Prompt Library

ChatGPT Prompts for Facebook Ads

30 copy-paste prompts

Thirty structured, copy-paste prompts to research audiences, write scroll-stopping hooks, draft primary text and headlines, brief creative, sharpen offers, and run a disciplined testing program on Meta.

In short: This page contains 30 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 15 prompts are free instantly โ€” no signup needed. Hand-curated and tested by the AI Academy team.

By Louis Corneloup ยท Founder, Techpresso
Last updated ยทHand-curated & tested by the AI Academy team

Audience & Targeting

5 prompts

Customer Avatar Builder

1/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Core offer: [OFFER] I advertise on Facebook and Instagram and need a sharp avatar to anchor targeting and messaging. </context> <task> 1. Build a single primary customer avatar: demographics, job/role, income band, and life stage. 2. List their top 3 pains, top 3 desires, and the trigger event that makes them buy now. 3. Capture their internal monologue: the exact words they use to describe the problem (5 verbatim phrases). 4. Identify 5 interests, pages, brands, and creators they follow that map to Meta detailed-targeting options. 5. Note 3 objections that would stop them clicking, and the proof needed to overcome each. </task>

A targeting-ready customer avatar with verbatim language and Meta interest signals.

๐Ÿ’ก

Pro tip: After the avatar, ask ChatGPT: "Now rank these interests by how specific they are to buyers vs. broad lookalikes" to avoid bloated audiences.

Detailed Targeting Stack

2/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Meta has narrowed detailed targeting, so I want layered interests plus behaviors that survive. </context> <task> 1. Propose 4 distinct interest/behavior clusters I can test as separate ad sets. 2. For each cluster, list the specific Meta detailed-targeting terms (interests, behaviors, demographics) to enter. 3. Suggest where to use AND (narrowing) vs. OR (broadening) logic and why. 4. Recommend a starting audience size band for a [OFFER] price point and budget. 5. Flag any clusters likely to be removed or too broad, with a fallback. </task>

Four testable detailed-targeting clusters with exact Meta terms and audience-size guidance.

๐Ÿ’ก

Pro tip: Paste your past winning audiences first so ChatGPT extends what already works instead of guessing cold.

Lookalike & Custom Audience Plan

3/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Data sources I have: purchasers, email list, site visitors (pixel), video viewers, engagers. </context> <task> 1. Recommend the priority order for seed audiences to build lookalikes from, highest-intent first. 2. For each seed, suggest lookalike percentages (1%, 1-3%, 3-5%) and when to use each. 3. Design a retargeting ladder from cold to warm to hot with the custom audience behind each rung. 4. Suggest exclusions to prevent overlap and wasted spend. 5. Note the minimum data thresholds Meta needs for each audience to populate. </task>

A full custom-audience and lookalike strategy mapped to your data sources.

๐Ÿ’ก

Pro tip: Ask ChatGPT to output the plan as a table with columns Audience | Type | Size | Stage so you can build it directly in Ads Manager.

Awareness-Stage Segmentation

4/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] I want to match messaging to where the prospect sits on the awareness spectrum. </context> <task> 1. Map my audience across the 5 awareness stages (unaware, problem, solution, product, most aware). 2. Estimate what share of cold Meta traffic falls into each stage. 3. For each stage, state the message the ad must lead with to move them one step. 4. Recommend which stages to target with cold prospecting vs. retargeting. 5. Give one example ad angle per stage tailored to [PRODUCT]. </task>

Audience segmented by awareness stage with the right lead message for each.

๐Ÿ’ก

Pro tip: Use this before writing any copy โ€” it tells ChatGPT which angle to lead with so cold and retargeting ads do not say the same thing.

Pain & Desire Mining

5/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] I want raw, emotional language to feed into hooks and primary text. </context> <task> 1. Generate 10 pains my audience feels, written in first person as they would say them. 2. Generate 10 desired outcomes, also in first person, emotionally specific not generic. 3. Pair the 5 strongest pain/desire combos as before-and-after tensions. 4. Flag which combos are most likely to stop the scroll and why. 5. Suggest the single most underused angle competitors ignore. </task>

A mined list of first-person pains and desires ready to drop into hooks and copy.

