Which Marketing Functions Are Next on the Automation-Eats-Execution Curve?

Which marketing functions are next on the automation-eats-execution curve?

TL;DR: Three domains have visibly shown the glossary/automation-eats-execution bifurcation pattern in 2026: paid media (Seufert framing), influencer marketing (Modash empirical data), and software development (Karpathy / vibe coding). Several other marketing functions look like candidates but the evidence isn’t yet conclusive. This page tracks working hypotheses for email/CRM/lifecycle, SEO/content, brand-building, B2B sales-marketing, and analytics — what we’d expect to see if each were on the curve, and what would constitute evidence either way.

Why this question

The comparisons/strategy-vs-execution-ai synthesis page named a cross-domain pattern: AI tooling commoditizes high-volume execution work first; strategy, judgment, and integration work stays human-leveraged. That framing was anchored by three independent data points across three domains. The natural follow-up: what does the framing predict for adjacent domains, and where would we look for evidence?

Answering this matters because the framework’s value compounds as more domains either confirm or refute it. If the pattern shows up in a fourth and fifth domain, “automation eats execution” becomes a general principle rather than a three-domain coincidence. If it fails to show up in a domain where we’d expect it, the boundary conditions get clearer — also useful.

How to evaluate a candidate domain

Three signals indicate a domain is on the curve (per the framework):

There’s a high-volume, structured execution layer that AI tooling is currently compressing.
Strategic, judgment-heavy work in the domain is well-defined and senior in the org chart.
Salary or job-market data shows compensation premium for strategy ownership versus execution ownership.

Below: working hypotheses for five candidate domains, with the signal-status for each and what observational evidence would settle the question.

Candidate 1: Email / CRM / Lifecycle Marketing

Working hypothesis: substantially on the curve, but the data is fragmented.

The execution layer is real and substantial: segment design, email copy variation, send-time optimization, A/B test setup, campaign orchestration across channels. AI tooling in this space (Klaviyo’s AI features, HubSpot Breeze, Iterable’s AI Optimize, Salesforce Einstein) has been shipping for 2-3 years and is mainstream in 2026.

The strategic layer is also clearly defined: lifecycle architecture, retention thesis, audience-segmentation logic, attribution model design. Senior CRM strategists exist as a distinct role from email marketing operators.

What we’d expect to see if it’s on the curve: salary data showing strategy roles (Lifecycle Director, CRM Strategist, Retention Architect) commanding a meaningful premium over email-marketing-operator roles. Job postings tilting toward strategic and away from execution-only.

What would constitute evidence: a study analogous to Modash’s salary survey, but for CRM/email roles. Or a public salary database (Glassdoor / Levels.fyi / Built In) cross-tabulated by role title and AI-tool usage. We don’t have this data yet.

Open sub-question: Is the email/CRM execution layer easier or harder to automate than influencer marketing’s? The intuition is easier (more structured input/output, longer-established AI tooling), which would predict a sharper bifurcation in this domain — but we lack the empirical anchor.

Candidate 2: Organic Content / SEO

Working hypothesis: partially on the curve, but the framing is moving fast and the picture isn’t settled.

The execution layer is real: content production at scale, on-page SEO, internal linking, schema markup, content variation across surfaces. AI tooling for SEO content is mature (the wiki has documented this in seo/agentic-search-optimization and seo/geo-aeo-benchmarks-2026). AI-assisted content production has been mainstream since 2023.

The strategic layer is also real: topical-authority architecture, glossary/super-niche selection, glossary/topical-authority mapping, GEO/AEO citation strategy.

The complication: the target of optimization is itself moving. SEO in 2026 is partly Google ranking + partly LLM citation. The execution work for “rank in Google” is more automatable than the execution work for “be cited by ChatGPT and Perplexity” — and the latter is still partly judgment-heavy (which sources to publish on, which schema to mark up, which AI assistants to test against).

What we’d expect to see if SEO is fully on the curve: agency pricing migration from “we’ll write 100 articles for you” (execution volume) to “we’ll architect your topical authority” (strategic outcome). Some of this is happening; some isn’t.

