Revision cycles are a margin problem

Most agency owners don’t need to be told that client revisions are frustrating. The bigger issue is what those revisions cost.

A campaign that should take two rounds turns into five. A client gives direction in a meeting, adds comments in a PDF, sends three Slack messages, then asks for one more version after someone senior sees the work. The designer can’t tell which note takes priority. The account manager spends an hour translating feedback. The strategist gets dragged back into a project they thought was done.

Nobody logs all of that time against the job. It disappears into “being responsive.”

For marketing and creative agencies in the $1M to $25M range, repeated revision work is often part of the $60K to $180K annual leakage band we see across delivery, account management, and reporting. The cost isn’t just design hours. It’s the account manager chasing approvals, the creative director resolving conflicting comments, and the opportunity cost of having capable people work on old jobs instead of profitable new work.

The goal isn’t to make clients feel rushed or unheard. Good clients should have a clear way to give meaningful feedback. The goal is to make feedback structured, decisions visible, and approvals final.

AI can help, but not by writing another generic email or making a prettier presentation. The useful application is an operating workflow that collects feedback in one place, identifies conflicts, controls versions, routes the work to the right person, and asks for a clear approval decision.

That is how an agency moves from five or more rounds toward two or fewer.

Why creative work gets stuck in revision loops

Revision cycles usually start going wrong before the first creative review.

The brief is incomplete. The account manager has context from calls that never makes it into the production brief. The client hasn’t identified who has final approval. The creative team has no firm acceptance criteria beyond “make it feel more premium.” Then the work goes out through email or a shared link with no guidance on how feedback should be given.

This creates four predictable problems.

First, feedback arrives in different formats. A client leaves timestamped video comments. Their colleague replies by email. An executive sends comments directly to the account manager. Someone pastes thoughts into Slack. The team now has to consolidate input before work can begin.

Second, clients comment on different versions. A client opens an old PDF from an earlier email chain and requests a change that was already made. Another person reviews the current version, but only after the designer has started work on the old comments.

Third, opinions conflict. One stakeholder asks for more copy. Another says the asset is too busy. The account manager may try to reconcile the contradiction without asking for a decision. That creates a sixth version instead of a two-minute client conversation.

Fourth, approval is vague. “Looks good from me” isn’t an approval if the decision-maker hasn’t reviewed it. Neither is silence. The job stays open, work-in-progress rises, and the team remains exposed to a late request.

The agency often treats this as a client-management problem. It is partly that. More often, it’s a workflow design problem.

A clear review system protects the client experience because clients know exactly where to comment, what decision they are making, and when their input is due.

Define what two rounds actually means

“Two rounds” sounds simple until the team tries to enforce it. Define it before you introduce any software or AI agent.

Round one is strategic and directional. The client checks whether the work meets the agreed brief, audience, message, channel, and brand requirements. The feedback should focus on decisions that affect the concept or core content.

Round two is refinement. The client checks the consolidated changes and flags factual errors, brand compliance issues, or small execution details. It is not another opportunity to change the campaign direction.

That definition needs to appear in four places:

• The proposal or scope of work
• The kickoff notes and creative brief
• The review request sent with every asset
• The approval workflow inside your project system

It should also spell out what happens when new direction arrives after round two. It may be handled under a change request, moved into the next production cycle, or approved as out-of-scope work. The exact commercial rule is your call. The key is that the team doesn’t quietly absorb it.

A two-round process doesn’t mean every project finishes in two touches. A new brand identity or major campaign may need more structured checkpoints. It means each checkpoint has a specific decision, an owner, and a boundary. That is very different from endlessly sending “v7 final final” files.

Build one source of truth for every review

Before AI enters the picture, clean up the review path.

Each client project needs one active review link or portal. That link should show the current version, the brief summary, the deadline for comments, the named approver, and the decision requested. All feedback should land there, even when it originates in email or Slack.

