Remember how everyone was wondering whether the GPT Store launch was OpenAI’s Apple App Store moment?
It’s safe to say that six months later, neither builders nor consumers feel like its lived up to the hype. Yet.
There is growing disenchantment towards OpenAI in the builder community. What happened to the usage-based revenue share that was slated for Q1? And despite the 100M+ weekly active ChatGPT users, builders are finding it hard to drive usage to their custom GPTs.
As for consumers, custom GPTs have not immediately captured public imagination the way ChatGPT did. Until June, only Plus subscribers could access the GPT Store. The explore page is static and generic making discovery hard. Seriously, Wolfram’s GPT has been pick of the week since January. Also, the name ‘custom GPTs’ does not help. (Is ChatGPT-powered apps any better? I experimented with ‘ChatGPT apps’ but my mum thought it meant using ChatGPT on a mobile app 🤷♀)
Yet, there is an undisputed consumer need for an application layer to harness the full potential of generative AI. I’ve run quite a few workshops this year, and know how few people know how to prompt well, can manage hallucination risk or have any idea what to do with generated output (okay cool, it generated a personalised greeting card message for my best friend… now what?).
Teething problems notwithstanding, we at Top Road are still bullish that OpenAI’s GPT Store will win. The competition in the generative AI space is stiffening in business and enterprise segments, ChatGPT is king in the consumer space.
Some tweaks are clearly necessary, but the OpenAI Store remains the strongest contender to be the watering hole for gen-AI powered apps because of the potential to convertChatGPT users into custom GPT users.
As for builders, the out-of-the-box functionality of custom GPTs is truly impressive. You can layer text generation, implement RAG via knowledge files or API connection to deliver more accurate and relevant results, and have access to useful pre-configured tools (DALLE, web browsing, ability to run Python code). For those who want to build more complex GPTs, custom actions are indispensable. It’s helpful that builders do not have to pay for API token usage, especially in these early days when monetisation is uncertain.
So what needs to happen for the OpenAI GPT Store to live up to its potential?
For OpenAI:
- Roll out revenue share — a lot of builders flocked to the platform in good faith that OpenAI would honour its promise to roll out usage-based revenue share. Builders need to a financial incentive to keep building.
- Improve discoverability — suggest relevant GPTs to the end user during a ChatGPT conversation, improve the search algorithm and make the GPT Store explore page for personalised and dynamic (pretty sure Wolfram’s GPT has been pick of the week since Jan 10th…)
- Champion independent builders — there is an evident bias towards big companies building GPTs. I am a huge fan of Canva but their custom GPT is subpar, and yet, it’s been cherrypicked as a trending GPT for months. I don’t blame Canva, it makes far more sense to focus on embedded AI features in their core product which generates revenue. We see so much more creativity and talent coming from individuals.
For builders:
- Keep the faith — we are in the early innings of the generative AI revolution and there are benefits from being early adopters. Also, know you’re not alone 🙂
- Don’t wait for OpenAI — you can monetise today! You can add in-chat payments with GPT Builder Tools (shameless plug #sorry #notsorry)
- Build GPTs people want — sorry, we’re YC alum, so we had to throw this in here. Start by knowing who your target customer is and talk to them IRL or get to know them online (we show user chat summaries and can enable you to request user email addresses) to get feedback and iterate.
Let’s go!
Are you a builder? Sign up for a free GPT Builder Tools account today to track chat history & search keyword results.