A single marketing email can involve four people, five software tools and three rounds of revisions. AI can write the copy in seconds, but the production process is proving much harder to speed up.
Artificial intelligence has made writing marketing copy considerably faster. Getting that copy approved and in front of customers is another matter.
Research from marketing production company Knak found that 85% of enterprise marketing teams missed at least one planned campaign launch date in the previous year. One in ten missed more than five.
The study, involving 333 marketing decision-makers in the UK, US and Canada, found that approvals, creative production and coordination between teams were the main reasons for delays.
That is a different problem from writing copy, and one that generative AI has yet to resolve.
Nick Millward, founder and CEO of customer engagement platform Lyberate, argues that the marketing technology industry has spent too much time promoting AI’s creative capabilities while paying less attention to the work required to launch campaigns.
His comments follow a series of 2026 studies that put figures on the time marketers lose to revisions, administration and disconnected software.
Four People, Five Tools and Three Rounds of Changes for One Email
The Knak research offers a revealing picture of what goes into producing a marketing email.
At the companies surveyed, 60% involve at least four people in the process. More than half, 54%, use between three and five separate software tools, while 69% need two or three rounds of revisions before an email is approved.
For 51% of teams, approvals are still managed primarily through email, Slack or Microsoft Teams conversations.
Knak estimates that the internal labour involved in producing a single email exceeds $300, based on US wage data. That is a modelled cost, not an independently measured average.
A short promotional email might take a copywriter 20 minutes to prepare. Yet it can still require a designer, campaign manager and brand reviewer before anyone presses send.
Adding AI to the copywriting stage will not necessarily shorten the remaining work.
Knak identified approvals and sign-off as the most common cause of missed campaign dates, cited by 47% of respondents. Design and creative production followed at 38%, with coordination between teams at 36%.
“The AI conversation in marketing has been dominated by content generation,” said Millward.
“The real drag is what happens after the idea exists.”
His argument is that businesses should be examining the steps between an approved idea and a live campaign, rather than concentrating on producing more drafts.
UK Marketers Are Spending Ten Hours a Week on Admin
A separate UK study from Optimizely, published in August 2026, found that 45% of marketers spend at least a quarter of their working week on low-value administrative tasks.
For a full-time employee working 40 hours, that represents at least ten hours each week.
Another 14% reported spending more than half their working time on administration.
The survey of 100 UK marketers identified slow approvals as the most common source of unnecessary work, cited by 38%. Poor or incomplete data followed at 34%.
A third also complained about unnecessary meetings, managing too many platforms and work related to justifying budgets.
The consequences are apparent in campaign delivery. Optimizely found that 60% of respondents often launched campaigns without enough time or data to optimise them properly.
The research does not establish how many hours AI could recover. It does, however, identify where marketers are spending time that could otherwise go towards campaign planning, testing and performance analysis.
AI Is Giving Marketers Another Editing Job
Automated content creation brings its own workload.
Knak found that 88% of enterprise marketing teams consider AI-generated material to require moderate or substantial editing before publication.
Optimizely’s June 2026 international study reached a similar finding.
Among 2,003 B2B marketing leaders surveyed across seven markets, 76% reported spending at least three hours a week editing, fact-checking or correcting AI output.
Almost half, 48%, said checking inaccurate or invented information created additional work. Another 40% identified transferring information between disconnected platforms as a drain on time.
Only 4% reported that AI saved them time at every stage of their work.
More concerning was the effect on quality. A quarter of respondents admitted frequently or always publishing AI-generated content they knew was not fully consistent with their brand when under deadline pressure.
That creates a problem for companies using content output as a measure of productivity.
A marketer who once prepared four versions of an email can now generate twenty. But those extra versions still need checking for factual accuracy, branding, links, product details and legal claims.
The software has accelerated drafting without necessarily reducing the overall workload.
For agencies and in-house marketing departments, the cost is not limited to the person generating the content. Editors, designers and compliance teams can inherit the additional work.
Where Campaign Production Gets Stuck
The research identifies several recurring problems.
| Production problem | Reported finding | Source |
|---|---|---|
| Missed campaign launches | 85% missed at least one date in 12 months | Knak, 2026 |
| Approval delays | 47% cited approvals as a cause of missed launches | Knak, 2026 |
| Email revisions | 69% require two or three rounds | Knak, 2026 |
| Email production tools | 54% use three to five tools | Knak, 2026 |
| UK administrative workload | 45% spend at least a quarter of their week on admin | Optimizely, August 2026 |
| Correcting AI content | 76% spend at least three hours weekly | Optimizely, June 2026 |
These are findings from separate surveys with different samples. They should not be added together or treated as a single measure of industry productivity.
