Strategy. Judgment. Specificity. Compliance. Coordination across disciplines. The marketing tasks that AI now handles are real and they are significant, but they are the production layer, not the strategic one. Practices paying for marketing in 2026 are paying for the work that determines whether all that AI assisted output actually moves the right metrics, holds clinical authority, and protects the practice’s reputation. The output is faster than ever. The judgment behind it has never mattered more.
I’ve been watching this shift play out across the practices we work with for the last eighteen months, and I’ll tell you what I’m seeing honestly. The practices that misread the moment and tried to use AI to replace marketing investment are now further behind than they were two years ago. The practices that read the moment correctly are using AI to expand what their marketing investment can produce. Both groups spent roughly the same amount of money. The outcomes are not close.
What did Anthropic’s research actually find about AI and marketing work?
Anthropic, the company that builds the Claude AI model, published research in 2025 measuring what AI is actually being used for in real-world work, not what it could theoretically do. The data came from observed usage across the platform.
The findings relevant to marketing:
- AI is currently performing roughly 65% of the tasks done in market research and marketing roles in real-world use (Anthropic Economic Index, 2025).
- The exposure for marketing is among the highest of any occupation studied, comparable to computer programming, customer service, and data entry.
- The 65% figure represents observed usage, not theoretical capacity. Theoretical capacity is significantly higher.
- Adoption is accelerating. The gap between observed and theoretical exposure is closing as organizations build out AI-assisted workflows.
This is not speculation. It’s measurement. AI is doing significant marketing work right now, and the trajectory is clear.
What marketing tasks is AI actually doing well?
A specific list, based on what’s measurably working inside real marketing operations:
- First-draft content production. Blog posts, social posts, email copy, ad variations, and meta descriptions, drafted from briefs.
- Content optimization for search and AI citation. Structure, schema, FAQ generation, and citation-ready formatting.
- Variation generation. Multiple headlines, subject lines, opening paragraphs, and creative variations produced quickly for human selection.
- Research and source synthesis. Compressing reports, articles, and source material into working notes a strategist can build from.
- Data analysis at scale. Pattern recognition across analytics, ad performance, search behavior, and content performance.
- Production scaling. Producing graphics, mockups, and asset variations from defined templates.
- Workflow automation. Routine content distribution, scheduling, tagging, and reporting.
Each of these is a real, measurable productivity gain. None of them is the strategic work.
What marketing tasks does AI fail at, every time?
A specific list of what AI cannot do reliably without strong human direction:
- Strategic positioning. Deciding what a practice should stand for, who it serves, and how it differentiates. AI cannot make this call. It can only execute against a position someone else has set.
- Audience definition. Understanding the specific behavioral health populations a practice serves and the language that resonates with each of them.
- Clinical accuracy. Verifying that content claims about diagnoses, treatment outcomes, medications, or crisis content are clinically sound and current.
- HIPAA and compliance judgment. Knowing when content, photography, reviews, or testimonials cross PHI lines and when they don’t.
- Brand voice ownership. Holding a recognizable, specific voice consistent across thousands of pieces of content over time.
- Cross-channel coordination. Operating website, social, email, paid, and intake as a coordinated system rather than disconnected outputs.
- Real client and stakeholder relationships. Conducting interviews, building case studies, and drawing on lived experience inside a real practice.
- Strategic prioritization. Deciding what not to do, given limited time, budget, and attention.
These are the parts of marketing that determine whether the AI assisted output is actually working. Without them, AI produces volume. With them, AI produces leverage.
What is the strategic work practices are actually paying for in 2026?
