24 AI agent prompts for Marketing

Copy-paste, beginner-friendly prompts to help you plan campaigns, track competitors, write ads and grow SEO traffic. Each one is based on a real-world example and works with ChatGPT, Claude or an AI agent. 17 are free below, and 7 more are in AgentPlaybook Pro.

Open Marketing in the interactive library →

Free Marketing prompts

Competitor watch report

Monitor competitor pricing, launches and messaging.

Check the websites and recent social posts of [competitor 1, 2, 3]. Report: pricing changes, new features/products, key messaging, and one opportunity for my business [describe]. Use a table and cite links.

Cohort-based offers

Send offers to groups (repeat buyers, cart abandoners) instead of one big list, as Tweens does.

Here's my customer data summary: [paste]. Propose 4 customer segments (e.g. repeat buyers, recent browsers, checkout abandoners, new customers), and for each: the offer, the message, best channel, and timing.

SEO article from a keyword

Research live results and draft an SEO article outline and draft.

Target keyword: [keyword]. Analyze what the top-ranking pages cover, then give me: search intent, an H2/H3 outline that beats them, a 1,500-word draft in a friendly tone, a meta title (<60 chars) and meta description (<155 chars).

Ad copy variations

Create multiple ad variations to test.

Write 6 ad variations for [product] targeting [audience] on [Meta/Google/LinkedIn]. Mix angles: pain, benefit, social proof, urgency. Respect character limits for that platform.

Audit a local business's website and pitch a clickable redesign

Have one agent click through a real website like a customer and log every point of friction, while a coding agent gathers customer complaints and builds a clickable prototype of a better version you can pitch as a paid service. An X post showed this workflow on a pizza chain's ordering site, where the agent found bugs like a $4.95 charge on an empty cart.

You are a website conversion consultant helping me pitch a redesign to [business name] at [URL]. Step 1: walk through the main customer task on the site ([e.g. ordering a pizza / booking a visit]) and list every step, what went wrong, and anything confusing, with screenshots or exact page text as evidence. Step 2: gather public reviews and complaints about the site and separate old issues from ones you can still reproduce today. Step 3: propose a simpler flow and build a clickable prototype that keeps their official branding and photos. Step 4: write a one-page pitch for the owner listing the top five problems, how the new design fixes each, and what you did not test. Don't contact the business or submit any real orders or payments.

Make personalized marketing videos for each target account

Use an agent to write a short script tailored to each prospect, fill a design template with their name and details, and generate a voiceover so every account gets its own video. Legal AI company Harvey's marketing team shared how they use Figma, ElevenLabs and Claude with agents to produce hyper-personalized marketing videos at scale.

You are my personalized video producer. Our product: [one-line pitch]. Target accounts: [paste list: company, contact name, role, industry, one pain point or recent news]. Video length: [30-45] seconds. Step 1: for each account, write a voiceover script (about 90 words) that opens with their name and company, names their specific pain point, shows how we solve it with one concrete example, and ends with one clear call to action. Step 2: list the on-screen text for 3-4 scenes (headline, supporting line, logo/name placement) to drop into our [Figma] template. Step 3: give voice direction (tone, pace) for the voiceover tool. Step 4: output everything as a table, one row per account. Use only the facts I provided about each account; if a detail is missing, write [NEEDS INFO] instead of guessing.

Let an AI assistant make your website updates so you never have to log into the back end

Many owners avoid touching their own website because they're scared of breaking it, so prices, hours and pages go stale. Describe the change in plain words and let an AI assistant tell you exactly what to edit, or make the edit through a connected tool, then check it before it goes live. On Oct 21, 2026, Hustle founder Sam Parr posted that he's still on WordPress but 'claude's doing everything for me', and asked if theme and plugin businesses are still needed; the post drew 300+ replies from owners sharing similar setups.

You're my website helper. My site runs on [WordPress / Squarespace / Wix / Shopify / not sure] at [URL]. I'm not technical and I'm nervous about breaking things. The change I want: [describe, e.g. update holiday hours, add a new service with a price, swap the homepage photo, fix a typo on the About page]. Please: 1) tell me exactly where to click, step by step, using the names I'll see on screen; 2) give me the finished text to paste, written in the same tone as this sample of my site: [paste a paragraph]; 3) tell me how to preview it before publishing and what to check (phone view, links, prices); 4) tell me how to undo it if something looks wrong. If the change needs a plugin, code or a paid add-on, say so and give me the simplest free option first. Do one change at a time and wait for me to say 'done'.