Slow Campaign Execution Is Costing Brands Opportunities
The problem extends beyond wasted employee time.
Adobe’s 2026 research into marketing workflows found that 84% of 150 surveyed marketing professionals reported missing at least one marketing opportunity during the previous quarter because their organisation could not respond quickly enough.
Some 29% reported missing six or more.
Adobe did not establish the financial value of those missed opportunities. Nevertheless, the findings point to a commercial consequence of slow campaign production.
A retailer responding to an unexpected rise in demand might have its creative material ready within hours. Approval delays, customer data checks and scheduling could still prevent the campaign from going live while that demand is strongest.
For companies managing time-sensitive promotions, a missed launch is not simply an internal inconvenience.
It can mean losing the chance to reach customers when interest is highest.
Why More Marketing Channels Mean More Production Work
The problem becomes harder when a campaign must run across several platforms.
A retail promotion may involve email, paid social, website content, WhatsApp and RCS messaging.
Each channel requires different formats and often different approval processes. Customer permissions must also be checked.
A message prepared for email may need rewriting for WhatsApp. Creative assets must meet the specifications of each platform, while links, offers and product details must remain accurate.
Millward believes these separate processes contribute to unnecessary duplication.
“Customers do not experience a brand in neat departmental silos,” he said.
His company, Lyberate, is developing its customer engagement business around automating the work involved in preparing and delivering communications across email, social media, WhatsApp and RCS.
The approach reflects a wider interest in connecting campaign production rather than treating each channel as an independent project.
However, Lyberate’s October announcement did not provide independently verified customer results showing how much time its platform saves or how many campaign delays it eliminates.
That evidence would be needed to establish whether its approach delivers measurable improvements.
There are also limits to what should be automated. Customer consent, sensitive claims and the final decision to launch a campaign still require appropriate controls.
What Would Actually Make Campaign Production Faster?
The Knak research offers one useful comparison.
It found that organisations with more advanced AI use were 85% more likely to produce a single email in under an hour than teams at earlier stages of adoption.
The report defines advanced users as those applying AI across production workflows rather than primarily using it for drafting.
This is an association, not proof that AI caused the improvement. Better-organised teams may also be more likely to adopt advanced automation.
Even so, it points towards a different way of evaluating marketing technology.
Rather than counting how many pieces of content employees generate, marketing departments could measure:
- The number of working hours between campaign brief and launch.
- The number of people involved in producing each campaign.
- The number of approval rounds and amendments.
- The staff hours required to produce an email or landing page.
- The proportion of campaigns launched on their planned dates.
- The impact of faster delivery on conversions, revenue and customer engagement.
A business that cuts email production from eight hours to four has achieved a measurable improvement, provided quality and performance do not suffer.
The same business has achieved considerably less if copywriting takes five minutes instead of half an hour but the campaign still spends two days waiting for approval.
Some of these delays could be addressed through better software integration. Others may simply require clearer responsibilities, fewer unnecessary reviews or changes to internal processes.
Not every production problem needs another AI subscription.
Marketing’s AI Productivity Claims Need a Different Test
The evidence from Knak, Optimizely and Adobe points to a problem that has attracted relatively little attention during the generative AI boom.
Campaign production involves considerably more than creating content. It includes checking, editing, building, approving, scheduling and coordinating work across teams.
In many organisations, these activities still depend on manual processes.
Millward believes the greatest productivity gains will come from removing that work.
“The best use of AI will be the least visible to the customer,” he said.
The claim is commercially convenient for a company selling customer engagement technology, but the underlying argument has merit.
AI suppliers have become very good at demonstrating how quickly their products generate content. Marketing buyers should ask equally detailed questions about what happens afterwards.
How many people are needed to launch a campaign? How much time disappears into revisions? How often does a promotion miss its planned date? And can the software show a reduction in those figures?
Until businesses start measuring those outcomes, faster AI-generated content will remain a poor substitute for faster marketing.
Sources
- Knak — Marketing Production in the Age of AI, July 2026
- Knak — Detailed Campaign Production Survey Findings, July 2026
- Optimizely — UK Marketing Administration Research, August 2026
- Optimizely — Global AI Productivity Research, June 2026
- Adobe — Accelerating Trend-to-Campaign Execution, September 2026
- Lyberate — Nick Millward’s comments from the company’s statement dated 28 October 2026.
The research cited comes from vendor-sponsored surveys. Reported delays and productivity problems reflect respondents’ answers, while financial estimates and comparisons between teams should be interpreted within each study’s methodology.