Marketers were promised that artificial intelligence would hand back hours for strategy. New data suggests it has mostly handed back faster first drafts — and then dumped them into the same clogged approval pipeline.
A report from marketing production platform Knak, Marketing Production in the Age of AI, found that 85% of marketing teams surveyed missed at least one campaign launch date in the past year because of workflow constraints. And 82% still spend at least half their time on production rather than planning or strategy. The analysis was highlighted in a MarTech column arguing that AI has accelerated the start of the process without touching everything that follows.
Where AI is actually being used
Adoption is not the issue. According to the study:
- 64% use AI to generate first drafts of email or landing page copy
- 56% use it to generate or edit images
- 56% use it to analyse performance and suggest optimisations
- 48% use it to produce subject line variations
- 70% have deployed AI somewhere in marketing production — yet 88% say the output still needs moderate or substantial human editing
In other words, the blank page has been solved. The campaign has not.
The bottleneck was never the copy
The Knak data points to where time genuinely leaks. The biggest reported delay is securing approvals and sign-offs, cited by 47% of respondents, followed by design and creative production (38%) and cross-team coordination (36%). Sixty per cent of teams need at least four people to ship a single email, and 69% go through two or three rounds of revisions before approval.
Those are process problems — unclear ownership, too many handoffs, approval-heavy governance. Adding a generative tool at the front of that chain simply gets you to the traffic jam sooner. The hour saved on a draft rarely converts into strategic thinking; it gets absorbed by another revision round, another stakeholder review, or a fresh request from someone who now knows you can turn things around faster.
More options, more decisions
There is a second, subtler cost. When producing alternatives was expensive, teams were selective — two subject lines, one creative route. Now they can generate ten subject lines, five opening paragraphs and multiple visual treatments in minutes. Every extra option, though, is another decision someone has to make.
The column frames this as choice overload and distinction bias: the more variants you place side by side, the more you hunt for trivial differences. Instead of asking whether the campaign is strategically sound, teams end up debating whether version seven reads marginally warmer than version four. AI cuts the cost of creating options; it does not cut the cognitive cost of choosing between them.
What separates the fast teams
One finding is worth pinning above your desk. Teams that can complete an email in roughly four hours tend to involve fewer people and use AI deliberately rather than experimentally. The advantage, in other words, comes from where AI is applied, not whether it is available.
The logic is blunt: AI-generated subject lines will not fix a six-person approval chain. Faster copy will not fix a model where every campaign is built from scratch. And no tool will prevent conflicting stakeholder feedback when brand guidance is vague.
A practical audit before your next tool purchase
Before buying more software, map how a single campaign actually travels from idea to launch. The report’s implied checklist is a good starting point: Who is involved? Who must approve? Which decisions reliably cause delays? Where is work duplicated? How many tools, documents, messages and meetings does one email require? What could be standardised through templates, modular design systems and clearer approval rules?
Crucially — does anyone measure the production process itself? Most teams track opens, clicks, conversions and unsubscribes with precision, but almost nobody tracks the effort spent getting the campaign out of the door. Without that visibility, it is easy to celebrate minutes saved on a draft while ignoring hours lost downstream.
As Knak Chief Marketing Officer Jennifer Delevante notes in the study, marketers were told AI would return time for strategy, and the data suggests that has not happened yet. The technology has not failed. The implementation has — because a fragmented, approval-heavy workflow was never going to be rescued by a faster first draft.
Source: MarTech
