A content approval process is the agreed route a piece of content takes from first draft to final sign-off. It sets who reviews it, who approves it, in what order and by when. AI has made the first draft cheap, so for many teams the approval queue is now the constraint. Optimizely's UK research found slow approval processes are the biggest source of unnecessary work, cited by 38% of marketers.
Running approvals is one step of marketing planning for 2027, and plans rarely spell it out. This guide shows how to build a content approval process that keeps pace with AI drafting. It covers the steps, the routes for each type of content, turnaround times and the checks software should run first. It then covers approving AI-generated content, UK rules and how to measure where work waits.
What is a content approval process?
A content approval process is the set of review stages, people and deadlines that every piece of content passes through before it goes live. It names who checks the work, who gives final approval and what happens when someone misses a deadline.
It is narrower than a content workflow, which covers the whole journey from idea to publication and reporting. A content approval process is only the part that decides whether the work is ready, and that is where most approval delays build up.
Two roles in the review process are easy to confuse. A reviewer advises, checking facts, tone or design and suggesting changes. An approver decides: they give the final sign-off and own the result. Good review and approval design keeps the two apart, so many people can comment but only one person can say yes.
A working content approval process covers four things for every piece of content:
- Stages: the order of checks, from first draft to final approval.
- People involved: the content creators, the reviewers and the final approver, each by name.
- Deadlines: how long each stage may take before the work escalates.
- Records: what was approved, by whom and when, kept in one place.
Types of content need different routes, because a social post and a pricing page carry different risks and should not queue behind the same approvers. We cover routing by content type later in this guide.
Why do marketing approvals stall?
Marketing approvals stall because nobody owns the queue, too many people are involved and feedback arrives in rounds rather than at once. Reviewers judge against personal taste rather than an agreed brief, and each round adds days of back-and-forth through email and chat.
The number of rounds is growing, according to a 2025 study by BetterBriefs and Flood + Partners, produced with the IPA and the WFA. It surveyed 1,034 agency, marketer and in-house respondents in 54 countries. It found creative work typically takes five rounds to reach a signed-off idea, up from three in UK research in 2007. Only 23% of agencies and 56% of marketers agreed the current approval process works well.
The same study suggests why: only 43% of agencies and 62% of marketers agreed that the right people sign off on creative ideas. Just 10% of all respondents said ideas are always evaluated against clearly defined criteria. Without criteria, every reviewer brings their own, and each new opinion means another round.
Respondents described the approval process as "inconsistent, slow, subjective and painful". That is the coordination tax at its most visible: time spent waiting, chasing and re-explaining rather than improving the work.
Larger organisations feel it too. The Content Marketing Institute's enterprise research was published in January 2025. It found 47% of enterprise marketers named managing workflow issues or the content approvals process as a challenge. And 61% named communicating across organisational silos, where every silo adds another set of reviewers.
The admin around approvals adds to the cost. Optimizely's UK research found a third of marketers point to meetings for the sake of meetings, managing too many tools and having to justify budgets. Each of those can slow down sign-off even when the work itself is ready.
Why has AI made approvals the bottleneck?
AI has made drafting fast, but approval still depends on a few people with fixed hours. When a team produces more drafts without adding review time, the extra work does not reach customers sooner. It waits in the queue, and launch dates start to slip.
Launch dates are already slipping, according to Knak's 2026 research with 333 enterprise marketing decision-makers in the US, the UK and Canada. It found 85% of enterprise marketing teams surveyed had missed at least one campaign launch date in the past 12 months because of workflow constraints. Getting approvals and sign-off was the single biggest reason.
The arithmetic is simple: say your approvers can give content 10 hours a month, and each asset takes 30 minutes to review. At that rate they can approve 20 assets a month, so if AI doubles your drafts to 40, the other 20 do not ship. They wait, and the wait grows every month.
Most plans count drafting capacity and forget review capacity, yet drafting now scales with tools while review still scales with people's calendars. A content approval process built for 20 assets a month will fail at 40, however good the drafts are.
There are only three ways out:
- Add review time: book approvers' hours into the plan, as you would book an agency.
- Cut review load: let software clear routine issues before a person looks, and route low-risk work past senior approvers.
- Cut rework: give writers and AI the context they need, so fewer drafts come back.
Most teams need all three, and the rest of this guide shows how to design them into your content approval process.
How do you build a content approval workflow step by step?
Build a content approval workflow by setting the route when the work is planned, not when the draft arrives. Name the owner, reviewers and final approver, then run machine checks, one review round, specialist checks when triggered and a single final approval.
- Set the route in the plan: when content enters the calendar, record its owner, reviewers, final approver and the date each must act by.