A practical review request might say:

Please review Version 1 against the attached brief by Thursday at 3 pm. Add comments directly to this link. We need one consolidated set of feedback from your team, approved by Jane Smith. This round is for direction, messaging, and priority changes.

That one message does more work than a long chain of polite follow-ups.

Version naming matters too. Don’t let files be called final, final-v2, or new-final-use-this-one. Use a consistent convention tied to the project and round, such as Client_Campaign_LandingPage_R1 and Client_Campaign_LandingPage_R2.

The production team should never need to guess which file is current. If a comment comes in on an older version, the system should flag it rather than allowing it to become a new task.

This is where connected tools matter. Your project platform, file storage, proofing tool, email, CRM, and communication channels need a workable handoff. We build these connected processes through Omni apps, so the workflow follows the job rather than forcing staff to rekey data across five systems.

What an AI feedback and approval workflow looks like

A useful AI workflow is not a chatbot sitting beside a messy process. It is a set of agents and rules that move review work forward without asking the account manager to become a traffic controller.

Here is the end-to-end flow.

When a project enters the review stage, the system checks that the brief includes the required fields. That includes target audience, deliverable format, campaign objective, brand references, review deadline, and final approver. If anything is missing, it prompts the account manager before work goes to the client.

The Content Production Agent takes the approved brief and prepares a first-pass draft in the required format. For some work, that might be copy, content variations, social captions, or an initial creative direction. The human team still owns judgment and craft. The difference is that they edit a strong starting point rather than starting with an empty document.

When the account manager sends the review request, the workflow creates a single review record. It captures the version number, date sent, reviewers, due date, requested decision, and link to the current asset.

As comments arrive, the feedback agent does five useful things:

1. Pulls comments from the agreed review source and any connected channels.
2. Sorts them into categories such as copy, design, legal, factual correction, brand, and strategic direction.
3. Groups duplicate comments so the team doesn’t respond to the same request three times.
4. Flags contradictions, such as one reviewer asking to shorten copy while another asks to add a new section.
5. Produces a concise feedback summary for the creative lead and account manager.

The summary should not make decisions on behalf of the client. It should make the decision points obvious. For example:

• Five comments are execution changes and can proceed.
• Two comments conflict on the headline direction.
• One new request is outside the agreed deliverable.
• No feedback has come from the named approver.
• The client has two business days left before the review deadline.

That is a real reduction in admin, but the bigger benefit is speed. The creative team receives a prioritized brief for the next version, not 27 loosely related messages.

After the team updates the asset, the workflow creates the next version, archives the prior one, and sends a focused second-round request. It can say exactly what changed and ask for one of two decisions: approve or identify any remaining factual or compliance corrections.

If a client sends broad new direction at this stage, the agent drafts a response for the account manager. It can reference the approved first-round direction, explain that the request changes scope, and offer the right next step. The account manager reviews and sends it. They don’t have to write that message from scratch while managing eight other accounts.

This is the kind of operational work covered in Omni ops. It makes the existing team more consistent without removing the human relationship from client service.

Use AI to improve client communication, not hide from it

Some owners worry that automation will make the agency feel impersonal. That only happens when AI is used to avoid an important conversation.

Use it for the repetitive coordination around the conversation.

An AI workflow can draft review reminders that reflect the project status. It can tell the client their input is needed from the final approver. It can create a short change log explaining what the team addressed in round two. It can draft a scope-change note when the request no longer matches the approved brief.

The account manager remains responsible for tone and judgment. They should review the message before it goes out, especially for a high-value account or a sensitive project. But instead of spending 30 to 50 percent of their week building reports, status updates, decks, and follow-up emails, they can spend more time helping clients make good decisions.

The Reporting Agent supports this from another angle. It pulls performance data from connected platforms, drafts the monthly report, and prepares the AM’s email summary. That means reporting doesn’t swallow the same hours needed to keep creative reviews moving. If an AM is rushing through client communication because reporting week is consuming their calendar, revision control will always suffer.

