The short answer
Point AI first at the places where a delay already costs you sales: replying to new enquiries, answering missed calls and chasing open quotes.
Content comes later. 75 percent of marketing teams now use some AI, yet BCG finds 60 percent of companies report minimal revenue and cost gains. Usage is not the problem. Where it is pointed is.
In this article 6 sections
The five lessons
Read these if you read nothing else.
- Rank every AI use by the revenue it moves, not by how impressive the demo is.
- Start with response: 69 percent of marketers say they struggle to reply to enquiries promptly (Salesforce, 4,450 marketers).
- Content is where most teams start (73.9 percent use AI for content creation) and the least likely place to move signed revenue first.
- Usage is high and payback is rare: 60 percent of companies report minimal gains (BCG, 1,250 executives, 2025).
- Set the guardrails before the tools: a person approves anything that spends money, contacts a customer or goes public.
How businesses use AI in marketing today, and what they get
The newest large studies, newest first. Survey means are what respondents report, not audited results.
| Finding | Sample | Source |
|---|---|---|
| 17 to 20 percent of US businesses use AI; 37 percent of firms with 250 or more employees | US Census Bureau business survey, December 2025 to May 2026 | US Census Bureau, May 2026 |
| CMOs put 15.3 percent of marketing budgets into AI, but only 30 percent have mature readiness to scale it | 401 CMOs and marketing leaders, January to March 2026 | Gartner CMO Spend Survey, May 2026 |
| 73.9 percent use AI for content creation, 65.4 percent for personalisation, 46.3 percent for data analysis and reporting | 188 US marketing leaders, January 2026 | The CMO Survey, 2026 |
| Generative AI is used in 22.42 percent of marketing activity on average | 187 US marketing leaders, January 2026 | The CMO Survey, 2026 |
| Reported gains: 14.64 percent lower marketing overhead and 14.06 percent higher sales productivity (means, self-reported) | 160 and 156 US marketing leaders, January 2026 | The CMO Survey, 2026 |
| 75 percent of marketing teams use at least one form of AI; 13 percent use agentic AI | 4,450 marketers, 26 countries, late 2025 | Salesforce, State of Marketing, 10th edition |
| 69 percent of marketers struggle to respond promptly to customer and prospect enquiries | 4,450 marketers, late 2025 | Salesforce, State of Marketing, 10th edition |
| 45 percent of martech leaders with AI agents say vendor-offered agents fall short of promised business performance | 413 martech leaders, June to August 2025 | Gartner, October 2025 |
| 60 percent of companies report minimal revenue and cost gains from AI; 5 percent are built to get value from it at scale | 1,250 senior executives, nine industries, 2025 | BCG, September 2025 |
| 27 percent of CMOs report limited or no generative AI in campaigns; 77 percent of adopters use it for creative work | 418 marketing leaders, July to September 2024 | Gartner, February 2025 |
- Read together: adoption is close to universal in marketing teams and still low across all US businesses, and the money mostly goes to creative work.
- The 14 percent gains are averages people report about themselves. Treat them as a ceiling for planning, not a promise.
Rank the uses by the revenue they move
Order matters. A faster reply makes every later step bigger, so fix the leak before you add volume. Each use below names the number it should move, so you can tell in four weeks whether it worked.
| Use, in order | What it fixes | The number to watch |
|---|---|---|
| 1. Instant reply to new web leads, and text back on missed calls | Enquiries that wait hours for a first answer | Median minutes to first reply; share of leads reached inside five minutes |
| 2. After-hours call answering with a named human backup | Calls that ring out at night and on weekends | After-hours calls captured; booked appointments from them |
| 3. Drafted follow-up on open quotes, sent by a person | Quotes that are sent once and never chased | Open quotes with a next action and a date; quote-to-sold rate |
| 4. A weekly report built from the CRM, not the ad platforms | Budget moved on opinion because nobody trusts the numbers | Signed revenue by source; cost per signed job |
| 5. Ad and landing page variations, reviewed before launch | Slow testing because every version takes a week | Cost per booked appointment, not cost per click |
| 6. Articles and posts for search and AI answers | Thin pages that never earn a citation | Enquiries from organic and AI search, confirmed in the CRM |
- Uses 1 to 3 recover demand you already paid for. They usually pay back first, because the enquiry already exists.