๐Ÿ’ก

Pro tip: Feed real customer reviews or support tickets into the prompt so ChatGPT mirrors actual voice instead of inventing it.

Prompts get you started. Tutorials level you up.

A growing library of 300+ hands-on AI tutorials. New tutorials added every week.

Start 7-Day Free Trial

Ad Angles & Hooks

5 prompts

Ten Angles, One Product

6/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Most of my ads fail because they all sell the same way. I want genuinely different angles. </context> <task> 1. Generate 10 distinct marketing angles for [PRODUCT] (e.g. problem-agitate, status, time-saving, social proof, contrarian, identity, fear-of-missing-out, transformation, novelty, anti-competitor). 2. For each angle, write one sentence on the core promise and who it resonates with. 3. Rank the 10 by likely cold-traffic performance for [AUDIENCE]. 4. Pick the top 3 and explain why they beat the rest. 5. Note any angle that risks Meta policy issues. </task>

Ten differentiated ad angles, ranked, with the top three justified.

๐Ÿ’ก

Pro tip: Run the same prompt twice with different audiences to see which angles are audience-specific vs. universal winners.

Scroll-Stopping Hook Generator

7/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Angle to use: [chosen angle] The first line and first 3 seconds decide everything on Meta. </context> <task> 1. Write 15 opening hooks for this angle: a mix of question, bold claim, callout, pattern interrupt, and curiosity gap. 2. Keep each hook under 12 words and front-load the most arresting word. 3. Label each hook with the psychological lever it pulls. 4. Rank the top 5 for thumb-stopping power on a fast mobile feed. 5. Flag any that overpromise or could trigger a policy review. </task>

Fifteen labeled hooks with the strongest five ranked for the feed.

๐Ÿ’ก

Pro tip: Tell ChatGPT "write these for sound-off viewing" so hooks work as on-screen text overlays, not just voiceover.

Pattern Interrupt Bank

8/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] My ads blend into the feed. I need openers that break the scroll. </context> <task> 1. Generate 12 pattern-interrupt openers: visual cues, unexpected statements, callouts, and mid-action starts. 2. For each, describe the matching opening shot or image so creative and copy align. 3. Explain in one line why each interrupts the scroll. 4. Sort them into native-feeling (UGC) vs. polished (brand) buckets. 5. Recommend the 3 best for a cold prospecting campaign. </task>

A bank of twelve pattern interrupts paired with matching visuals.

๐Ÿ’ก

Pro tip: Ask ChatGPT to also write the first frame of B-roll for each โ€” pattern interrupts live in the visual, not just the words.

Hook-to-Body Bridge

9/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Hook I am using: [hook] Great hooks die if the second line breaks the promise. </context> <task> 1. Write 5 second-line bridges that keep the curiosity or tension the hook opened. 2. Ensure each smoothly hands off to a benefit or proof point about [PRODUCT]. 3. Avoid restating the hook or going generic in the bridge. 4. Show the full hook + bridge as a 2-line unit for each option. 5. Mark which bridge sustains momentum best and why. </task>

Five tested transitions that carry the hook into the body without losing attention.

๐Ÿ’ก

Pro tip: Paste a hook that already performed and have ChatGPT write only the bridges โ€” fixing the drop-off line is often a cheap win.

Angle Stress-Test

10/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Proposed angle and hook: [paste angle + hook] I want a skeptical review before I spend money. </context> <task> 1. Critique the angle as a skeptical first-time viewer would: where does belief break? 2. Identify the single weakest claim and the proof needed to support it. 3. Check for clarity: would a stranger know what the product does in 3 seconds? 4. Suggest 3 sharper rewrites of the hook keeping the same angle. 5. Predict the most likely reason this ad underperforms and how to pre-empt it. </task>

A skeptical teardown of your angle and hook with sharper rewrites.

