Now That You Use AI, Do You Still Need a Task Management Tool? What Changed in Creative Client Work

AI has made the work itself faster. And yet, it does not really feel like projects finish any sooner.

This article looks at what changed, and what did not, once a creative team started using AI. The short answer first: you still need a task management tool. What changes is what you record in it. The focus moves from a list of work items to a record of decisions and agreements.

Note: this article is written by the developer of GeckoFlag.

What AI sped up: the “making”

Using generative AI at work is no longer unusual. In Japan, the Ministry of Internal Affairs and Communications’ 2026 White Paper on Information and Communications (in Japanese, published July 2026) reports that, in its FY2025 survey, 86.4% of Japanese companies were using generative AI in at least one area of their business. The year before, the figure was 55.2%.

Note: this figure excludes companies that answered “don’t know” about their AI policy.

In creative work, too, plenty of tasks have become faster with AI.

• Rough wireframes: describe the purpose of a page and what should go on it, and layout ideas appear right away.
• Copy options: you can produce taglines or button labels ten or twenty at a time.
• Image variations: you can make several versions of a hero image with a different color or mood.
• Meeting summaries: you can pull the decisions and the action items out of a long transcript.

Work that used to take half a day now takes shape in minutes. That is a big change.

How long the work takes is now hard to predict

At the same time, a new problem has appeared: it is harder to tell how long a piece of work will take.

Sometimes the first result is usable. Other times, no matter how often you regenerate, nothing feels right. You change the wording of your instructions, add reference material, and try again. Before you know it, two hours have gone by. That is not unusual.

Until now, teams estimated “this will take three hours” and tracked progress by the hours spent. Once AI is involved, those estimates stop being reliable. Estimating and tracking progress in units of working time is gradually becoming a poor fit.

What AI did not speed up: the “deciding”

When AI gives you more options, you also have more choices to make. Picking one tagline out of twenty takes longer than picking one out of three.

And the following things do not get any shorter with AI:

• Agreeing on a direction inside the team: the director and the designer talk it through and decide which option to go with.
• Getting the client to review: you show them the options and wait for a reply. Sometimes the client has to circulate them internally first.
• Deciding whether to accept a change request: you decide whether a change the client asked for falls within what was agreed at the start.

All of these are things people think through and decide. AI can gather the material for a decision, but people make the decision.

When a project stalls, look at the reasons one by one and you will often find that it is waiting for someone’s decision rather than behind on work. Is it waiting for someone inside the team, or for the client? Because AI has made the work faster, this waiting now stands out more than before.

Tracking work items in detail matters less

Until now, task management mostly meant breaking a job into work items and giving each one a time estimate, an owner and a deadline. Something like “Wireframes, 4 hours, Alex, by Wednesday.”

When AI makes the work shorter and harder to predict, that level of detail means less. A task estimated at four hours might be done in twenty minutes, or it might take half a day of regenerating.

What you want to record instead looks more like this:

• What needs to be decided
• Who decides it
• By when it needs to be decided
• What was decided

Side by side, the old and the new look like this:

This does not mean you stop recording work. It still helps to see who is working on what. But what you really need to follow to move a project forward is the column “What to record from now on.”

Why you still need a tool

You might think, “If deciding is what matters, can’t we just talk it through?” Even so, you still need a tool, for three reasons.

1Dates and handoffs still need managing

Launch dates and deadlines do not move because you have AI. Client reviews and handoffs with outside photographers or developers are still part of the job. When, to whom, and what to hand over: this still needs managing after you start using AI.

2More options mean more “Wait, which one did we pick?”

The more options you generate with AI, the more options pile up. Thirty taglines, ten banners. Which one did you choose? And why that one?

You remember when you decide, but two weeks later you will not. If the client says “Actually, let’s go back to the first one,” and nobody can trace which option that was or why it was dropped, you end up discussing it all over again.

3In chat alone, decisions get washed away

Many teams make their decisions in chat. “Let’s go with option B.” “Got it.” It is quick and easy, but the decision is soon pushed out of view by other topics.

When you look for it later, you cannot find the message where it was decided. Keeping decisions in a set place is exactly what a task management tool is good for, including for teams that use AI.

RelatedWhy requests get lost in chat is covered in detail in Part 3 of our beginner series “Your First Project Management Tool,” “Why Do Requests in Email and Chat Get Lost?”

What this series covers

Across six parts, this series looks at how creative teams that use AI can record their decisions and agreements. Here is what Parts 2 to 6 cover:

• Part 2, “Frame Tasks Around What to Decide, Not Just What to Make”: break “Design the home page” into decision-sized tasks such as “Choose the hero image direction from two options.”
• Part 3, “Make ‘Waiting for a Decision’ Visible”: separate tasks waiting on someone inside the team from tasks waiting on the client. Start your regular meeting with “What we need the client to decide this week.”
• Part 4, “AI Drafts Are Decision Material, Not Finished Work”: record which AI-generated options you kept and which you dropped, so you can trace the reason when someone says “Go back to the first one.”
• Part 5, “Keep Work Moving While You Wait for Client Approval”: instead of sending review requests piecemeal in chat, send them together with deadlines. Also, when to invite the client into the tool and when to send them a list instead.
• Part 6, “With Good Records, AI Can Help Even More”: use your records with AI, for example to draft tasks from meeting notes or a weekly report from your task list, and decide how much client information to give an AI tool.

For what the AI features built into task management tools cost, see our separate article, “How Much Do AI Features in Project Management Tools Cost?”

Summary

• What AI sped up is the “making”: wireframes, copy options and the like
• The “deciding” does not get faster: agreeing on direction inside the team, client reviews, judging change requests
• Projects stall more often because they are waiting for someone’s decision than because work is behind
• What you record shifts from working time and percent complete to decision deadlines, who decides, and what was decided and why

What AI changed is not the amount of work but where the time goes. When a task management tool grows from a list of work items into a record of decisions and agreements, it makes work easier to move forward alongside AI.

Next time, we will look at how to frame tasks around what to decide rather than what to make. From Part 2 onward, we will show practical examples using GeckoFlag screens.