OpenAI is winning the GenAI race, but is it leaving everyday users behind?
Enterpret is in the customer feedback business; we help companies across a variety of industries (I work with a bus company and a dating app) understand what their customers are saying in real time, mostly through internal surveys, review, and support tickets. Sometimes, we like to scrape public data to test out our models and see what we can learn.
OpenAI is one of the buzziest companies around, and they have generated a mountain of feedback and commentary. Their iOS and Android apps receive thousands of reviews each month. On top of that, there’s an official Discourse community and a sub-Reddit, both of which are replete with developers comparing notes on esoteric topics like how to manage rate limits.
What are users saying about OpenAI?
We wanted to dig into OpenAI's publicly available feedback to see what we could find in the last three months.
Top OpenAI Feature Requests
Back in December, CEO, Sam Altman, asked all users what they wanted OpenAI to build/fix in 2024. We decided to look at a holistic view of feedback related to feature requests across all channels.
We generated a summary to dig into the feedback behind the top 10 requests.
Here's the summary from Enterpret. Note: Each summary in Enterpret includes annotations that allow us to dig into the underlying feedback organized by the top reasons.
The feedback for OpenAI over the past three months indicates a strong desire for improvements across various aspects of the application. Users have suggested enhancements in the app's user interface, image generation capabilities, ChatGPT functionality, and the frequency of updates. They have also expressed a need for additional features and improvements in the search function.
- Improving App Functionality: Users have expressed a need for improvements in various aspects of the OpenAI app, including the user interface, regular expressions, handling of limitations, and model performance. They have also mentioned that the app is verbose and could benefit from further adjustments.
- Enhancing ChatGPT: Users have provided various suggestions for the Ability To Enhance ChatGPT. They have requested specific improvements such as accommodating desktop users with a wider text area, providing a free trial, including international voices, information, and vocabulary, and addressing the issue where they have to ask multiple times to get a satisfactory answer.
- Improving User Experience: Users have expressed a strong desire for the Ability To Enhance User Experience. They have suggested improvements to the user interface of OpenAI's products and expressed frustration with the complexity of basic queries and the need for improved user-friendliness in the software.
- Increasing Update Frequency: Users have expressed a desire for the Ability To Increase Update Frequency. They hope for regular updates and improvements in the AI's level. Some users have specifically mentioned that the last update was in 2022 and have requested for more frequent updates .
The Impact of OpenAI
The launch of ChatGPT made a big impact because it made a technology that had quietly been building momentum suddenly available to a huge swath of people. That’s why we had to endure all the think pieces about the various institutions ChatGPT would disrupt/change/ruin and why Twitter/X is full of blue-checks trying to give you AI tips.
Behind all that noise, the OpenAI really did have a big impact on a lot of people’s lives, and you can find some pretty touching testimonials to that effect.
Looking under the hood of OpenAI feedback
We decided to look at the feedback beyond feature requests. What we found based on the available data is increasingly, OpenAI seems to be leaving their everyday users behind. You can see some rumblings of this in a couple of seemingly unrelated topics when you look at the Anomalies (spikes in feedback) within the top 10 complaints:
- Content Restrictions
- Customer Support
Understanding feedback around OpenAI Content Restrictions
Let’s talk about the content restrictions first. While it’s inevitable that any product company will get loads of vague “it got worse” reviews, there’s a real increase in the number of users sharing how they are getting tripped up by tightening content restrictions. One user noted that “friendly memes” just aren’t funny to them, to which I have to say: my dude – the whole rest of the internet is there for you, don’t worry.
But there are also a growing number of users sharing examples of how totally benign prompts, including sometimes one they have used in the past, are now hitting up against content limits. The examples can be pretty absurd.
Why is this an issue for everyday users? Because the restrictions, by and large, are on the ChatGPT tool, but not on the underlying APIs. Larger corporate customers or tech-savvy developers are already plugged into the APIs, and they can mostly sidestep these issues (or some have come up with creative ways to “jailbreak” the bot).
Taking a Closer Look at OpenAI Support Issues
And then there is the issue of support. There are loads of users writing in and sharing experiences of their challenges reaching OpenAI support, even if they are paying customers.
Is OpenAI leaving everyday users behind?
This feedback isn’t surprising since OpenAI blew up pretty suddenly and is being pulled in multiple directions, supporting some major large corporate customers while also providing services to legions of individual users. (No shade – I know that the OpenAI team doing their best. I’m part of the customer success team here at Enterpret, and I know the struggle is real.).
Both of the issue of content restrictions and support responsiveness are now in the top 10 complaints for the last few months, and they are on the increase.
Is the dissatisfaction of these everyday users a critical business threat to OpenAI? Maybe not, and in fact, we agree with the idea that you should consider relative revenue impact when you look at your customer feedback and recently launched a feature that helps you do just that.
But, in the feedback I reviewed, but there are indicators that there is a narrative forming that the app is losing utility if you are a chat user and that if you have issues (with chat or the API), and it’s increasingly challenging to get any attention unless you’re one of the big guys.
To the OpenAI team we’ve got an account ready for you all with some of your public feedback so you can see Enterpret in action. Email us at firstname.lastname@example.org or get in touch - we’d love to show you more of what we found.