint (step-by-step)

Be direct, avoid generic advice, and make assumptions explicit.

One Idea, Three Agents (How They Work Together)

Let’s run one safe example through all three “agents”:

Example idea:
“I will deliver a market research report + competitor snapshot for new ecommerce product ideas, delivered in 48 hours.”

What Market Intel should reveal

• ecommerce sellers already pay for product research tools

• pain triggers: “I’m not sure this will sell,” “I need proof before inventory”

• risk: a lot of tools exist, so you need a wedge (speed, clarity, or niche)

What Competitive Intel should reveal

• main competitor types: tools, templates, freelancers, agencies

• buyer complaints: generic results, too much data, no decision

• your wedge: “decision-ready report” with a single recommendation and next actions

What Idea Validation should decide

• Go with constraints: choose one ecommerce category first

• tighten deliverables: don’t promise “everything,” promise a clear decision

• test: sell 3 pilots, collect reviews, refine template

How to Customize Scope Without Breaking the Prompt

Narrow the market (the fastest improvement)

Pick one:

• niche (dentists vs “small business”)

• channel (Amazon sellers vs “ecommerce”)

• geography (US only)

• buyer stage (pre-launch founders)

Narrow the offer

Pick one:

• “market + competition only”

• “go/no-go scorecard”

• “customer personas + ad angles”
  Don’t sell “everything.” Sell “decision clarity.”

Add constraints

Constraints create realism:

• “No paid ads”

• “10 hours/week”

• “must deliver in 48 hours”

Force hard outputs

Add these lines:

• “Provide a table.”

• “Provide a scorecard.”

• “List assumptions.”

• “List unknowns and how to verify.”

Common Mistakes That Cause Generic Fluff (And Fixes)

Deliverable Templates Readers Can Copy (Client-Ready)

If you’re doing this for clients, here’s a simple “report skeleton” that makes you look like you own a fancy clipboard.

1-Page Executive Summary Template

• What you asked

• What we found (3 bullets)

• What it means (2 bullets)

• Recommendation (Go / No-Go / Go with constraints)

• Next 3 actions

Competitor Matrix Template

Competitor | Buyer | Promise | Price range | Weakness | Your wedge

Go/No-Go Scorecard Template

Demand (1–5)
Buyer budget (1–5)
Competition intensity (1–5)
Differentiation (1–5)
Reachability (1–5)
Feasibility (1–5)
Profit potential (1–5)

Total /35:

• 28+ = Go

• 20–27 = Go with constraints

• <20 = No-Go

Final Thoughts: Briefing Beats Chatting

The advantage isn’t that you can ask AI questions.

The advantage is that you can brief AI like a team and produce the kind of output that helps someone make a decision.

If you want the full “how to get paid doing this” workflow, read:
https://aivillereview.com/ai-market-research-how-to-get-paid/

If you want to watch the training in action, go here:
https://aivillereview.com/workshop/

And for the hub of everything:
https://aivillereview.com/

Before you build, validate.
Because building first and validating later is like cooking a 12-course meal before asking your guests if they’re allergic to everything. You might still get applause, but it’ll be from paramedics.

This guide shows a time-boxed, step-by-step validation sprint you can run in 60 minutes using AIville as your “research cockpit.” You’ll follow a safe, broad example idea, fill out a scorecard, and finish with next actions you can execute this week.

If you want the home base for all my AIville notes and reviews, start here
And if you want to watch the workshop training that this workflow supports, here’s the workshop page.

Before You Build, Validate (Or Enjoy Your New Hobby: Regret)

Most business ideas don’t fail because the founder is “bad.” They fail because:

• The pain wasn’t real enough
• The buyer couldn’t be reached affordably
• The market was crowded in a way that required a big brand budget
• The offer was vague
• Or the idea was fine but the positioning was not

Validation is just a fast way to reduce “unknowns” before you spend money, time, and emotional energy.

