AI prompts for Demand Generation Marketers

For demand generation marketers. Know who every campaign is for, test ads small and show up on Google. Turn the people who engage into warm leads, and make sure every lead gets a reply in 5 minutes.

7 tested prompts and 5 guides. Free. Copy them into Claude.

7 tested prompts for Demand Generation Marketers

1 of 7

Check your prices against competitors

Anyone who runs marketing, before a price change or when sales asks how your prices compare.

Claude

What to paste in

  • Your prices plus the competitor prices you found, with where and when you found each one.
PromptPaste into Claude
I do marketing for a [type of business] in [town or area]. Here are our prices: [paste each item or service with our price, size and any fees].

Here are competitor prices I found: [paste each competitor's name, items, prices, sizes, fees and deals, with where and when you found each price, such as their website, price list, flyer or a visit]. Add how each one sells, if you know, such as loyalty deals, delivery, online booking, an app or sales reps.

1. Make a table: item, our price, and each competitor's price. Only match items that are truly alike. Mark a match "close" if the items are a bit different. If sizes differ, compare price per unit and show that math.
2. For each match, show how much more or less we charge, in dollars and as a percent of their price.
3. List where we charge more by over [5]%, where we charge less, and where we are about the same.
4. Sum up how each competitor sells, in 1 or 2 lines each.
5. Flag any price older than [3] months, and any deal that ends soon or has fine print.
6. Give 2 or 3 ideas we could test, based only on my notes.

Use plain words. Show your math. Only use what I pasted. If a price is missing, leave the cell blank and say so. Do not guess.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Gather data on competitors and analyze their prices, sales, and method of marketing and distribution.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

All names and prices are made up. Today is October 2026. Us, Bright Lane Car Wash: Basic wash $12. Wash and wax $20. Inside clean $35. Unlimited washes $30 a month. Suds Corner, from their website, Oct 2026: Basic wash $10. Wash and wax $22. Unlimited $25 a month. Sells through an app, first wash free. Shinewell Auto Spa, from a flyer, April 2026: Basic wash $14. Full detail inside and out $45. No monthly plan. Ads on local radio. Flag prices older than 3 months.

What Claude wrote

Item | Us | Suds Corner | Shinewell Basic wash | $12 | $10 | $14 Wash and wax | $20 | $22 | none Inside clean | $35 | none | $45 full detail (close) Unlimited | $30 a month | $25 a month | none Gaps (gap / their price): Basic wash: $2 over Suds ($12 − $10), 2/10 = 20.0%. $2 under Shinewell, 2/14 = 14.3%. Wash and wax: $2 under Suds, 2/22 = 9.1%. Inside clean: $10 under Shinewell (close), 10/45 = 22.2%. Unlimited: $5 over Suds, 5/25 = 20.0%. Over 5% more: basic wash and unlimited, vs Suds. Less: the rest. Same: none. Suds: app, first wash free. Shinewell: radio ads. Old: Shinewell prices are from April 2026. Recheck them. No fine print found. Ideas: 1. Lead ads with wash and wax, $2 under Suds. 2. Push our unlimited plan. Shinewell has none.

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Recheck each competitor price at its source and date before you change yours or share the table.

2 of 7

Compare what competitors offer and promise

Anyone who runs marketing, before new ads, a new web page or a new service.

Claude

What to paste in

  • Text copied from your website and from 2 to 4 competitor websites or ads.
PromptPaste into Claude
I do marketing for a [type of business] in [city or area]. Here is text from our website or ads: [paste]. Here is text I copied from 2 to 4 competitor websites, ads or public price pages: [paste each one, with the company name on top].

1. Make a table with one row per company, including us. Columns: what they offer, prices if shown, who they sell to, area served, main promise, and how people buy or reach them (phone, form, online booking, store, reps, ads or social media).
2. Write "not shown" when the text does not say. Do not fill gaps with guesses.
3. Under the table, say where we look the same, where we look stronger, and where we look weaker.
4. Give 2 or 3 things we could say in our marketing that no competitor says, based only on this text.
5. Name one thing a competitor does that we should not copy, and say why.

Use plain words. Only use what I pasted. Do not guess prices or sales numbers.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Gather data on competitors and analyze their prices, sales, and method of marketing and distribution.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

All names are made up. Us, Maple Note Music Studio: Piano, guitar and voice lessons for kids and adults. 30 minute lessons $35. First lesson free. Book online at maplenote.example. Two recitals a year. Riverbend Music School: "Piano and violin for ages 5 to 12. Call for prices. Learn from pros." Phone only. Strum Shack: "Guitar lessons, $30 for 30 minutes. Rock band program for teens. Follow us on Instagram." Sign up in store.