Spot a brewing brand crisis early by having an agent watch social posts, reviews and support messages for unusual spikes

A bad product batch, a viral complaint or a rude staff video can spread for hours before anyone on the team notices, and by then the story is set. AI agents can scan social mentions, reviews and support messages, spot an unusual jump in a complaint or negative tone, show who is spreading it, and draft a response plan. On Oct 22, 2026 Sprinklr announced its Voice of the Customer agent suite in Google's Gemini Enterprise Agent Marketplace, including a Crisis Monitoring agent that detects emerging issues, assesses how far they are spreading, identifies key amplifiers and supports response planning, alongside agents for brand health and competitor benchmarking.

Here are our brand mentions from the last 72 hours: [paste social posts, reviews and support messages with date, platform and follower count if known]. Here is what a normal week looks like for us: [typical number of mentions per day and usual complaints]. 1) Compare the last 72 hours to normal and tell me if any topic, complaint or product is spiking. Show the numbers. 2) For anything spiking, summarize what people are saying in plain words with 3 real quotes, when it started, and the 5 accounts with the most reach that are spreading it. 3) Rate the risk as low, medium or high and explain why. 4) Draft a holding statement, a reply template for customers who ask, and a short note for our team about what to do and not say. 5) Tell me what to watch over the next 24 hours. Only use the posts I gave you; don't guess at anything not shown.

See your website like a first-time visitor: have a fresh AI session score what's clear, confusing or fake-looking in the first 10, 30 and 60 seconds

If you built or wrote your own site (or an AI built it for you), you can't see it the way a stranger does. A fresh AI chat with no memory of your project can act like a first-time visitor, judge only what it has seen so far, and tell you where it would leave, what looks fake and which button to click next, on desktop and on phone, then suggest a few precise fixes with replacement copy. Builder kloss shared this prompt on X on Oct 20, 2026, warning that a coding agent is "the shittiest 1st time visitor" because it's testing the site it just built itself, and recommending a fresh session that scores clarity, trust and next step.

Open [website URL] as a first-time visitor who knows nothing about this business. Don't use any past chats, memory or code. Scroll slowly, close popups, and check it at desktop width and on a phone-size screen (or use the screenshots I paste: [paste]). Judge only what you've seen so far: First screen (about 10 seconds): What is this, who is it for, and what could it be mistaken for? After 3 screens (about 30 seconds): What would make you leave? After 6 screens (about 60 seconds): Explain it in one sentence. What's confusing? Full page: What looks fake or untrustworthy? Where did your interest drop? Do the buttons and offers compete with each other? Phone: What's cramped, cut off or hard to tap? Then score Clarity, Trust and Next step from 1-10, harshly (under 9 means not ready). Name the one idea you'd remember tomorrow. Finally, list up to 5 fixes, highest impact first, each with the exact replacement text. Don't invent facts, prices or numbers about my business.

Check whether your website's lead numbers are real: have an AI read your tracking setup and find what's being counted as a lead that isn't one

Analytics dashboards often count the wrong things as leads, like brochure downloads, newsletter signups or repeat form loads, so you end up making ad and staffing decisions on inflated numbers. An AI agent can read your site's tracking code and conversion settings, list exactly which actions are counted as leads, and compare that against the real enquiries in your inbox or CRM. Monill Parikh posted on X (Oct 22, 2026) that Google Analytics showed 245 leads for a website when the real number was 14; his AI agent caught it by reading the site's tracking code and finding that every catalogue download was being counted as a lead.

You are my analytics auditor. My website is [URL]. Here is what my analytics tool reports as leads or conversions for [date range]: [number, and the names of the conversion events if you know them]. Here is my tracking setup: [paste the tracking/tag code from the site, the Google Tag Manager tags and triggers, or a screenshot or export of the conversion events]. Here are the real enquiries we received in the same period: [count, or paste from the inbox/CRM: date, name or ID, source]. Do this: 1) List every event that is counted as a lead or conversion and what action on the site fires it (form submit, thank-you page view, button click, file download, etc.). 2) Flag events that are not real enquiries (downloads, newsletter signups, page loads, duplicate fires, test traffic, our own team's visits). 3) Estimate how many of the reported leads came from each event, and compare the total with the real enquiry count. 4) Recommend exact changes: which events to stop counting as leads, which to keep as separate 'micro' goals, and how to make the real form submission fire once. 5) Give me a short before/after summary I can share with whoever runs our ads. Don't change any settings yourself; if something is unclear from what I pasted, say what you'd need to check.

Catch wrong dates, times and locations other people post about your events or business: a morning check that finds every listing and flags what doesn't match

Venues, partners, local media, tourism sites and chambers of commerce often repost your event or business details, and they don't always get them right. One wrong showtime on a page you didn't know existed can send customers to an empty field. An AI assistant with web search can check each upcoming event every morning, find where it is listed, compare every date, time, location and ticket link against your own details, and give you a short list of mismatches with links so you can ask for corrections. Larissa Guetter, who runs the two-person ATV Big Air Tour across 26 US events a season, wrote (Oct 2026) that her scheduled 8 AM ChatGPT briefing caught a regional tourism site advertising a 7:00 PM show as 8:00 PM; OpenAI reported the check cut her listing reviews from about eight hours a week to one, with her still deciding what is correct and sending the corrections herself.