What would constitute evidence: agency-pricing benchmark data showing strategic-outcome pricing growing faster than per-article pricing. Or salary data showing GEO/AEO-strategist roles commanding a premium over content-writer roles.

Open sub-question: Is the strategic layer in SEO actually scarce, or is it just not yet productized? The risk is that a fourth-generation AI tooling cycle (agentic SEO planners) automates parts of the strategic layer that are currently human-only.

June 2026 update — a small piece of evidence that the SEO strategic layer is genuinely scarce, not just unproductized. A wiki source-calibration pass found that the SEO domain’s headline statistics were almost entirely single-vendor numbers, and that distinguishing what a primary institution actually measured (Pew’s AI-Overview clickstream) from vendor estimates (the 64.82% / 96% / 3× figures) was a non-trivial judgment call — see seo/zero-click-strategy § calibration. Two strategic-layer tasks surfaced that current AI tooling does not do: (1) adjudicating evidence quality — knowing that a confidently-cited “96% of AI citations come from strong-E-E-A-T sources” is a vendor artifact, not a measured fact; (2) deciding where to earn third-party authority — the earned-media bias in AI search means the GEO strategic move is choosing which authoritative publications to be cited by, a market-judgment task. Both are judgment under low-validity conditions (the RPD/jagged-frontier boundary), which is exactly the kind of work the framework predicts stays human-leveraged. Weak evidence — one domain, observational — but it tilts the sub-question toward “scarce,” not merely “not yet productized.”

Candidate 3: Brand-Building (Sharp’s framework)

Working hypothesis: explicitly NOT on the curve, despite surface appearances.

This is the cleanest “no” in the candidate list. Per marketing/brand-vs-content-layers and Sharp’s How Brands Grow, brand-building works through years-long mental-availability investment — consistent distinctive-asset deployment, broad reach, frequency. AI tools improve production efficiency for individual brand-touchpoints, but they don’t compress the underlying mechanism. Building glossary/mental-availability in 2026 still requires the same multi-year discipline it required in 2010.

The work involves judgment per-decision (asset system design, consistency rules, long-horizon investment thesis), but the cumulative output over time is what matters — not the speed or volume of any single execution.

What we’d expect to see if brand-building were on the curve (it’s not): senior brand-strategist compensation flatlining while execution roles compress. Empirically: senior brand-strategist compensation is holding or rising, not flatlining. The role’s market value isn’t being compressed by AI.

This is useful negative evidence for the framework. “Automation eats execution” doesn’t predict that AI eats all marketing work — it predicts that AI eats execution work in domains with a clear execution-layer / strategic-layer split. Brand-building doesn’t have that split in the same shape; the prediction holds.

Candidate 4: B2B Sales-Marketing (ABM, demand gen, sales enablement)

Working hypothesis: weakly on the curve. The execution layer is real but smaller; the strategic layer dominates more in B2B than in DTC.

The execution layer exists: account research, outbound sequence personalization, lead-scoring rules, content-asset production, sales-enablement collateral. AI tooling has flooded this space (Apollo, Outreach, Gong, Clari, Salesloft).

The strategic layer is unusually dominant: ICP definition, account-strategy-by-tier, deal-coaching, multi-stakeholder navigation, channel-partner architecture. The judgment-per-account is large compared to DTC; the volume-per-week is smaller.

What we’d expect to see if B2B is partially on the curve: SDR/BDR roles compressing or being augmented heavily; AE compensation holding or rising; CRO/Head-of-Sales compensation rising more steeply.

What would constitute evidence: SDR/BDR labor-market data showing role compression or AI-augmentation as the dominant trend. Anecdotal evidence from 2024-2026 hiring patterns suggests SDR roles are indeed getting AI-augmented and partially compressed (tooling has clearly improved); the salary-premium-for-strategy story isn’t as cleanly visible as in Modash’s influencer data.