- Start from an agreed brief: reviewers judge the draft against the brief and your brand guidelines, not personal taste.
- Run machine checks first: software checks spelling, links, banned words and house style before anyone reads the first draft.
- Hold one review round: reviewers comment on the latest version at the same time, and the owner settles conflicting comments.
- Add specialists on a trigger: legal, product or a subject expert reviews only when the content contains a new claim, figure, customer name or comparison.
- Give final approval to one person: the named approver signs off, and the approved version is locked.
- Record and go live: keep an audit trail of what was approved, by whom and when, then publish.
Step 1 is the easiest to skip, because it is tempting to plan publish dates and leave approvals to chance. Put the approvers' names and dates into the plan, and their time becomes a commitment rather than a favour.
Step 4 does the most to cut rounds in a content approval process. When team members review one after another, each sees a different version and the comments multiply. When they review together and centralise feedback in a single thread, the owner can settle conflicts once.
Use automated reminders for every stage with a deadline, so nobody has to chase. Chasing is work about work, and it is the first thing a good workflow removes.
Keep the number of review stages as low as the risk allows, because every extra stage adds a queue and every queue adds days.
Should every type of content follow the same approval route?
A good content approval process matches the route to the risk. Routine social media posts need a quick check and one approver, while claims, comparisons and regulated content need a specialist. One route for everything means low-risk work queues behind high-risk work.
| Content type |
Risk |
Route |
Suggested window |
| Social media posts and routine updates |
Low |
Machine checks, then the owner approves |
One working day |
| Blog posts and guides |
Medium |
Machine checks, one editor, then the owner |
Two working days |
| Design assets and campaign creative |
Medium |
Creative team review, then the campaign owner |
Two working days |
| Case studies and customer quotes |
Medium |
As blogs, plus the customer's written approval |
Five working days |
| Product claims, comparisons and pricing |
High |
Machine checks, a subject expert, then the final approver |
Three working days |
| Regulated or financial content |
High |
Legal or compliance signs off every time |
Agreed with legal |
The windows are a starting point, not a standard, so set yours from the wait times you measure and shorten them as the checks improve.
Write the triggers down too: a blog post that quotes a new figure or makes a comparison takes the high-risk route, whatever its format. That rule lets content creators route their own work without asking.
The point is to protect senior attention: when every piece of content goes to the Head of Marketing, the queue forms at the top. Our guide to marketing team structure sets out who should own each decision. The route then sets how work moves between those owners, and how fast.
A creative team can run its own route for design assets, as long as the final approver and the window are agreed. What matters is that each route is written down before the work starts, not negotiated asset by asset.
How fast should content approvals be?
Set a review window for each stage of your content approval process, short enough to keep the plan on time. Most routine content can be reviewed within one or two working days. When a reviewer misses the window, the work goes to a named backup rather than waiting.
A window only works if the time is real, because an approver with no hours booked for review cannot meet it. So book review time into calendars when you set the plan, in the same way you book production time.
Three rules keep windows honest:
- A backup for every approver: holidays and busy weeks should not stop the queue.
- Escalation, not a pile: when an approver's window passes, the work goes to the backup or up a level.
- A recorded yes for every approval: no content goes live on silence, so each sign-off is logged with a name and a date.
Plan approvals backwards from launch dates. If a campaign goes live on 1 March and its route takes eight working days, the final draft must be ready by mid-February. Put those dates in the plan, so a late draft shows up as a risk before it becomes a missed launch.
The project management tool matters less than the rules. Any board that shows each asset's stage, owner and deadline can run this. What matters is that everyone can see where work is waiting and why.
Speed also depends on what reaches the approver, since someone who receives a clean draft against a clear brief can decide in minutes. A reviewer who receives a rough draft with no brief will send it back, and the window starts again.
Which checks should run before a person reviews?
Let software run every check that has a right answer before a person reads the draft. Spelling, house style, banned words, broken links and missing disclaimers can all be checked automatically. People then spend their review time on judgement, not typos.
A useful split is between checks with a right answer and checks that need judgement:
| Check |
Who runs it |
| Spelling, grammar and house style |
Software |
| Banned words and required disclaimers |
Software |
| Links, sources and image rights |
Software, then a person spot-checks |
| Figures matched against your approved proof |
Software flags, a person confirms |
| Brand voice against your brand guidelines |
Software scores, a person decides |
| Positioning, claims and risk |
A person |
The more routine issues software clears, the fewer rounds the work needs. A reviewer who no longer marks typos can read for meaning, which is the part only they can do.
Machine checks in a content approval process need written rules to check against: your style guide, banned words, approved claims and proof. If those rules live in people's heads, no tool can apply them. Our AI compliance framework sets out the approval gates and audit trail that sit around those checks.