๐Ÿ’ก

Pro tip: Add "respond only as the most cynical version of my buyer" to get harder, more useful objections.

Primary Text & Headlines

5 prompts

Primary Text โ€” PAS Framework

11/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Write Meta primary text using Problem-Agitate-Solution. </context> <task> 1. Open with a hook line that names the problem in the buyer's words. 2. Agitate: 2-3 lines on the cost of leaving it unsolved, concrete not abstract. 3. Solution: introduce [PRODUCT] and the core mechanism that makes it work. 4. Add one proof element (result, stat, or social proof) and the [OFFER]. 5. Close with a single clear CTA and keep total length tight for mobile (under 125 words before the See More cut). </task>

Mobile-optimized primary text built on the PAS framework with proof and CTA.

๐Ÿ’ก

Pro tip: Ask ChatGPT to show you exactly where the "See More" truncation lands so your hook and key promise sit above the fold.

Headline Variations Pack

12/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] The 40-character headline under the image needs to earn the click. </context> <task> 1. Write 12 headline options under 40 characters each. 2. Cover distinct styles: benefit, offer, curiosity, social proof, urgency, and outcome. 3. Keep them punchy and skimmable; no truncation on mobile. 4. Label each with its style and rank the top 4. 5. Suggest the best pairing of headline + CTA button (Shop Now, Learn More, etc.). </task>

A dozen short headline variants, labeled, with the strongest four ranked.

๐Ÿ’ก

Pro tip: Have ChatGPT count the characters for each headline so none get cut โ€” Meta truncates around 40 on mobile placements.

Benefit-Stacked Copy

13/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Features: [paste 5-8 features] Prospects do not buy features. Translate them. </context> <task> 1. Convert each feature into the concrete benefit it delivers to [AUDIENCE]. 2. Convert each benefit into the emotional outcome behind it. 3. Write primary text that leads with the strongest emotional outcome, not the feature. 4. Weave in 2-3 supporting benefits without listing them robotically. 5. End with [OFFER] and a CTA. </task>

Feature-to-benefit-to-outcome copy that leads with what buyers actually want.

๐Ÿ’ก

Pro tip: Ask for the feature-benefit-outcome chain as a table first, then have ChatGPT write copy from only the rows you approve.

Social Proof Ad Copy

14/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Proof I have: [paste reviews, numbers, testimonials, logos] </context> <task> 1. Pick the single most credible, specific proof point to lead with. 2. Write primary text framed around the proof (e.g. "12,000 [AUDIENCE] switched because..."). 3. Quote or paraphrase a testimonial naturally inside the copy. 4. Add a second proof layer (quantity, authority, or recency) without bragging. 5. Close with [OFFER] and a low-friction CTA. </task>

Proof-led primary text that converts skepticism into clicks.

๐Ÿ’ก

Pro tip: Specific numbers beat round ones โ€” tell ChatGPT to keep "12,400" rather than rounding to "12,000+" for credibility.

Copy Refresh & Variation

15/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Winning ad copy to refresh: [paste current copy] The ad is fatiguing and I need fresh variants that keep the winning DNA. </context> <task> 1. Identify what is doing the heavy lifting in the original (hook, proof, offer, or angle). 2. Write 5 variations that keep that core element but change the framing and opening. 3. Vary length: 2 short punchy, 2 medium, 1 long-form storytelling. 4. Keep the same offer and CTA intent across all. 5. Note which variant to test first against the control. </task>

Five fresh copy variants that preserve a winner while fighting ad fatigue.

๐Ÿ’ก

Pro tip: Paste the metrics (CTR, CPA) alongside the copy so ChatGPT knows which element is the proven winner to protect.

Creative Briefs

5 prompts

UGC Video Script

16/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] I need a native-feeling UGC script for a creator to film on a phone. </context> <task> 1. Write a 30-45 second script with a hook in the first 3 seconds. 2. Structure it: hook, problem, product reveal, demonstration, proof, CTA. 3. Write it in casual spoken language, not ad-speak. 4. Add bracketed direction for shots, on-screen text, and B-roll at each beat. 5. Include 2 alternative hooks the creator can A/B film. </task>

A film-ready UGC script with shot directions and alternative hooks.