What Claude wrote

Company | Offers | Price | Sells to | Promise | How to buy Maple Note (us) | Piano, guitar, voice, recitals | $35 per 30 min, first lesson free | Kids, adults | not shown | Online Riverbend | Piano, violin | not shown | Ages 5 to 12 | "Learn from pros" | Phone Strum Shack | Guitar, teen band | $30 per 30 min | Teens for band, rest not shown | not shown | Store, Instagram Area: not shown for anyone. Same: piano like Riverbend, guitar like Strum Shack. Stronger: only we list adults, voice, a free first lesson, online booking and recitals. Weaker: $5 more than Strum Shack ($35 − $30). No promise line. No violin or teen band. Things to say: "Your first lesson is free." "Lessons for adults too." "Book online." Do not copy: Riverbend hides its prices, so people must call.

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Open each competitor's page again to make sure the claims are still true before you use them in an ad.

3 of 7

Sort customer comments into themes

Anyone who runs marketing or customer care, when a batch of reviews or survey comments comes in.

Claude

What to paste in

  • Reviews or survey comments, one per line, with names and private details taken out.
PromptPaste into Claude
Here are customer comments from [where they came from, such as online reviews, a survey or comment cards] for our [type of business], from [time period]: [paste comments, one per line, with names, contact details and any private details taken out]. If there are scores, include them.

1. Group the comments into themes, such as price, quality, service, wait times or billing. Give each theme a short, plain name.
2. For each theme, give how many comments fit it and their numbers, say if it is mostly good, mostly bad or mixed, and quote one short line word for word.
3. If a comment fits two themes, say so, and say what the counts add up to.
4. If there are scores, give the average and show the math.
5. List the top 2 or 3 things to fix, with one small fix to try in the next 30 days.
6. List the top 2 things customers like that we could talk about in our marketing.
7. List any comment that sounds like a legal or safety risk, such as a charge they did not agree to or someone getting hurt, so I can send it to the right person.

Only count what is in the comments. Do not add themes that no comment supports. Do not guess at numbers. If a theme has fewer than 3 comments, say the sample is too small to judge. Use plain words and short lines.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Measure and assess customer and employee satisfaction.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

Ledgeway Climbing Gym (made up), October member comments, names taken out: 1. Staff are friendly and helped me pick shoes. 2. Way too crowded after 5 pm. 3. Got charged for a month after I froze my membership. 4. New routes every week keep it fun. 5. Could not find parking again. 6. A hold spun on the wall and I almost fell. 7. Crowded on weekends. Waited 20 minutes for a wall. 8. Coaches are great with my kids. 9. The 6 pm rush is too much. No room to warm up. 10. Love the friendly front desk.

What Claude wrote

Crowding: 3 (2, 7, 9). Mostly bad. "Way too crowded after 5 pm." Staff and coaches: 3 (1, 8, 10). Mostly good. "Coaches are great with my kids." Too small to judge, 1 comment each: Billing (3). Bad. "Got charged for a month after I froze my membership." New routes (4). Good. "New routes every week keep it fun." Parking (5). Bad. "Could not find parking again." Wall safety (6). Bad. "A hold spun on the wall and I almost fell." Each comment fits one theme. Total 10. No scores. Fix first: 1. Crowding. Post the busy hours so members can plan. 2. Wall safety. Check holds before opening each day. 3. Billing. Check frozen accounts before billing day. Talk about: friendly staff and coaches, new routes every week. Send to the right person: comment 3 (charged after a freeze) and comment 6 (a hold spun).

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Take out names and private details, use an AI tool your workplace allows, and recount two themes by hand.

4 of 7

Find which marketing brings in customers

Owners or marketing leads, at the end of each month or quarter, before setting the next budget.

Claude

What to paste in

  • Spend, leads and new customers for each channel, all for the same dates.
PromptPaste into Claude
Here are our marketing results for our [type of business] from [dates]: [paste each channel or lead source with money spent, leads (calls, forms, bookings or messages) and new customers or signed jobs]. Our goal was [goal, if you have one].

1. For each channel, work out cost per lead, cost per new customer, and the percent of leads that became customers. Show your math in a table.
2. Rank the channels from best to worst on cost per new customer.
3. Point out any channel with spend but no customers, and any channel with too few results to judge.
4. Say how we did against our goal, if I gave one.
5. Give 2 or 3 plain next steps for the next month or quarter, such as where to move money or a small test, and the number that would tell us a test worked.
6. Say what data is missing that could change the answer.