- Use 4 decides whether the rest worked. Without it, every channel claims the same sale.
- Uses 5 and 6 add volume. Add them once the path from enquiry to signed job holds.
Where AI does not pay
Five places where the cost lands on you, not on the tool.
- Brand-defining copy published without a human edit. One wrong claim costs more than a month of saved writing time.
- Decisions run on a CRM nobody trusts. AI repeats the error faster: clean the fields it reads first.
- Replacing a person on the phone with no human backup. Customers want a way to reach a person, and a dead end loses the job.
- Agents bought on a vendor's promise. 45 percent of martech leaders with agents say they fall short (Gartner, 2025).
- Tools bought before the process exists. Software is the smallest cost line; staff time and setup decide the bill. See what AI implementation costs.
The guardrails to set before the first tool
Write three rules into the plan and give each one an owner. They keep the speed and remove most of the risk.
- Approval gates: anything that spends money, contacts a customer in a new way or goes public waits for a named person.
- Your data stays yours: no customer data trains shared models, and every tool's access is logged and can be removed.
- One number per use, read every week. A use that does not move its number in four weeks is switched off.
What to check this week
Pull last week's new enquiries from your CRM and phone system. Count how many waited more than five minutes for a reply from a person, and how many calls went to voicemail.
If either number is high, start with use 1. It is the cheapest test with the clearest result. The lead response time statistics show what those minutes cost, and AI for lead generation covers the setup.
Questions leaders ask
How should a small business use AI for marketing?
Start where slow replies already cost sales: instant replies to web leads, text back on missed calls and drafted quote follow-up. Measure the reply time and booked appointments for four weeks before adding content or ad tools.
What percentage of marketers use AI?
75 percent of marketing teams use at least one form of AI (Salesforce, 4,450 marketers, late 2025). In US marketing leaders, 73.9 percent use it for content creation (The CMO Survey, 2026). Across all US businesses, use is 17 to 20 percent (US Census Bureau, 2026).
Does AI in marketing increase revenue?
Sometimes. US marketing leaders report a 14.06 percent average gain in sales productivity (self-reported, The CMO Survey 2026), but BCG finds 60 percent of companies report minimal revenue and cost gains (2025). The gain follows where AI is pointed: response and follow-up before content.
What should AI not do in marketing?
Publish brand-defining copy without review, run decisions on untrusted CRM data, replace a person on the phone with no human backup, or spend money and contact customers without a named person's approval.
How much of the marketing budget goes to AI?
15.3 percent on average among CMOs in Gartner's 2026 CMO Spend Survey of 401 marketing leaders, with only 30 percent reporting mature readiness to scale it.
SOURCES
- US Census Bureau, Large Firms With at Least 20 Employees Biggest AI Users (May 2026): 17 to 20 percent of US businesses use AI; 37 percent of firms with 250+ employees
- Gartner, 2026 CMO Spend Survey (May 2026): 401 CMOs; 15.3 percent of marketing budgets to AI; 30 percent mature readiness
- The CMO Survey, 35th edition, Topline Report (Duke Fuqua, Deloitte, AMA, 2026): 308 US marketing leaders
- Salesforce, State of Marketing, 10th edition (fielded late 2025): 4,450 marketers, 26 countries
- Gartner (October 2025): 413 martech leaders; 45 percent say vendor AI agents fall short
- BCG, AI Leaders Outpace Laggards (September 2025): 1,250 senior executives; 60 percent report minimal gains
- Gartner (February 2025): 418 marketing leaders; 27 percent limited or no GenAI in campaigns
Where’s the money leaking?
Pick one. Be honest.
Cite this article
Grigorchuk, T. (2026, October 17). How to Use AI for Marketing: Rank Each Use by the Revenue It Moves. Megawebvision. https://megawebvision.com/insights/how-to-use-ai-for-marketing