๐Ÿ’ก

Pro tip: Ask ChatGPT to mark where on-screen captions should appear โ€” most UGC ads are watched on mute.

Static Image Ad Brief

17/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Brief a designer for a single high-converting static image ad. </context> <task> 1. Define the one message the image must communicate in under 2 seconds. 2. Specify the visual concept, focal point, and emotional tone. 3. Provide the exact on-image text (headline overlay) under 7 words. 4. Recommend layout, color contrast for the feed, and where the product sits. 5. List 3 variations to test (different visual concepts, same message). </task>

A designer-ready static ad brief with copy, layout, and three test variations.

๐Ÿ’ก

Pro tip: Tell ChatGPT to keep text under 20% of the image area as a habit, even though Meta relaxed the rule โ€” it still reads cleaner.

Video Ad Storyboard

18/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Storyboard a 15-second polished video ad for cold traffic. </context> <task> 1. Break the 15 seconds into scenes with timestamps. 2. For each scene specify: visual, on-screen text, voiceover/audio, and purpose. 3. Front-load the hook and product value in the first 3 seconds. 4. Build to a single CTA frame at the end. 5. Suggest a sound-off version that works with captions only. </task>

A timestamped 15-second storyboard with sound-on and sound-off versions.

๐Ÿ’ก

Pro tip: Have ChatGPT design the first frame to make sense as a static thumbnail โ€” it doubles as your image ad.

Carousel Ad Structure

19/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Design a multi-card carousel that tells a story across swipes. </context> <task> 1. Decide the carousel narrative: problem-to-solution, features, steps, or proof gallery. 2. Write card-by-card copy (headline + description) for 5 cards. 3. Make card 1 a hook strong enough to earn the first swipe. 4. Give each card a distinct visual direction. 5. End on a card with the [OFFER] and CTA. </task>

A five-card carousel with per-card copy and visual direction.

๐Ÿ’ก

Pro tip: Ask ChatGPT which single card should also work as a standalone single-image ad โ€” carousels often lose to one strong static.

Creative Concept Sprint

20/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] I need a batch of fresh creative concepts to feed my testing pipeline. </context> <task> 1. Generate 8 distinct creative concepts (format + angle + hook in one line each). 2. Span formats: UGC, static, founder talking-head, demo, comparison, testimonial montage. 3. Tie each concept to a specific awareness stage or audience segment. 4. Rank by expected production effort vs. likely payoff. 5. Recommend the 3 to produce first and why. </task>

Eight ready-to-produce creative concepts ranked by effort and payoff.

๐Ÿ’ก

Pro tip: Save the output as your creative backlog and ask ChatGPT to regenerate monthly so you never run dry on concepts to test.

Offers & Funnels

5 prompts

Irresistible Offer Builder

21/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Current offer: [OFFER] I want to make the offer feel like a no-brainer in the ad. </context> <task> 1. Restate [OFFER] as a value equation: dream outcome, perceived likelihood, time, and effort. 2. Propose 4 ways to strengthen it (bonus, guarantee, urgency, payment terms, or framing). 3. Write the offer as a single punchy line for the ad and headline. 4. Add a risk-reversal statement that removes the main hesitation. 5. Flag any claim that needs proof or could draw Meta scrutiny. </task>

A strengthened, ad-ready offer with risk reversal and a one-line pitch.

๐Ÿ’ก

Pro tip: Ask ChatGPT to compute the offer with Hormozi's value equation explicitly so you can see which lever (time, effort, likelihood) is weakest.

