ChatGPT's success in enterprise IT comes as no surprise to Miqdad Jaffer. The GenAI giant's head of product says enterprise has always been at the core of its product offering, but the stakes are higher as companies race to capitalize on their competitive advantage. I am.
Generative artificial intelligence (GenAI) went from being the talk of the tech world to a global household phenomenon in 2022 when OpenAI announced ChatGPT. Business leaders quickly recognized the value of finding use cases, and GenAI has been on the C-suite's radar ever since.
So it was no surprise when OpenAI started being released. Last year, we launched our own enterprise product.first target large companies, then Expand your services to small and medium-sized businesses. Highlighting our growing portfolio of enterprise products. This week's blog postOpenAI touts security improvements, an assistant API, improved management controls, new pricing options, and more.
Jaffer told InformationWeek that the company's efforts will continue to rapidly evolve to keep pace with the rapidly changing enterprise AI landscape. More than 2 million developers from hundreds of companies use OpenAI's API platform, Jaffer said.
(Editor's note: The following interview has been edited for brevity and clarity.).
Can you give our readers a 10,000-foot overview of OpenAI's corporate efforts and how the service is evolving?
We've been working on Enterprise since the beginning, and I think it was important that we wanted to focus on Enterprise as a package because it just wasn't obvious. [as a standalone enterprise product] To as many people as possible. What we've already seen since ChatGPT's launch is that 92% of Fortune 500 companies are already using ChatGPT in some way. What we are currently announcing in this version are changes from a pricing and cost perspective. We introduced a batch API that allows you to perform asynchronous requests. We also introduced provisioned throughput, which allows you to commit to a certain amount of throughput and receive a discount. . We want more people to be able to build with our APIs and not have to worry about spending more money. Another part of the release is about actual functionality. The new Assistant API improves your ability to follow instructions and search functionality. File limit changed from 20 to 10,000, allowing users to build more real-time applications. In terms of actual company management, projects are a way to create a hierarchy within an organization, allowing each company and even individual developers to isolate their work and set separate fee limits and separate cost limits. Now you can actually monitor your management. Towards the introduction of AI.
Recently, new enterprise AI products have been released from various vendors. How does OpenAI differentiate its enterprise products from its competitors?
There are several different ways of thinking about it. When you think about your overall enterprise platform, think in terms of models, products, and platforms. If you think about the models, I think we are quite differentiated in that respect. [OpenAI’s competitors] We are catching up with a model that completed training three years ago. And we intend to release more and more new models that push the boundaries of what's possible. And what's important for us is to keep in mind the full level of intelligence of these models. In some cases, it increases the intelligence level on some frontier models. In other cases, it improves the efficiency of existing models and reduces costs by improving latency. We have a model that differentiates us. But from an enterprise packaging perspective, it's the extras that come with it that make the difference.
I know that the use of GenAI in the enterprise is still in its infancy, but could you talk a little bit about how the market is already changing? What about growth? Are small businesses keeping pace with some of the Fortune 500 companies?
The way we think about our products is we don't build one thing for enterprises and another thing for startups and individual developers. We've strived to build a platform that gives you the tools to keep growing as your landscape changes, whether you're an individual developer or a very large company. Therefore, as the situation becomes more complex, all the capabilities that are available to the company need to be made available to everyone else. From a market perspective, we recognize that every aspect of the market is growing, and we are building ways to ensure our customers, no matter who they are, get what they need. .
How are you approaching responsible AI for enterprise products? Can small businesses expect the same level of security as their large enterprise customers?
All of our APIs, all of our safety checks, all of our compliance and safety requirements are for everyone across the board. And we spend countless hours making sure they're perfect before we show them to anyone. And it's important to us that AI is deployed safely. If that means it takes longer to release the model, so be it. We're not trying to push the boundaries of what people can build if they can't build it safely.
I know that OpenAI is working carefully on its Text-to-Speech voice cloning model (Voice Engine). But do you think it will become a big part of enterprise services in the future?
I think this will be part of the overall product roadmap. Obviously, that's where things can get a little dicey from a release standpoint. Safety is paramount here, and we are working closely with government agencies to define what use cases exist and how to deploy them. And we're not going to let it out as is until it's completely safe.
Where do you see enterprise AI growing in the future?
We are very bullish on the concept of agent workflow. [Agentic AI refers to AI systems that can autonomously pursue workflows with limited human supervision]. We think this will be a major change in the way people adopt and think about AI. These agents will be able to help organizations with employment issues expand. Being employable in all these areas is not always easy. That's why it's so helpful to have an agent who can augment your existing work base. Imagine having a software engineering agent when you need help with a deadline. It helps if you have that ability. We also see the industry's push towards text intelligence. This is core to our belief that improved text intelligence will improve the overall capabilities of these models. We will continue to invest there.