If these are job ads, count hires instead of customers. Use only my numbers. If a number is zero or missing, say so and do not fill it in. If a number looks wrong, flag it.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Measure the effectiveness of marketing, advertising, and communications programs and strategies.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

Pawsome Suds Dog Grooming (made up). Made up numbers, July 1 to Sept 30. Goal: 40 new customers. Search ads: spent $900, 45 calls or bookings, 18 new customers Instagram boosted posts: spent $450, 30 messages, 6 new customers Flyers at a vet office: spent $200, 8 calls, 4 new customers Referral card, $10 off for both people: $160 in discounts, 10 calls, 8 new customers Local magazine ad: spent $600, 4 calls, 0 new customers

What Claude wrote

Channel | Per lead | Per customer | Leads that bought Referral card | $160/10 = $16 | $160/8 = $20 | 80% Search ads | $900/45 = $20 | $900/18 = $50 | 40% Vet flyers | $200/8 = $25 | $200/4 = $50 | 50% Instagram | $450/30 = $15 | $450/6 = $75 | 20% Magazine | $600/4 = $150 | none | 0% Rank: referral card, then search ads and vet flyers tied at $50, then Instagram, magazine. Spend, no customers: magazine, $600. Too few to judge: vet flyers (8 leads), referral card (10), magazine (4). Goal: 36 of 40. Short by 4 (40 − 36). Next: 1. Give every customer a referral card. It worked if referrals beat 8 next quarter. 2. Move the magazine's $600 to search ads. Check cost per customer stays near $50. Missing: repeat visits and spend per customer.

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Redo two lines of math by hand, and make sure every channel covers the same dates and counts each customer once.

5 of 7

Check if a sale or promo paid off

Anyone who runs marketing, a week or two after a sale, discount or ad push ends.

Claude

What to paste in

  • Sales from before, during and after the promo, plus what it cost and your usual profit margin.
PromptPaste into Claude
We ran this promo at our [type of business]: [what it was, the dates, the discount, and where we posted it, such as email, social posts, signs or flyers].

Here are the numbers: [paste sales and customer counts for the same number of days or weeks before, during and after the promo, the promo cost, promo code uses if you have them, and our usual profit margin if you know it]. Notes on weather, holidays or local events: [notes].

Tell me in plain words:
1. Did customers and sales go up during the promo? By how much, in numbers and percent?
2. Did the average sale per customer change?
3. After the discount and costs, did we likely make or lose money? Show the math.
4. Did sales dip after the promo ended?
5. What else could explain the change, using only my notes?
6. A plain verdict: worked, did not work, or too early to tell.
7. Two things to track next time.

Compare like with like, such as the same weekdays. Only use my numbers. If something is missing, tell me what to find. Do not make up numbers.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Measure the effectiveness of marketing, advertising, and communications programs and strategies.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

Lockbox Escape Rooms (made up). Made up numbers. Promo: 25% off weeknight bookings, Oct 6 to Oct 12. Posted on social media and in one email. Ad cost: $150. Weeknights only (Mon to Thu): Week before: 20 bookings, $2,400 Promo week: 34 bookings, $3,060 (after discount) Week after: 18 bookings, $2,160 Usual profit margin before discount: 70%. Notes: local schools were on fall break during the promo week.

What Claude wrote

1. Yes. Bookings: 20 to 34, up 14. 14/20 = 70% more. Sales: $2,400 to $3,060, up $660. $660/$2,400 = 27.5% more. 2. Yes. $2,400/20 = $120 per booking before. $3,060/34 = $90 during. Down $30, or 25%. 3. Full price value: $3,060/0.75 = $4,080. Costs: $4,080 x 30% = $1,224. Profit: $3,060 − $1,224 − $150 = $1,686. Normal week: $2,400 x 70% = $1,680. Gain: $6. 4. Yes. The week after fell $2,400 − $2,160 = $240, or 10%. Lost profit: $240 x 70% = $168. Two weeks net: $6 − $168 = −$162. 5. Fall break may have brought in more families that week. 6. Verdict: did not work. Bookings rose, but profit over the two weeks fell about $162. 7. Track new versus repeat customers, and promo code uses. Find: any extra staff cost from the busy week.

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Check that sales are after discounts and returns, that the days match, and that your margin is right.

6 of 7

Write a short customer survey

Anyone who runs marketing, before a new product or price change, or when sales slow down.