Landing Page Message Match

22/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Ad copy + hook: [paste] Landing page headline + first section: [paste] Clicks are wasted when the page does not match the ad. </context> <task> 1. Score the message match between ad and landing page (1-10) and explain gaps. 2. Rewrite the landing page headline to mirror the ad's promise and hook. 3. Ensure the first scroll restates the offer the ad sold. 4. Flag any disconnect in tone, audience, or claim. 5. List the 3 highest-impact fixes to lift conversion rate. </task>

A message-match audit and rewrite to stop post-click drop-off.

๐Ÿ’ก

Pro tip: Paste a screenshot description or the live URL text โ€” ChatGPT catches mismatches in tone that you stop noticing on your own page.

Funnel-Stage Ad Mapping

23/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Map which ads run at each funnel stage. </context> <task> 1. Define my funnel stages from cold to purchase to retention. 2. For each stage, specify the ad goal, message, and offer (lead magnet, trial, full price). 3. Recommend the Meta campaign objective and audience for each stage. 4. Suggest the CTA and next step for each ad. 5. Identify the stage most likely leaking and the ad to fix it. </task>

A full-funnel ad plan mapping message, audience, and objective to each stage.

๐Ÿ’ก

Pro tip: Ask ChatGPT to format this as a funnel table so you can hand it straight to a media buyer or build it in Ads Manager.

Lead Magnet Ad Copy

24/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Lead magnet: [free guide / quiz / template / webinar] I want cheap leads I can nurture toward [OFFER]. </context> <task> 1. Write primary text selling the lead magnet, not the product, with a curiosity-driven hook. 2. Make the value of the free resource feel concrete and immediate. 3. Write 3 short headlines for the lead-gen creative. 4. Specify the single field the form should ask for and why. 5. Outline the first nurture message that bridges from magnet to [OFFER]. </task>

Lead-magnet ad copy plus the bridge into your paid offer.

๐Ÿ’ก

Pro tip: Tell ChatGPT to make the lead magnet solve a problem the paid product creates next โ€” that is what turns free leads into buyers.

Retargeting Sequence

25/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Warm audiences: site visitors, add-to-cart abandoners, video viewers, past buyers. </context> <task> 1. Design a 4-ad retargeting sequence escalating from reminder to objection-handling to urgency to offer. 2. For each ad specify the warm audience, message, and creative type. 3. Write the hook and CTA for each of the 4 ads. 4. Recommend frequency caps and exclusion logic to avoid fatigue. 5. Note when to stop retargeting and let a prospect go cold. </task>

A four-step retargeting sequence with messages, audiences, and frequency caps.

๐Ÿ’ก

Pro tip: Ask ChatGPT to write a distinct angle per ad โ€” repeating the same creative to warm audiences is the fastest way to burn them out.

Go from copy-pasting to actually mastering AI.

AI Academy: 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ other tools. New tutorials added every week.

Start Your Free Trial

Testing & Scaling

5 prompts

Testing Roadmap

26/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Monthly ad budget: [budget] I test randomly and learn nothing. Give me a disciplined plan. </context> <task> 1. Define the testing hierarchy: test offer and angle before creative before copy before audience. 2. Propose a 4-week testing roadmap with one primary variable per week. 3. Recommend how many concepts/ad sets to run given my budget. 4. State the minimum spend and conversions needed before judging a test. 5. Define the kill and scale rules for each test. </task>

A structured four-week testing roadmap with clear kill and scale rules.

๐Ÿ’ก

Pro tip: Ask ChatGPT to base statistical-significance thresholds on your actual conversion volume, not generic "wait 3 days" advice.

Hypothesis Generator

27/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Recent results: [paste CTR, CPA, ROAS by ad] Turn my data into testable hypotheses. </context> <task> 1. Read the results and identify the 3 clearest patterns. 2. Convert each into a hypothesis in the form: "If we change X, then Y will improve because Z." 3. For each hypothesis design the specific test (variable, control, success metric). 4. Rank the 3 by expected impact and ease of testing (ICE). 5. Recommend the single test to run first. </task>

Data-driven, ICE-ranked test hypotheses ready to run.

๐Ÿ’ก

Pro tip: Export your Ads Manager table and paste it raw โ€” ChatGPT reads the columns and spots patterns you scroll past.