Five categories. This is what real marketing investment looks like now:
| Strategic Work | What It Looks Like in Practice |
|---|---|
| Positioning and audience strategy | Defining what the practice stands for, who it serves specifically, and how it differentiates in its market. |
| Brand and voice ownership | Building and stewarding the visual identity, voice document, and content standards that carry across every AI assisted output. |
| Clinical and compliance review | Operating the workflow that verifies clinical accuracy, HIPAA compliance, and ethical standards on every piece of content. |
| Cross-channel coordination | Managing website, social, email, paid, search, and intake as a single coordinated system instead of disconnected channels. |
| Strategic measurement and iteration | Defining what success looks like, measuring against it, and adjusting strategy based on what’s actually working. |
This is the work AI does not do. It is also the work that determines whether everything AI does produces a return.
Why is this work harder to operate in-house than it used to be?
Because the disciplines required to operate marketing well in 2026 have multiplied, not consolidated. A coordinated marketing operation now requires:
- A brand strategist who can hold positioning and voice across an expanding content surface.
- A content lead who can operate AI assisted production at quality without losing voice or accuracy.
- A clinical reviewer who can verify behavioral health content for accuracy and compliance.
- A designer and visual strategist who can hold brand identity across six to ten platforms.
- A paid media operator who can run AI-augmented advertising without burning budget on the wrong audiences.
- An SEO and AIO specialist who can structure content for both traditional search and AI citation.
- A compliance reviewer who understands HIPAA and behavioral health marketing standards.
- An analytics lead who can connect activity to clinically meaningful outcomes.
I’ll be the first to tell you, no in-house team at a behavioral health practice should be running all of this. That’s not what the practice is built to do, and trying to staff it internally is one of the most common reasons practice owners burn out on marketing entirely. The capacity gap is real, and it’s gotten wider, not narrower, since AI tools became widely available.
What does this mean for your marketing investment in 2026?
It means the value of marketing investment has shifted decisively toward strategy and judgment, and away from raw production. AI handles the production. The work that protects the practice (the positioning, the brand, the compliance, the cross-channel coordination, the clinical accuracy, the strategic measurement) is more important than it has ever been, and it requires more specialized expertise than most practices realize.
The honest version of what we’re all paying for now: we’re paying for the people who make the AI assisted output actually work. The strategists, designers, writers, clinical reviewers, and compliance leads who turn fast production into compounding marketing investment. The output is the easy part. The judgment is the entire game.
That’s exactly the kind of cross-disciplinary work our team operates inside marketing strategy, branding, content marketing, SEO and AIO, and website design for behavioral health practices. If you’ve been reevaluating what your marketing investment should look like in an AI-assisted environment, let’s talk.
Frequently Asked Questions
How much marketing work is AI actually doing in 2026? AI is currently performing roughly 65% of the tasks done in market research and marketing roles in real-world use, according to Anthropic’s 2025 economic research. Adoption is accelerating, and the gap between observed and theoretical capacity is closing as organizations build out AI-assisted workflows.
What marketing tasks does AI handle well? First-draft content production, content optimization for search and AI citation, variation generation, research and source synthesis, data analysis at scale, production scaling for graphics and assets, and workflow automation. Each is a real productivity gain when used inside a strong strategic and editorial framework.
What marketing tasks does AI fail at? Strategic positioning, audience definition, clinical accuracy, HIPAA and compliance judgment, brand voice ownership, cross-channel coordination, real stakeholder relationships, and strategic prioritization. These are the parts of marketing that determine whether AI assisted output actually moves the right metrics.
What should a behavioral health practice be paying for in marketing in 2026? Strategy and judgment, primarily. Positioning and audience strategy, brand and voice ownership, clinical and compliance review, cross-channel coordination, and strategic measurement. AI handles the production. The strategic work determines whether the production is producing a return.
Should a behavioral health practice operate marketing in-house in 2026? Most can’t, and it usually isn’t a good use of the practice’s time even when they technically can. Operating marketing well now requires brand strategy, content production, clinical review, design, paid media, SEO and AIO, compliance, and analytics, often coordinated across multiple platforms. Practices that try to staff this internally typically end up with burned-out internal teams and inconsistent output. The capacity gap is the most common reason practices partner externally for this work.
If AI is doing 65% of the marketing tasks, what is the 35% your practice is most exposed on right now?
