There is no honest guarantee about jobs

A confident declaration that AI will never replace marketers would be impossible to defend. Tasks can disappear, teams can shrink and employers can choose cost reduction over better work. The headline is an argument about the substance of marketing, not a forecast that protects every role.

OpenSkool's position is that a marketer whose value is defined only by producing assets or operating an interface is vulnerable when those activities become easier to automate. The stronger professional contribution is to understand which problem deserves attention, choose a credible response and establish whether it helped the business and its customers.

Execution has always contained judgment

It would be a mistake to divide the profession into strategic thinkers and replaceable button-pushers. A skilled practitioner notices that a campaign is attracting the wrong people, that a creative claim is misleading or that a tracking change makes a result incomparable. Those judgments often emerge through doing the work.

The distinction we need is between following a procedure without understanding its purpose and being able to explain, challenge and improve that purpose. A junior marketer asking why a conversion matters can contribute more strategic value than a senior presentation full of untested assumptions.

Automation changes where that understanding must be applied. Google Smart Bidding already automates auction-level optimisation towards specified conversion goals. The marketer's question is whether those goals represent useful business outcomes and whether the evidence justifies the next investment.

Evidence & context: Google Ads Help

Customer understanding cannot be replaced by a plausible customer

A synthetic persona can speak fluently about a problem the business has never investigated. A dashboard can display a segment without explaining why people buy, hesitate or leave. Marketers need contact with evidence that can contradict their preferred story.

Imagine an agency asked to generate more enquiries for a service with a fulfilment problem. Faster ads and better targeting could worsen the customer experience. The valuable intervention may be to narrow the promise, repair the service or delay expansion. That decision requires enough business understanding to question the brief.

The practical discipline begins with audience research that keeps observation separate from invention, then a strategy that makes real choices.

Creativity needs a point of view about the problem

Producing alternatives is useful, but choosing among them requires a reason. A distinctive campaign might emerge from recognising an overlooked customer tension, reframing a category assumption or explaining an unglamorous truth that competitors avoid. The asset is the expression of that insight, not its substitute.

AI can participate in exploration. It can suggest objections, help articulate competing ideas and expose inconsistencies in a brief. The team still has to judge which direction fits the customer, the brand's actual capability and the circumstances. A large choice set is not the same as a strong choice.

What if the organisation does not reward better judgment?

This is the serious counterargument. An employer may reward content volume and lower costs even when customer understanding deteriorates. Knowing more will not automatically protect a person's job. Nor should workers alone bear responsibility for redesigning an organisation that measures the wrong things.

Leaders have to make room for investigation, experimentation and disagreement. They also need a new apprenticeship: if routine execution disappears, how will less experienced marketers encounter real errors and learn to diagnose them? Give them supervised decisions, access to customer evidence and responsibility for explaining results, rather than only responsibility for producing more drafts.

Risk guidance such as NIST's reminds organisations to assign accountability. A nominal human in the loop is not a substitute for someone with context and the authority to challenge an action.

Evidence & context: NIST

Make the reasoning part of the work

  • Before a campaign: state the customer problem, the commercial objective and the assumption being tested.
  • During execution: record what changed and why, including rejected alternatives.
  • Afterwards: distinguish observed results from causal claims and identify what remains unknown.
  • For the next decision: explain what evidence would make you stop or change direction.

The profession should not defend repetitive work simply because it once required skill. It should defend the opportunity to develop and exercise judgment. If AI creates capacity, the worthwhile question is whether that capacity goes into understanding customers and improving decisions—or merely into producing more activity.

That is the connection to critical thinking: an answer, an asset or a recommendation becomes useful only when someone can justify relying on it.

Sources & further reading

  1. Smart Bidding: definition

    Google Ads Help. Official description of auction-time optimisation. Checked 11 September 2026; not evidence of guaranteed business results.

  2. Generative Artificial Intelligence Profile (NIST AI 600-1)

    NIST. Risk-management guidance, including confabulation. It does not establish a universal error rate.

Examples and exercises are illustrative unless attributed to a source. No independent expert review is claimed.

A correction, a counterexample or an experience worth sharing?

Join the conversation ↗