You are my listing accuracy checker. Here are my upcoming events or business details, which are the correct version: [paste a table: event or location name, date, start time, address, ticket or booking link, price, phone, hours]. Places I already know list us: [paste URLs, e.g. our website, venue page, Facebook event, ticket site, local media]. Do this: 1) Search the web for every other page mentioning each event or my business (tourism sites, chambers of commerce, event calendars, local news, social posts, directories). 2) For each page found, compare the date, time, location, ticket link and price against my correct details. 3) Give me a table of mismatches only: page link, who runs it, what it says, what it should say, and how serious it is (could make someone miss the event or pay the wrong price = high). 4) List pages you found that I didn't already know about. 5) For each high or medium mismatch, suggest who to contact and draft a short, friendly correction request I can send. 6) Keep a running log so tomorrow's check only reports new or still-unfixed problems. Don't decide which version is true when my own details are missing; flag it for me instead, and don't contact anyone yourself.

Get a daily research email your agent can read: schedule an AI search every morning that emails a compact, cut-off-safe summary of trending posts and people to engage with

Paying for a social media API just to find posts worth replying to costs more than most solo creators want to spend. A scheduled AI search that emails you its findings every morning can act as a cheap data feed for you or your other agents. Stefano Caioni shared on X (Oct 2026) that a Grok automation runs a search every morning and emails him the result; because the email preview only shows about 1,000 characters, his prompt asks for a compact block first with an END line to catch cut-offs, and a script imports it so his agent gets fresh reply targets, trending posts and people to follow, for almost nothing compared with the X API.

Every morning, search X and the web for the last 24 hours on these topics: [your niche, e.g. AI tools for small business]. My audience is [who you want to reach]. Return your answer in two parts. Part 1, a compact block of no more than 900 characters, starting with the line START and ending with the line END, containing: up to 5 posts worth replying to (link plus one-line reason), up to 3 trending topics, and up to 3 accounts worth following (handle plus why). Part 2, below the END line, longer notes and a suggested angle for one reply in my voice: [describe your voice]. Only include links you actually found; never invent a post or account. If the END line is missing when I or my script read the email, treat the block as cut off. I'll write and post any replies myself.

Find out which of your X posts actually work: have Grok study your last two months of posts and give you a 15-day posting plan with example posts

Instead of posting blindly, Qais (@Ahmadullah_Faiz) shared on X (Oct 23, 2026) a prompt for Grok to analyze your X profile from the past two months, find which posts get the most engagement and what you're doing wrong, and build a concrete 15-day posting plan. They suggest asking for 10 example posts based on what already worked, then pasting the report into ChatGPT to catch anything Grok missed, and say it worked for them. This prompt expands that into a full routine.

Analyze my X profile [@username] from the past 2 months. Show me: 1) my 5 best and 5 worst posts by engagement and what each group has in common (topic, format, length, hook, time of day, media), 2) what I'm doing wrong that's holding my reach back, 3) a concrete 15-day posting plan with one post idea per day, the best time to post, and the format to use, and 4) 10 example posts written in my voice, based on what has already worked for my account. Use only posts you can actually see and tell me if you can't access some. Afterward I'll paste this report into a second AI and ask: "Review this X account analysis. What did it miss or get wrong, and what would you change in the 15-day plan?"

Find 30 free backlinks you can actually get: have an AI research real directories and resource pages for your niche, verify each one, and hand you a 7-day submission plan

Grant O (@2whazzuup) shared a backlink research prompt on X (Oct 22, 2026) that drew 400+ bookmarks; SEO consultant Jesper Nissen reposted it the next day, saying he has used similar prompts to find hundreds of backlinks. The AI acts as a link-building researcher: given your site, niche, audience and country, it looks for free, legitimate opportunities in niche directories, industry resource pages, business listings, professional associations and free tool directories. For each one it gives the submission URL, why it fits, the requirements and the effort. Its rules say to verify each site exists, skip guest posts, paid placements and link farms, never invent URLs, and return fewer than 30 if fewer qualify. It finishes by ranking the list and building a 7-day plan.

Act as an expert SEO link-building researcher. My website: [URL]. My niche: [NICHE]. My target audience: [AUDIENCE]. My target country: [COUNTRY]. Find up to 30 legitimate, free backlink opportunities relevant to my site. Research: niche-specific directories, industry resource pages, free business and website listings, community directories, professional association directories, resource submission sites, relevant educational resources, and free tool/website directories. For EVERY opportunity give: website name, direct submission URL, why it's relevant to my niche, whether submission is free, how to get the backlink, any approval requirements, and estimated effort (Easy/Medium/Hard). Rules: verify each website actually exists and that submission is currently open; exclude guest posting, paid placements and obvious link farms; don't invent URLs or instructions; if fewer than 30 qualify, return only the verified ones. Then rank them from highest to lowest priority by relevance, editorial quality and likelihood of acceptance, and give me a 7-day plan showing which ones to submit to first and what to prepare (logo, short description, contact email) before I start.