What “validation” means (and what it doesn’t)

Validation is not:

• Asking your friends if they like it (friends are kind, not accurate)
• Posting a poll (polls measure opinions, not purchases)
• Reading a few Reddit threads and declaring victory

Validation is:

• Identifying who pays and why
• Mapping what they buy today instead
• Finding proof they spend money on solutions
• Checking if you can reach them without burning your budget
• Running a small test that gives real signals

The 60-minute promise

In 60 minutes you can’t “prove” a business will succeed.
But you can absolutely decide:

• Go
• No-Go
• Go with constraints (pivot the niche, price, or offer)

That alone can save you weeks.

The Setup: What You Need Before the Timer Starts

Tools and tabs (keep it simple)

You need:

• AIville (or ChatGPT plus your own templates)
• a note doc (Google Doc is fine)
• a timer
• one page to write the idea brief

The one-paragraph idea brief (copy/paste)

Fill this in before you start:

Idea:
Target buyer:
Problem:
Proposed solution:
Business model (how you get paid):
Price range:
Geography:
Decision I want to make by the end of this hour:

Rules of the hour

• No rabbit holes
• No “just one more search”
• If something is uncertain, label it as an assumption
• End with next tests, not fantasies

The Case Study Idea We’ll Validate (Safe, Broad Niche)

We’ll validate a simple idea that is safe, broad, and easy to understand:

The idea in one sentence

“A weekly AI-generated content plan for local service businesses (dentists, chiropractors, HVAC, plumbers), delivered as ready-to-post social captions + ad angles.”

Who it’s for

Local service businesses that want leads but don’t want to become full-time content creators.

Business model and target price

• Monthly subscription: $149–$399
• Optional upgrade: ads + landing page copy

This is broad enough to be realistic but specific enough to evaluate.

The 60-Minute Validation Map (Minute-by-Minute)

Minutes 0–5: Define the decision

What decision will you make today?

Minutes 5–20: Problem validation

Is it real? Is it urgent? Do people pay for help?

Minutes 20–35: Solution validation

Does your approach beat alternatives? Can you deliver it fast?

Minutes 35–50: Market + competition reality check

What are competitors charging? What are they missing?

Minutes 50–60: Scorecard + next actions

Go/no-go and the 7-day sprint.

Step 1: Define the Decision (5 Minutes)

The exact question you’re answering today

“Should I test this offer for local service businesses in the next 7 days?”

Kill criteria (automatic No-Go)

If any of these are true, it’s a No-Go or pivot:

• Buyers don’t pay for anything similar today
• Acquisition would require expensive ads to start
• Delivery requires huge customization per client
• Competitors already dominate with cheap pricing + strong brand

Green light criteria

• Buyers already pay for marketing help
• A clear pain exists (leads, consistency, time)
• A simple entry wedge is available (one niche first)
• You can run a quick test within a week

Step 2: Problem Validation (15 Minutes)

You’re hunting for signals, not vibes.

What to look for

• Existing spend: agencies, tools, freelancers
• Urgency: “we need leads now”
• Frequency: weekly marketing problem
• Frustration: “posting is hard,” “I don’t know what to say,” “ads don’t work”

Fast proof of willingness to pay

Look for:

• Agencies serving this niche
• Freelancers offering content plans
• Tools built for local business marketing
• Pricing pages, packages, “done-for-you” offers

Copy/paste prompts for problem validation

Use prompts like:

Prompt A
“List the top pain points local service businesses have with marketing and content. Include evidence they pay for solutions and what they buy today.”

Prompt B
“Find the common reasons local businesses hire marketing help instead of doing it themselves. Give me decision triggers and urgency indicators.”

Prompt C
“What objections stop local businesses from paying for content services, and what words do they use to describe those objections?”

Step 3: Solution Validation (15 Minutes)

What makes your solution meaningfully different

The goal is not “better.” The goal is easier to choose.

