TL;DR

McKinsey's June 2023 report "The economic potential of generative AI" estimated that 63 use cases could add $2.6 trillion to $4.4 trillion a year, with about 75% of that value in four functions: customer operations, marketing and sales, software engineering, and R&D. For marketing it estimated productivity value of 5–15% of total marketing spend, and 3–5% of sales spend for sales. For SMEs, the practical lesson is to concentrate on a few high-volume, measurable marketing and sales workflows instead of scattering pilots.

What McKinsey's report says

"The economic potential of generative AI: The next productivity frontier" was published on 14 June 2023 by McKinsey & Company (also listed in its report overview). Its authors include Michael Chui, Lareina Yee and Alex Singla. It remains one of the most cited attempts to size the value of generative AI by business function.

The headline numbers, all based on the 2022 global economy:

  • Across 63 use cases in 16 business functions, generative AI could add $2.6 trillion to $4.4 trillion in value annually.
  • Including wider productivity effects on knowledge work, total benefits rise to $6.1 trillion to $7.9 trillion a year.
  • About 75% of the use-case value falls in four areas: customer operations, marketing and sales, software engineering, and R&D.
  • Current generative AI and other technologies could, in principle, automate work activities that take up 60–70% of employees' time, up from around 50% in earlier estimates.

Where marketing and sales fit

The report expresses the marketing and sales opportunity as a share of spending rather than as a single dollar figure:

  • Marketing: productivity value of 5–15% of total marketing spending.
  • Sales: roughly 3–5% of current global sales expenditure.

The marketing use cases it describes include faster ideation and drafting, keeping a consistent brand voice and format, personalising campaigns for segments (including translation), drawing insight from unstructured data such as reviews, SEO support for page titles, image tags and URLs, and personalised product discovery and search. It specifically mentions generative AI "producing first drafts of brand advertising, headlines, slogans" and similar content. On the sales side, it points to improving lead development and raising the probability of a sale.

Two caveats from the report itself matter here. First, the marketing and sales estimates capture only direct productivity effects, not knock-on gains such as revenue growth. Second, it warns that models trained on public data can create intellectual property and brand risks, and that significant human oversight is needed for strategy specific to each company.

Use case value

in McKinsey's method, the estimated productivity gain from applying generative AI to a specific business activity, expressed as a share of what that function currently costs.

Why this still matters in 2026

A 2023 forecast is not a result, and much of the early hype has been tested since. McKinsey's own State of AI survey published in August 2026 (1,719 respondents) found that nearly nine in ten organisations report regular AI use in at least one function, yet only 37% attribute any EBIT impact to AI. Revenue gains were most often attributed to marketing and sales. A small group of "high performers" (6% of respondents) stood out, and nearly three-quarters of them had fundamentally redesigned workflows, compared with about a quarter of other respondents.

Read together, the two reports point the same way: marketing and sales is where generative AI value is most likely to show up, but value comes from changing how work is done, not from giving everyone a chatbot licence.

A playbook: how SMEs should prioritise use cases

SMEs cannot run 63 use cases. The task is to pick the two or three that match where McKinsey sees value and where your business has volume. In our work with European SMEs, the following sequence keeps effort focused:

  1. List your marketing and sales activities by volume. How many product descriptions, ads, emails, quotes, translations and customer questions do you produce a month? Value scales with repetition.
  2. Score each on three questions. Is it high volume? Is the quality bar clear enough to check quickly? Is the risk of an error low or easy to catch? Prioritise activities that score well on all three.
  3. Start with content production and localisation. First drafts, variations for testing, and translation for multiple markets map directly onto the report's marketing use cases and are easy to measure in hours saved.
  4. Then move to insight from unstructured data. Summarising reviews, support tickets and sales call notes can sharpen messaging and is low risk because the output is internal.
  5. Treat personalisation and sales support as phase two. They depend on clean customer data and integrations, so they cost more to do well.
  6. Redesign the workflow, not just the task. Decide who briefs, who reviews and who approves, and remove steps that AI makes redundant. This is the behaviour the 2026 survey links with high performers.
  7. Measure against spend. McKinsey frames marketing value as a share of marketing spend, so track cost per asset, time to launch and agency or freelance spend before and after.

Limits and common misreadings

The trillion-dollar figures are global, economy-wide estimates of potential, not a forecast for any single company. They exclude entirely new products, and McKinsey excluded use cases whose only benefit was natural-language ability. They also assume workers are redeployed into productive work, which depends on management choices.

A common misreading is to take "5–15% of marketing spend" as a guaranteed saving. It is an estimate of potential productivity value across the economy, and the 2026 survey shows most organisations have not yet captured enterprise-level gains. Another is to assume that because marketing is a top function, every marketing activity benefits equally. The report's value comes from specific, high-volume activities, and so should your priorities.

Key takeaways
  • McKinsey (2023) sized generative AI at $2.6–4.4 trillion a year across 63 use cases, based on the 2022 economy.
  • About 75% of that value sits in customer operations, marketing and sales, software engineering and R&D.
  • Marketing productivity value was estimated at 5–15% of marketing spend; sales at 3–5% of sales spend.
  • SMEs should focus on a few high-volume, checkable workflows and redesign how that work gets done.

FAQ

Does McKinsey give a dollar figure for marketing and sales?
The report's article text gives marketing and sales value as percentages of spending (5–15% for marketing, 3–5% for sales) and puts the function among the four that account for about 75% of total use-case value. We have not quoted a separate dollar figure here.
Is a 2023 report still useful?
Its function-level map of where value concentrates remains a useful starting point. Treat the figures as 2023 estimates of potential, and check them against more recent evidence such as McKinsey's 2026 State of AI survey.
What is the first use case an SME marketing team should try?
Usually content drafting and localisation for a high-volume asset type, such as product descriptions or email campaigns, because it is easy to measure and to review before publication.

Sources

Find your highest-value AI use cases

JRK can help you score your marketing and sales activities and choose the two or three generative AI workflows worth building first.

Book a free growth session →

Want this working for your brand?

Book a free 30-minute growth session. We'll look at your funnel and tell you what we'd do first — honestly.

Book a growth session →