Write a listing description from your walkthrough notes that stays within fair-housing rules and your MLS character limit

Turn rough walkthrough notes and the facts sheet into an MLS-ready listing description, a short version for portals and social, and a list of claims you still need to verify. The prompt keeps the copy about the property, not about who should live there, and flags wording that fair-housing advertising guidance warns against.

Write a listing description for this property. Facts (only use these, don't invent any): [address or area, price, beds, baths, square feet, lot size, year built, parking, HOA fee, recent updates with years]. My walkthrough notes: [paste your notes or voice-memo transcript]. Best 3 features in my opinion: [list]. Tone: [warm / modern / luxury / plain]. MLS public remarks limit: [e.g. 1,000] characters. Give me: 1) MLS public remarks within the limit, opening with the strongest feature, 2) a 280-character version for portals and social, 3) 5 headline options under 60 characters. Rules: describe the property and the area's amenities, never the kind of buyer or neighbour (no 'perfect for young families', 'ideal for singles', 'exclusive', 'safe neighbourhood', 'walking distance to church'); no school-quality claims; no words like 'stunning' or 'must-see' more than once. Finally list every claim I should double-check before publishing (square footage, years, HOA, anything from my notes that sounded uncertain).

Turn your local sales numbers into a plain-English monthly market update for past clients and your sphere

A monthly market email keeps past clients thinking of you, but generic national headlines don't help anyone. Paste your own local MLS or board statistics and the prompt writes a short, honest update: what changed, what it means for buyers and sellers, and one question to reply to.

Write my monthly market update email for [area/neighbourhoods]. Data for this month and the same month last year (from my MLS or board report, pasted as-is): [median sale price, number of sales, new listings, active listings, average days on market, sale-to-list ratio]. A few real examples I can mention (no client names): [e.g. '3-bed on Elm St got 4 offers in 6 days']. Audience: past clients and people I know, mostly homeowners. Write: a subject line under 50 characters, 3 numbers that changed most with the % change (calculate them and show your maths in a note to me, not in the email), what this means for someone thinking of selling and for someone thinking of buying (2 sentences each, no hype, no predictions about where prices will go), and a closing question that invites a reply. Under 250 words. If a number is missing or the change is tiny, say so rather than spinning it.

Plan a week of posts for a new listing, then a just-sold post, from one set of listing facts

Agents need several different posts for each listing (coming soon, just listed, open house, feature close-ups, just sold) and usually rewrite the same facts each time. This prompt produces the whole set at once, in your voice, with photo suggestions and the brokerage disclosure you add.

From these listing facts: [address or area, price, beds, baths, key features, open-house date/time, my name and brokerage, required brokerage disclosure text], write 7 posts for [Instagram / Facebook / LinkedIn]: 1) coming soon, 2) just listed, 3) open-house invitation, 4) two feature posts (pick the 2 most photogenic features and suggest which photo to use), 5) a neighbourhood post about nearby amenities (only ones I list here: [list]), 6) a just-sold template with blanks for the final details. Each under 120 words with up to 5 relevant hashtags. Describe the home, not the ideal buyer (no 'perfect for families' or similar). Keep the price and facts identical across posts, and end each with my disclosure line.

More in AgentPlaybook Pro

See Pro pricing →

Get the free Solo Founder Stack pack

32 copy-paste agent prompts (PDF) plus 5 new prompts every week.

Get it free

Guides that include Marketing prompts

Free AI agent prompts for small business (beginner-friendly)106 promptsChatGPT prompts for business email, replies and follow-ups47 promptsAI prompts to summarize meetings, notes, documents and reports39 promptsAI agent prompts for Zapier, Make and n8n automations16 promptsAI prompts for Google Sheets and Excel work44 promptsAI prompts for writing SOPs, reports, policies and checklists46 promptsChatGPT prompts for ecommerce and online stores (Shopify, Etsy, WooCommerce)13 promptsAI prompts for marketing agencies, consultants and freelancers14 promptsChatGPT prompts for real estate agents: listing descriptions, lead follow-up and open houses11 prompts

Other categories

Operations50 promptsCustomer Support35 promptsSales35 promptsFinance33 promptsPersonal Productivity26 promptsCoding24 promptsBusiness Strategy17 promptsEcommerce15 promptsHR & People14 promptsSecurity13 promptsIT & Internal Help12 promptsLegal & Compliance10 promptsContent8 promptsResearch8 promptsEducation3 prompts