Possible angles:

• “weekly plan in 24 hours”
• “post-ready captions + ad angles”
• “one niche specialization (dentists only)”
• “includes seasonal promos and local hooks”

Feasibility check

Ask:

• Can I deliver this in 60–90 minutes per client per week?
• Can AI handle 80% and I handle 20% polishing?
• Can I keep quality consistent?

Profit check

Example:

• Charge $199/month
• If delivery takes 2 hours/week = 8 hours/month
• That’s $24.88/hr before taxes, tools, and revisions
  Not great.

Fix it:

• Tighten the deliverable
• Reduce customization
• Raise price
• Create tiered packages

Copy/paste prompts for solution validation

Prompt A
“Compare this solution against common alternatives and list the main reasons a buyer would switch.”

Prompt B
“Design a minimal deliverable that still feels valuable at $199/month and takes under 60 minutes/week to deliver.”

Prompt C
“Suggest a positioning hook that feels specific and credible for local service businesses.”

Step 4: Market + Competition Reality Check (15 Minutes)

“Too crowded” vs “validated demand”

Crowded can be good if:

• Competitors are generic
• Reviews show dissatisfaction
• You can specialize
• You can wedge in with a tight offer

Competitor scan

Identify:

• Agency packages
• Freelancer offers
• Templates
• Local marketing SaaS

White space (entry wedge)

Instead of “local businesses,” pick:

• Dentists
• Chiropractors
• Med spas
• HVAC

One niche first makes messaging sharper.

Copy/paste prompts

Prompt A
“List the main competitor categories and typical pricing for local business content/marketing services.”

Prompt B
“Summarize competitor weaknesses based on customer complaints and gaps.”

Prompt C
“Propose 3 entry wedges and explain why each is easier to win.”

Step 5: Customer + Reach Validation (10 Minutes)

Where buyers actually are

• Local business Facebook groups
• LinkedIn
• Chambers of commerce
• Niche forums
• Local marketing communities

First 10 conversations plan

Your best test is human:

• Message 10 businesses
• Offer a free “one-page content plan sample”
• Ask what they currently do and pay for
• Ask what would make them switch

Messaging prompt

“Write a short outreach message for [niche] offering a free one-page content plan sample, with a friendly tone and no hype.”

The Go/No-Go Scorecard (Copy/Paste)

Score 1–5:

• Pain level:
• Buyer has budget:
• Competition intensity:
• Differentiation clarity:
• Reachability (cheap + direct):
• Delivery feasibility:
• Profit potential:

Total score (out of 35):

• 28–35 = Go
• 20–27 = Go with constraints/pivot
• under 20 = No-Go for now

Case Study Results (Example)

For the local business content plan idea:

• Pain level: 4
• Budget: 4
• Competition: 3
• Differentiation: 3
• Reachability: 4
• Delivery feasibility: 3
• Profit: 3

Total: 24/35 = Go with constraints

What we’d adjust

• Start with one niche (dentists or HVAC)
• Raise price or tighten deliverables
• Add upsells for landing page + ads

Next Actions: The 7-Day Validation Sprint

Day 1–2: Offer + page

• One niche only
• One clear deliverable
• One clear promise
• One clean price

Day 3–4: Outreach + 10 conversations

• 10 messages
• 5 calls
• Take notes on exact phrases

Day 5–7: Sell or test

• Offer pre-sale discount for first 3 clients
• Deliver a pilot week
• See if they renew

Common Ways People Cheat (And Still Call It Validation)

Asking friends

Your friends will support you even if your idea is “gluten-free ice cubes.”

Confusing interest with payment

Likes are not money. Money is money.

Overbuilding before proof

Your logo will not fix your offer.
Your offer will fix your logo.

Final Thoughts: Validation Is a Gift to Your Future Self

Before you build, validate. You’re not killing your dream. You’re protecting it from wasted effort.

If you want more workflows like this, you can always start at the main hub.
And if you want to watch the training workshop that pairs with this process, use the workshop page