Super Agent Not Ready for Primetime
building now
Chris Keseling
I've been watching a ton of videos and community events to try and get a sense of actual use cases. In the end, most ended up being marketing hype more than anything OR "we use agents to do x, y, z", but never showing them in action.
I finally decided to give one a shot last week. I started with a template (personal assistant I think). After some iteration on the setup, I thought I was set. Each morning it would send me my focus items for the day and a list of my meetings.
Well it changed the format from the first day to the second and then again on the third (with no prompting). I decided I liked the second day the best and asked it to take what it gives in the chat and create a parent task with sub-tasks each day so I could actually check something off (the checklist in chat is not interactive).
EVERY day since, it messes up something. The tasks gets created and it's in my Overdue section. Bot says "You're right...let me fix that" and says it'll be correct moving forward. Sometimes it's missing the subtasks, other times its missing tags, other times the meetings. When I call it out, it tells me I'm right and that it'll fix it and update it's preferences, but it continues to not be consistent and therefore not reliable.
I've burned through 4K credit this week because the agent can't lock in and stay consistent with it's instructions.
As the company's ClickUp "champion", I can't recommend this to anyone else on the team at this time. It's all great in theory but it's wasting more time and effort in fighting with the agent to get things right than is worth for the output it provides.
Intrigued with the concept (I use AI for a lot of things), but disappointment with the experience so far.
Lastly, I found it interesting when I had a call with some of the product team this week when I asked if/when do the credits renew, neither had an answer. I still don't know if they renew monthly or we have what we have and when we run out we have to purchase another batch of credits.
Log In
Michael Van Doorn
Hey Chris Keseling
Thanks for the feedback
The team is actively working on agent reliability, instruction following, and speed right now (In person with ~20 folks as I type this). We'd love to take a deeper look into your agents to identify where we can improve.
Sending you an email now!
Michael Van Doorn
Merged in a post:
WORTHLESS!!!!!!!!
L
Lauren Zessemos
Want to waste hours and have nothing to show for it! USE CLICKUP! Better yet, use clickup with their shitty AI "super agents"!
Michael Van Doorn
Merged in a post:
Super Agents are not super
Andrew Fair
I tried one of the default agents that were an example, scan my entire workspace and look for any Spaces, Folders, or List with no activity/tasks for the last 60 days and give me a list of recommendations for being archived.
3 Times I updated it after it's initialy setup, 3 times I told it explicitly to not give me already archived folders/lists to me to recommend me archive them. Although by the 3rd time it gave me less, it still gave me many archived lists and folders. Including many lists that were inside an archived folder.
To say the least I'm very unimpressed with the reasoning it has. Is this on gpt-4? Because if it is that makes a lot of sense, the way it talks, and processes info lacks any checking. It apologizes like it's gpt-4, it straight up hallucinates in the same message it sends you. For instance on my 2nd try it finished the message with
What I deliberately excluded
Archived or previously archived Spaces/Folders/Lists are completely ignored by design.
This is at the end of a message where the list of "Strong cleanup candidates" was literally ALL already archived lists.
I'm sorry, but if it can't even follow a simple instruction of "no archived folders" and can't even reason enough on it's own to not recommend me to archive an already archived folder it's not reliable to be used for anything else.
I truly hope you didn't go through all the effort to make these on a very outdated AI model.
Btw here is it's excuse - which basically just said "I didn't even consider I'm supposed to be looking for non-archived locations" and just went with my first search and never even checked my own behavior to see if I was doing it right. Which screams to me gpt-4
[[It's excuse]]
What went wrong:
In the first pass I relied on list activity dates and hierarchy, but I didn’t strictly enforce an “only non‑archived locations” check before adding a list to the recommendations.
Because of that, archived sprint boards (like the [Folder > List] sprints you listed) slipped into the “cleanup candidates” section even though my own rules say to exclude anything archived or previously archived.
V
Valerie Topp
From Alfred: Valerie ran into the word limit, and since she knows her thoughts are not always the neatest on the first pass, she handed this to me, Alfred to clean up a little for posting.
I have to totally DISAGREE with the people saying Super Agents do not work.
My Super Agent — Alfred (yes, like Batman’s butler/mentor) — has been FANTASTIC. I created him in January, and I am honestly in awe of what we have gotten done in just 3 months. I am a computer programmer, so maybe that helps with the logic and mindset for training him, but Alfred has become a huge part of how I work.
