Why marketing AI needs to reduce administrative work

Marketing’s AI push is colliding with an older productivity problem

AI spending is rising while administration still consumes a large part of marketers’ working week. Optimizely research covering 100 UK marketers and 1,000 UK shoppers found that 45% of marketers spend at least a quarter of their week on low-value administrative tasks.

That changes the question for marketing leaders evaluating AI. They need to measure whether AI removes work from existing workflows and gives teams more time for planning, analysis and creative decisions.

Administrative friction is reaching the campaign itself

Optimizely found that one in seven marketers said low-value administration takes more than half of their working week.

Optimizely also found that six in 10 marketers said they often launch campaigns without enough time or data to optimise them.

The shopper findings point to a separate problem. Among 1,000 UK shoppers surveyed by Optimizely, 59% said the quality of marketing content has declined in recent years. These findings do not establish causation. But they give executives two issues to examine together: marketers report insufficient time or data for optimisation, while consumers report weaker content.

Tara Corey, Senior Vice President of Marketing at Optimizely, argues that when administration takes a quarter of the week, marketers have less time to consider what customers want and produce work they value. Optimizely sells marketing technology, giving it a commercial interest in arguments that support investment in more effective marketing systems.

AI adoption needs a productivity test

Marketing teams are already investing in AI. Optimizely reported the following investment and training responses:

For executives, acquisition and productivity are separate measures. An assistant may speed up content generation while a campaign still passes through the same approvals, data collection and platform handoffs. Leaders need to measure the whole process rather than infer productivity from the speed of one task.

The constraints sit across workflow, data and technology

Optimizely found that respondents cited slow approvals and poor data as causes of unnecessary work. Slow approvals were cited by 38%, while 34% pointed to poor data.

Each constraint requires a different management response. An organisational approval requirement may require a process change. A data problem may require changes to collection, access or quality. Adding an application may also increase the number of systems employees must handle.

Corey argues that standalone AI layered over disconnected systems and slow processes can become another thing for marketers to manage. This is Optimizely’s interpretation of research that supports its commercial interests. Leaders can turn the vendor claim into a measurable question: does an AI investment remove existing handling and waiting from a campaign process?

That test matters when teams expand their technology stack to address productivity pressure. Executives can compare the workflow before and after deployment. These are proposed management tests rather than findings from the Optimizely research.

Measure AI by the work it removes from the workflow

Executives can test AI investments against the workflow they already operate. For each investment, they should ask:

• Which administrative step will it eliminate or shorten?
• Which existing systems and data will it connect to?
• Will it reduce handoffs or approval delays?
• Will it create another interface that employees must manage?

Then measure the targeted step before and after deployment. If the investment is intended to shorten approvals, measure approval time. If it is intended to remove administrative handling, measure the time employees spend on that work. That gives executives a defined operational result against which to judge the investment.

Key executive takeaways

• Administrative friction is affecting campaign execution: 45% of UK marketers spend at least a quarter of their week on low-value administration, while six in 10 often lack enough time or data to optimise campaigns. Leaders should identify where administrative work is reducing capacity for higher-value marketing.
• AI adoption needs a productivity test: Investment in AI assistants and training is rising, but faster individual tasks do not necessarily mean more productive workflows. Measure whether AI reduces total handling, waiting and administrative time across the campaign process.
• Workflow, data and technology need different fixes: Marketers cite slow approvals and poor data as sources of unnecessary work. Leaders should address each constraint directly rather than assume another application will solve process or data problems.
• Measure AI by the work it removes: Define the operational result expected from each AI investment, such as shorter approvals or less administrative handling, and measure it before and after deployment. This provides a concrete basis for judging productivity gains.