Claude

What to paste in

  • What you want to learn, who will answer, and how you will send it.
PromptPaste into Claude
I do marketing for a [type of business]. I want to learn: [what you want to find out, such as why people stop buying or what they want next]. Who will answer: [who, such as current customers or past customers]. How we will ask: [email, a QR code on the receipt, a link on the invoice, or in person].

Write a survey that takes under [3] minutes:
1. 5 to 7 questions, in plain words a busy person gets in one read.
2. Easy questions first.
3. Mostly multiple choice or 1 to 5 scales, with 1 or 2 open questions at the end.
4. One idea per question. No leading questions.
5. Add an "Other" choice where the real answer might be missing.
6. Do not ask for names, birth dates, phone numbers, account numbers or other private details unless I say I need them.

Then tell me:
1. What each question will help us decide.
2. The best time to ask, and one way to get more people to answer.
3. One way the results could be wrong, like who might skip the survey, and how to fix it.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Devise and evaluate methods and procedures for collecting data, such as surveys, opinion polls, or questionnaires, or arrange to obtain existing data.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

Tallyleaf (made up), a scheduling app for dog walkers. Want to learn: why people stop after the 14 day free trial. Who: people who tried it and did not pay. How: one email with a link to tallyleaf.example/survey. Must take under 3 minutes.

What Claude wrote

1. How did you hear about Tallyleaf? (Search, a friend, social media, other) 2. How often did you use it in the trial? (Every day, a few times, once, never) 3. How easy was setup? (1 = very hard, 5 = very easy) 4. Why did you stop? Pick one. (Price, missing something, hard to use, went back to my old way, other) 5. What would bring you back? Pick up to two. (Lower price, longer trial, setup help, nothing, other) 6. What is one thing you wish Tallyleaf did? Decides: Q1, where users come from. Q2 and Q3, if they got started. Q4 and Q5, what to fix. Q6, what to build. Best time: a few days after the trial ends. More answers: say "6 quick questions" in the subject. Could be wrong: people who barely tried it may skip it. Fix: one reminder.

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Try the survey on one or two friendly customers, and check it against your privacy rules before it goes out.

7 of 7

Turn survey results into a short report

Anyone who runs marketing, when the owner or a boss wants the main points from a survey or data table.

Claude

What to paste in

  • Your survey results or data table, plus the question your boss asked.
PromptPaste into Claude
I do marketing for a [type of business]. Here is my data: [paste survey results or a data table]. My boss asked: [paste the question].

Write a one page report in plain words:
1. The main answer in two sentences.
2. Three to five findings, each with the number that backs it up.
3. One or two charts I should make. Say the chart type, what goes on each axis, and why.
4. What the data does not tell us.

Use only my numbers. Check your math and show percents. Do not add trends or causes the data does not show.

Built on this O*NET task for Market Research Analysts and Marketing Specialists: “Prepare reports of findings, illustrating data graphically and translating complex findings into written text.” O*NET, U.S. Department of Labor, CC BY 4.0.

See a real example

What went in

Brightpath Tutoring (made up). Made up numbers. Boss asked: Why do families leave after the first term? Survey of 50 families who left. Main reason (pick one): schedule did not fit 18, price 12, child did not improve 8, moved away 7, other 5. What would bring you back (pick up to two, all 50): weekend times 22, lower price 15, progress reports 14, online sessions 13, nothing 6.

What Claude wrote

Main answer: Schedule is the top reason families leave. 18 of 50 (36%) said our times did not fit, and weekend times were the top pick to bring them back (22 of 50, 44%). Findings: 1. Reasons: schedule 18 (36%), price 12 (24%), child did not improve 8 (16%), moved away 7 (14%), other 5 (10%). 2. Weekend times: picked by 22 of 50 (44%). 3. Lower price: 15 of 50 (30%). 4. Progress reports: 14 (28%). Online sessions: 13 (26%). 5. 6 of 50 (12%) said nothing would bring them back. Charts: 1. Bar chart of reasons. Reasons on one axis, families on the other. Shows schedule leads. 2. Bar chart of what would bring them back, same layout. Shows weekend times lead. Not told: how many families stayed, so we do not know what share of all families leave.

Tested on Claude Opus 5.5, October 2026

Check before you use it

  • Check every number and percent against your raw data before your boss sees it.

Step by step guides

Tips from Demand Generation Marketers

Share a tip for Demand Generation Marketers

New AI tools for this job

Don't see your work here?

Tell me what you do every day, and I'll make guides and prompts for it.

Got a prompt that works? Add it here.

Want this built for you? Get it built.