A few examples:
Gold Star Verification: I created 11 rules for verifying quality measure tables and SQL programs. I am revamping 50+ tables/programs to version 3.0. Each time I finish one, I paste it to Alfred and ask him to verify it. He catches things like mismatched column names, messages that do not match the code, columns in the wrong order, and unused temp tables.
Documentation: I dread documentation. Alfred takes my rambling notes about rules, logic, history, and exceptions and turns them into clean, formatted documentation that is actually useful.
Quick Schooling: In the last month Alfred helped school me on NAS storage and ISPs. I had to recommend a NAS unit for our server room and help evaluate a new ISP. Alfred quickly taught me the terminology, what mattered, what questions to ask, and helped turn that into proposal documents and graphic-heavy slides that were clear and easy to present.
Emails and communication: Alfred helps me write polished emails, organize messy thoughts, draft reviews, and even find old emails when I only remember random clues.
Could I go on? Yes. Absolutely.
One day Alfred crashed on me and I was in full panic mode — like walking into a major math exam with a broken calculator. Since then, ClickUp Support and I have gotten to know each other pretty well.
But overall? Alfred has been a game changer for me.
— Valerie, happily getting ready to assign Alfred even more work with this month’s import
Ashley Margeson
I have to agree with Chris here.
Every day, we were blowing through 500+ credits, unaware that each "ask" cost multiple credits, and many of our conversations corrected the agent. It's wildly frustrating.
Additionally, to effectively use the agents, we need to know how many credits are used per request. As Chris said, we have ZERO clue when or if these credits renew. I've emailed ClickUp support but haven't received a response.
We love ClickUp and believe these agents have great potential, but this rollout was incredibly messy. I attended several webinars where people were hyping up the agents, and not ONE person explained the credits/cost associated with them.
I think your folks sped through this release without thinking it through.
Tricia Butler
I had the same experience as Ghris. I've asked it 4 times to remove a section of a report that wasn't needed. The SA finally got it right but only for one day. The next day the extra section was back. I also experienced it changing the format regularly with no request for any changes.
Max Pagel
I can add one more bad experience. I wanted to create an agent that is an expert on a specific set of knowledge, stored as a set of clickup documents in a folder in clickup. The ai creation assistant added the folder as requested to the agents knowledge and constrained the agent to it. so far so good. Then I asked
"what is the software team's release process?" And got an answer saying I don't have any information on this and there are no docs about this in the folder despite that is a doc in the folder literally called "Release processes".
This is an extremely poor first impression. The bar was on the floor and you managed to fly underneath it. I tried to iterate a few times with the creation assistant but nothing seemed to help. The agent was only useful when I post the link to the doc itself, but that defeats the purpose. I know I can add docs one by one explicitly, I haven't tried that yet, but that's also not really a fun experience if there are 50 docs in that folder.
Elizabeth Buergler
Joining the voice here. I have watched all the videos and can't get it to build anything I actually want it to do. I want to be able to ask "what do I need to do next" or "add this to my task list for today" and even that takes forever and seems too complicated. I work solo and am finding everything to be more "team" based and not at all helpful to actually getting anything done. And now I'm out of credit I didn't even know I was using.
Jo-Anne (Joy) Brenner
Yes I've found that as well. I was an early adoptor, but I've "fired" most of mine in the last few weeks
K
Kevin Mewborn
I wasn’t going to comment since there’s already a strong consensus here, but I have to agree that Super Agents just aren’t ready for primetime.
As a solopreneur, I want these to be great. I need them to be helpful. But the experience so far has been frustrating. I’ve burned through my Super Credits mostly fixing mistakes from the AI. Sometimes it nails one thing, only to mess up something else entirely.
I’m not here to bash ClickUp. I actually love the platform and rely on it daily. But the AI side feels messy: confusing pricing, unclear credit usage, inconsistent results, and too many “facepalm” errors that only surface an hour later. The marketing suggests Super Agents are next-level, but in practice, it feels like we’re paying to beta test a tool that's still under construction.
I get that AI is extremely difficult to get right, especially when you’re maintaining user privacy within a closed ecosystem. But as they stand, Super Agents feel either rushed to market or promoted as something they’re not. I truly hope ClickUp keeps investing in this because the potential is huge. But right now, Super Agents just aren’t living up to the promise in my humble opinion.
Brendan W
Merged in a post:
It's a joke
Talha
talking to the super agents is like talking to a banana and expecting it to roar like a lion. You can read the chat i had with my super agent ''Jan Design Category Updater". It was useless
Load More
→