We designed the Agentic Marketing Platform® (AMP) around that split. Agents draft from a Shared Marketing Brain that holds your approved positioning, audience and proof. Brand QA runs on every output before it reaches a reviewer. The final approval stays with a named person on your team, because accountability for what you publish should not move to software.
That is what human in the loop should mean for approvals: machines clear the routine, and people make the call.
How should AI-generated content be approved?
Approve AI-generated content through the same route as any other content, but check it harder. Confirm every figure, quote and source against the original, and name the person accountable for the published version. The approver signs off the content, not the tool that drafted it.
The risk is real. In IAB research with Aymara, based on a July 2024 survey of 125 US advertising executives, 70% reported at least one AI incident. The consequences included 40% having to pause or pull ads, and over a third dealing with brand damage or PR issues.
The report also found most teams rely on human review and brand integrity checklists, which it calls "important but basic steps". A checklist catches what a reviewer is looking for, but AI errors can be ones nobody thought to check. Think of an invented statistic, or a source that does not say what the draft claims.
Add three checks to the route for AI-generated content:
- Source check: open every cited source and confirm it says what the draft says.
- Claim check: compare every figure and claim with your approved proof, and remove anything you cannot back.
- Context check: confirm the draft fits the audience, the buyer's stage and the decisions already made.
Flag AI-assisted drafts in the workflow, so reviewers know where to look harder. The flag is not a mark of lower quality; it tells the reviewer which checks matter most.
AI does not change who is accountable: the named approver owns what goes live, whoever or whatever wrote the first draft. That is the line to hold in your content approval process as agents take on more of the drafting.
What do UK rules mean for B2B content approvals?
UK advertising rules apply to B2B marketing, not just consumer ads. The CAP Code covers marketing emails and the claims on your own website. So your content approval process needs a step that confirms you can prove every objective claim before it goes live.
The CAP Code counts anyone likely to see a marketing communication as a consumer, including people who see it at work. The ASA's guidance on its remit confirms the Code covers emails, direct mail and claims on a marketer's own website.
The rule that matters most for approvals is rule 3.7. Before a marketing communication goes out, "marketers must hold documentary evidence" for claims people are likely to read as objective. So the evidence for each claim belongs in the approval record, not in someone's inbox.
The regulator is using automation too. The ASA's annual report for 2025, published in April 2026, says its Active Ad Monitoring system processed 60 million ads. Through its work, 22,383 ads were amended or removed.
Two more checks belong in a UK route:
- Financial promotions: in financial services, the FCA lets authorised firms known as section 21 approvers approve financial promotions for businesses that are not authorised. If you market a financial product, put compliance sign-off on the route before anything goes live.
- Customer names and quotes: get the customer's written approval before you publish a case study, logo or quote, and keep it with the approved version.
None of this needs a separate process: add a claims and evidence check to the high-risk route, and store the evidence with the approval.
How do you measure a marketing approval workflow?
Measure how long work waits at each stage of your content approval process, not just how long it takes overall. Track wait time and work time separately, by stage and by approver. Then count rounds and on-time launches, and review the numbers with the plan each quarter.
Five measures give a clear view of your marketing approval workflow:
- Wait time by stage: the hours from when a stage could start to when someone starts it.
- Work time by stage: the hours spent reviewing or revising.
- Rounds per asset: how many times work goes back before final approval.
- First-pass approval rate: the share of work approved in its first review round.
- On-time go-live rate: the share of content that goes live on the planned date.
Wait time is the measure to start with, because overall cycle time hides where the delay sits. A two-week approval could be one hour of review and thirteen days of waiting. When waiting is far larger than working, faster reviewers will not help. You need more review capacity, a shorter route or fewer people in it.
Rounds tell you about the brief: if most work needs three rounds, reviewers are probably judging against different criteria. Fix the brief and the criteria before you add reviewers.
You do not need new software to start. The one-week audit in our guide to workslop already marks where work waits in your content approval process. Extend it to record the wait at every stage and for every approver. To see what the queue costs, our guide to AI marketing ROI shows how to measure the whole cycle from brief to approval.
Review the measures every quarter, when you review the plan. Move slow decisions down a level, cut stages that never change the work and adjust the windows to match.
What to do before you add more AI drafting
AI has made drafting fast, so the content approval process now sets the pace. Before you add more AI drafting, put approvers, review windows and backups into the plan. Let software clear the routine checks, route work by risk and measure where it waits. That is how marketing planning for 2027 keeps the coordination tax out of the approval queue. Then book a demo to see brand QA run before your team reviews.