OpenAI has launched its GPT Store, a library of ChatGPT-powered bots – or “custom GPTs” – that developers have designed to perform particular tasks.
Already, OpenAI offers custom GPTs via its paid ChatGPT Plus service. However, the store also allows citizen developers to build, market, and monetize their own GPTs, which will – in theory – improve the more people use them.
As developers build these, they may “hard code” the GPT to have a particular personality and perform specific tasks within various applications.
Some of those already available on the store include GPTs that design presentations and social posts in Canva, suggest personalized trail recommendations in AllTrails, and search and synthesize results from 200MN academic papers with Consensus.
Yet, there are many more. Indeed, OpenAI suggests that users have already created three million custom versions of ChatGPT over the past two months.
Meanwhile, the company noted that it will spotlight “useful” GPT bots within the store every week.
What does the GPT Store Mean for Contact Centers?
In the future, OpenAI’s recommended bots may include custom GPTs for contact centers.
There are already several such non-endorsed GPTs, as the following screenshot suggests.
Currently, they seem little more than domain-specific knowledge bases – not actionable GPTs that will generate useful content inside popular customer service applications.
Some may suggest that could be what’s to come, as custom GPTs may help to automate various contact center tasks within popular customer service applications.
Yet, the providers of the apps will likely get there first. So, the store isn’t likely to have a significant impact on the space – especially in the short term.
Also, remember that contact centers aren’t environments where businesses can throw new innovations at the wall and see what sticks. Everything requires meticulous planning, and most brands will only work with trusted tech providers, not random citizen developers.
What the store does perhaps suggest is that vendors may begin to build custom LLM-driven bots to help customers achieve specific tasks with their platforms – similar to what Verint is already doing.
Yet, they won’t always use ChatGPT. After all, GPT is only one LLM, and other models will provide greater accuracy rates across various use cases.
Many CX vendors have already cottoned onto this and opened up their GenAI applications so contact centers can plug in their preferred LLM.
Zoom has perhaps gone the furthest with its federated GenAI approach that ranks LLMs based on their performance across several use cases on its platform.
Nonetheless, contact centers may still struggle with the massive computing power these conventional LLMs require.
As such, smaller LLMs, customized for specific tasks, which contact center providers host in their own data centers, is the most likely direction this is heading in.
These may then come under a use case library that contact centers can pick and choose from – similar to what Salesforce has created with its Einstein Copilot studio.
But don’t expect a brand like OpenAI to suddenly release a recommended GPT that will transform the industry or a ubiquitous chatbot that can respond to any customer query across any sector.
Indeed, OpenAI doesn’t even use a natural language chatbot for customer support. Instead, the company puts a stilted, option-tree-based bot in front of its customers – as the following LinkedIn post exemplifies.
However, the release of the GPT store isn’t insignificant. Instead, it provides a signpost for the market’s direction that seemingly validates Verint and Salesforce in their respective GenAI-infused, specialized bot and Copilot Studio strategies.