Open sub-question: Does the framework apply differently to roles where the execution layer is itself relationship work (cold outreach to a specific human, building trust over a 6-month sales cycle)? The intuition is that relationship execution is harder to automate than creative or content execution, so B2B would be less on the curve than DTC.

Candidate 5: Marketing Analytics / Attribution / Insights

Working hypothesis: substantially on the curve, with an interesting twist — the analyst role itself is bifurcating.

The execution layer is huge: data extraction, dashboard building, regular performance reports, ad-hoc data pulls, basic SQL queries. AI tooling for marketing analytics (text-to-SQL agents, dashboard generators, attribution platforms with built-in AI insights) has shipped massively in 2025-2026.

The strategic layer is also substantial: choosing which metrics matter, designing measurement frameworks, attribution-model architecture, hypothesis-driven analysis design.

The bifurcation is showing up in 2026 hiring patterns: junior data-analyst roles (Tableau monkey, SQL puller, dashboard builder) are getting compressed; senior analytics-strategist roles (measurement architect, growth analyst, head of analytics) are commanding premium compensation. Anecdotal but consistent across multiple firms.

What would constitute evidence: salary data showing senior analytics roles commanding $20K+ premium over junior data-analyst roles, with the gap widening 2024-2026. We don’t have this data formally documented yet but the labor-market signal is strong.

Synthesis: scoring the candidates on the three signals

Operationalizing the three signals from “How to evaluate a candidate domain” turns the prose hypotheses into a comparable matrix. Each domain is scored on: (1) is the execution layer high-volume, structured, and AI-compressible; (2) is the strategic layer well-defined and senior in the org; (3) is there salary/market evidence of a strategy-over-execution premium.

Domain	(1) Execution compressible	(2) Strategic layer defined	(3) Market-premium evidence	Verdict
Paid media (anchor)	High	High	Confirmed (Seufert)	✅ On curve
Influencer marketing (anchor)	High	High	Confirmed (Modash data)	✅ On curve
Software dev (anchor)	High	High	Confirmed (Karpathy + market)	✅ On curve
Marketing analytics	High	High	Strong-anecdotal	🟡 Likely (needs salary anchor)
Email / CRM / lifecycle	High	High	Fragmented	🟡 Likely (needs salary anchor)
Organic content / SEO	High (rank) / Med (GEO)	High, partly scarce	Fragmented	🟡 Mixed / moving
B2B sales-marketing	Med (relationship execution resists)	Very high	Weak	🟠 Weakly
Customer success / RevOps	Untested	Untested	Untested	❓ Next probe
Brand-building (Sharp)	Low (mechanism not compressible)	High	Holding/rising — no compression	❌ Not on curve (scope-defining)

What the matrix reveals: signal (1) is the discriminating variable. Every confirmed and likely domain shares a high-volume, structured execution layer that AI compresses. The two domains that fall off the curve fail specifically on signal (1): brand-building’s “execution” is cumulative mental-availability investment (not a high-volume task AI can speed up), and B2B sales-marketing’s core execution is relationship work (trust built over a months-long cycle with a specific human), which resists automation in a way creative or content execution does not. So the sharper version of the framework is: AI eats execution where execution is high-volume and structured; it does not eat execution that is cumulative (brand) or relational (enterprise sales). Signals (2) and (3) describe what’s left for humans; signal (1) decides whether the domain bifurcates at all.

This also flags the single most useful next data point: Customer Success / Revenue Operations is the one adjacent domain that is completely untested and sits at the intersection of all the others — its execution layer (health-score dashboards, renewal-risk flags, QBR prep, playbook drafting) looks highly compressible, while its strategic layer (which accounts to save, how to sequence an expansion) looks relationship-heavy like B2B sales. It may be the cleanest test of the high-volume-vs-relational boundary the matrix just surfaced.