The Account Health Agent also watches for review risk. It can flag projects with overdue approvals, repeat late feedback from a stakeholder, growing out-of-scope requests, or a client who has reopened approved work several times. It then drafts the next-step message before the AM has to ask.

That gives the agency an early warning system. You don’t find out the account is unhappy only when the renewal is at risk.

For a clearer view of how those workflows fit an agency model, review the AI audit for marketing and creative agencies.

Calculate the financial case before buying tools

Don’t buy a new review platform because its demo looks polished. First, calculate what your current process costs.

Take a sample of 10 recently completed jobs. For each job, record:

• Planned creative hours
• Actual creative hours
• Number of revision rounds
• Account management time spent collecting and clarifying feedback
• Senior review time
• Unbilled out-of-scope changes
• Days between first submission and final approval

You will probably find that the visible design changes are only half the cost. A typical mid-sized client account can burn several unplanned hours per month in chasing responses, translating vague notes, and managing late changes. Across 15 to 30 active accounts, it adds up quickly.

Say your team avoids just three unplanned hours per account per month across 20 accounts. At a blended internal cost that many agencies would recognise as material, that is 720 hours a year reclaimed. Not all of that becomes direct profit. Some will be used to improve quality and client service. Still, it gives you capacity to handle more work without immediately adding another account manager or producer.

That matters because most agencies hit a scaling ceiling when each AM can only manage six to 10 accounts well. If growth requires a new hire every time the client list expands, margin stays under pressure. Better approval operations don’t replace good people. They allow good people to carry more valuable work without drowning in administration.

If you want an outside view of the number, Book a 60-min Omni Audit. We use the session to map the handoffs, identify where time and margin are leaking, and prioritise the workflows worth fixing. You get three practical outputs in 60 minutes, with no deck.

Roll this out without disrupting active clients

Start with one repeatable service line. Monthly social content, paid campaign creative, email production, or recurring design work are usually better candidates than a major rebrand.

Choose a small group of clients who have predictable stakeholders and enough job volume to produce a useful sample. Run the process for 30 days, then compare it against the prior month.

Track four measures:

• Average revision rounds per deliverable
• Time from submission to approval
• Unplanned hours per job
• Percentage of reviews completed by the named approver

Don’t try to automate every exception on day one. First, establish the one review link, one current version, one feedback deadline, and one clear approval state. Then add AI-based feedback summaries, reminders, conflict detection, and scope-change drafts.

Your team needs a short operating rule too. If feedback is outside the review system, the AM’s job is to put it into the system, not forward it as another unstructured message. If someone other than the named approver comments, their input is welcome, but the approver still needs to consolidate the final decision.

Clients usually adapt quickly when the process is explained as a way to protect turnaround time and quality.

Your internal team may need more coaching. Designers and strategists often carry work around in their heads because they don’t trust the process to capture the nuance. Show them that the feedback summary links back to the underlying comments, and make it easy to correct categorisation. AI should reduce handoffs, not create a black box.

You can find more practical operating ideas in our agency operations guides and the broader Enterprise DNA insights library.

The better outcome is fewer decisions, made earlier

Reducing revision cycles isn’t about telling clients no more often. It is about helping them make the right decisions at the right moment.

The best agencies set direction before production begins. They make it easy for clients to give feedback. They expose conflicting views quickly. They preserve a clean version history. They know when a request is a refinement and when it is new work.

AI makes this easier because it can handle the tedious coordination that usually falls to account managers and producers. It can gather the evidence, summarise the decisions, flag risks, and prepare the next communication. Your people can focus on the conversation that needs judgment.

If your agency is losing margin to five, six, or seven rounds of revisions, it is worth mapping the workflow in detail. The leak is rarely one dramatic failure. It is dozens of small handoffs repeated across every account.

See Omni for marketing and creative agencies to understand where these agents can fit in your delivery model. When you’re ready to identify the first workflow to fix, Book a 60-min Omni Audit.