Performance Diagnosis

28/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Offer: [OFFER] Metrics: CPM [x], CTR [x], CPC [x], landing conversion [x], CPA [x], ROAS [x] The ad is underperforming and I need to know where it breaks. </context> <task> 1. Diagnose the funnel: is the problem the creative (CTR), the click cost (CPC/CPM), the page (conversion), or the offer (CPA/ROAS)? 2. Pinpoint the single weakest link with the metric that proves it. 3. Recommend the 3 highest-leverage fixes for that link. 4. Suggest the diagnostic test to confirm the fix before scaling. 5. State what good benchmarks look like for [AUDIENCE] and this objective. </task>

A funnel-level diagnosis that isolates the real bottleneck before you spend more.

๐Ÿ’ก

Pro tip: Give ChatGPT your numbers and ask it to compare ratios (CTR vs. conversion) rather than absolute values โ€” the bottleneck shows in the ratios.

Scaling Decision Framework

29/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Winning ad set: [paste metrics and current spend] Target ROAS: [x] I want to scale without breaking the winner. </context> <task> 1. Decide whether to scale vertically (raise budget) or horizontally (duplicate into new audiences) given the numbers. 2. Recommend a safe budget-increase cadence to protect the learning phase. 3. Propose 3 horizontal expansion audiences from the proven winner. 4. Set the ROAS floor that triggers pulling back. 5. List the early-warning metrics that signal fatigue before ROAS drops. </task>

A scaling playbook that grows winners without resetting the algorithm.

๐Ÿ’ก

Pro tip: Ask ChatGPT for both the 20%-every-3-days vertical path and a CBO horizontal path, then choose based on how stable your CPA has been.

Ad Fatigue Early Warning

30/30

<context> Product: [PRODUCT] Audience: [AUDIENCE] Metrics over time: [paste frequency, CTR, CPM, CPA trend] </context> <task> 1. Read the trend and tell me whether the ad is fatiguing, with the metrics that prove it. 2. Identify the frequency threshold where performance starts decaying for this audience. 3. Recommend whether to refresh creative, rotate copy, or expand audience. 4. Provide a 3-step refresh plan that keeps the winning angle. 5. Set a monitoring cadence so I catch the next fatigue cycle earlier. </task>

An early-warning read on ad fatigue with a refresh plan that protects the angle.

๐Ÿ’ก

Pro tip: Paste a week-over-week table โ€” fatigue is a trend, so ChatGPT needs the trajectory, not a single snapshot, to call it.

Frequently Asked Questions

Copy a prompt, replace the bracketed placeholders like [PRODUCT], [AUDIENCE], and [OFFER] with your real details, and paste it into ChatGPT. The structured context and task format gives ChatGPT everything it needs to return targeting plans, hooks, copy, and creative briefs you can drop straight into Meta Ads Manager.
These prompts ask ChatGPT to flag claims that could trigger a policy review, but no AI guarantees approval. Always review generated copy against Meta's advertising standards yourself, especially around health, finance, and before-and-after claims, and avoid personal-attribute targeting language.
Start with the Audience & Targeting category to build your avatar and targeting stack, since every hook, headline, and offer downstream depends on knowing exactly who you are talking to. From there move to Ad Angles & Hooks, then writing, creative, offers, and finally testing.
Yes. Several prompts, like Hypothesis Generator and Performance Diagnosis, are built to take your real metrics. Export your Ads Manager table and paste the raw numbers into the prompt, and ChatGPT will surface patterns, isolate bottlenecks, and turn them into ranked, testable next steps.
Yes. Facebook and Instagram ads run through the same Meta Ads Manager and share copy, creative, and targeting logic, so every prompt applies across feed, Reels, and Stories placements. For Reels and Stories, lean on the UGC and video prompts and ask ChatGPT to optimize for vertical, sound-off viewing.

Prompts are the starting line. Tutorials are the finish.

A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.

7-day free trial. Cancel anytime.