What we know vs. what we’d want to know

The current state:

3 domains confirmed on the curve with hard data (paid media, influencer marketing, software)
2 candidates that look strongly on the curve but with fragmented evidence (email/CRM, analytics)
1 candidate that’s mixed/moving (organic content/SEO)
1 negative case that strengthens the framework (brand-building per Sharp)
1 candidate that’s weakly on the curve (B2B sales-marketing)

To turn the working hypotheses into firm conclusions, we’d want:

Modash-equivalent salary surveys for adjacent roles: CRM/lifecycle marketers, SEO specialists, marketing analysts, B2B SDRs/BDRs. Specifically: salary delta for strategy ownership vs execution ownership.
Job-posting linguistic analysis: trend in job titles 2022-2026 toward “Strategist” / “Architect” / “Senior” descriptors and away from “Specialist” / “Coordinator” descriptors. Indeed/LinkedIn data would surface this if dug into.
Agency-pricing benchmarks: per-asset / per-deliverable pricing trends versus strategic-outcome pricing trends, by domain.
Failure-mode documentation: domains where AI tooling shipped but the bifurcation didn’t show up. Would clarify boundary conditions.

Why this matters as a tracked question

The framework is currently anchored by three data points. Three is enough to name a pattern; it isn’t enough to call it a general principle. Each adjacent-domain confirmation pushes the framework toward “general principle” status. Each disconfirmation clarifies the scope. Either outcome is useful.

The pragmatic concern for Primores work specifically: most clients operate in CRM/email or content/SEO, not in the three already-confirmed domains. If the framework applies to those domains too, the strategic-vs-execution lens is the right way to advise them. If not, it’s a misapplication risk. Worth getting right.

Suggested next moves

If a future session wanted to advance this question, the highest-leverage actions:

Find a CRM/email salary survey analogous to Modash’s. If one exists (LinkedIn’s annual marketing salary report? Klipfolio’s annual benchmarks?), ingesting it would give the email/CRM domain a hard anchor.
Document a 2026 case study where a Primores client’s CRM/email or content/SEO operation got bifurcated through AI tooling adoption — execution-layer compressed, strategic-layer expanded. A real worked example would be more valuable than three more salary surveys.
Stress-test the framework against a 2024-2026 failure case — a marketing function where AI tooling shipped but the bifurcation didn’t materialize. Boundary-condition discovery.
Check whether the framework predicts anything in 2026 customer-success / revenue-operations roles, which sit adjacent to all of these.

glossary/automation-eats-execution — The framework this question explores
comparisons/strategy-vs-execution-ai — Full synthesis with diagnostic signals and honest limits
glossary/creative-is-new-targeting — Domain anchor 1 (paid media)
marketing/influencer-marketing-task-overload — Domain anchor 2 (influencer marketing)
glossary/vibe-coding — Domain anchor 3 (software development)
marketing/brand-vs-content-layers — The brand-building negative case
seo/agentic-search-optimization — Adjacent: SEO under agentic search
automation/ai-implementation-patterns — Adjacent: empirical anchor on what AI eats well

Key takeaways

The sharpened framework: signal (1) is the discriminating variable. AI eats execution where execution is high-volume and structured; it does NOT eat execution that is cumulative (brand-building) or relational (enterprise sales). The strategic-layer and salary signals describe what’s left for humans; the compressibility of the execution layer decides whether a domain bifurcates at all.
The framework is anchored by 3 domains; it’s natural to ask which adjacent domains are next.
Strong-but-fragmented evidence: email/CRM/lifecycle, marketing analytics.
Mixed/moving: organic content/SEO (the optimization target itself is shifting).
Weakly on curve: B2B sales-marketing (relationship execution is harder to automate).
Useful negative case: brand-building — the framework doesn’t predict the same pattern, and empirically it doesn’t show up. This is feature-not-bug for the framework’s scope.
The highest-leverage way to advance the question: a Modash-equivalent salary survey for adjacent marketing roles, OR a documented Primores-client case study showing the bifurcation in real time.

Sources

The three confirmed-domain anchors (Seufert / Modash / Karpathy + supporting Google Cloud data) are documented in glossary/automation-eats-execution and comparisons/strategy-vs-execution-ai.
Hypotheses for the candidate domains here are working extrapolations from the framework — not yet anchored in domain-specific data. Treat as questions to track, not conclusions reached.