Will ChatGPT Enterprise cause a disruption in Large Language Models (LLMs) space?
On Monday, 28th August 2023, OpenAI (OpenAI 2023) had a press release launching ChatGPT Enterprise, a comprehensive ChatGPT variant designed for enterprise-level applications which aims at top-tier security and privacy, unlimited higher-speed GPT-4 access, longer context windows for processing longer inputs, advanced data analysis capabilities, various customization options, and lots more. This has major improvements from the earlier versions aiming at helping teams be more novel and efficient. However, currently they are offering APIs if you choose a custom solution similar to Microsoft. Work in progress includes on pushing customizations (integrating existing applications that the business use) with added focus on security and privacy regardless of team size. They want to offer a more powerful and improved version of Advanced Data Analysis and allow data browsing that is finely tuned for professional tasks. In addition to providing tools catered to specialised roles like data analysts, marketers, customer support, and beyond.
The alpha and beta users of ChatGPT Enterprise included industry spearheads such as Block, Canva, Carlyle, The Estée Lauder Companies, PwC, and Zapier. They are restructuring their operational methods and employing ChatGPT to create more lucid communications, speed up coding assignments, swiftly investigate solutions to intricate business inquiries, aid in creative tasks, and accomplish various other purposes.
Now-a-days, the safeguarding of data using top-tier privacy, security, and implementation solutions at an enterprise level is a non-compromise point. The business data remains under the host organisation’s ownership and authority within ChatGPT Enterprise. The system does not utilize the host’s data or dialogues for training, ensuring that ChatGPT’s Enterprise models don’t gain that knowledge. Adhering to SOC 2 standards, ChatGPT Enterprise maintains compliant, while conversations are safeguarded through encryption both during transmission and storage. The latest administrative console is designed to facilitate effortless team management.
In ChatGPT Enterprise, all limitations on usage are eliminated, and performance is elevated by up to twice the speed. The 32k context capability (already available in OpenAI) within the Enterprise version enables users to handle inputs or files that are four times longer. Additionally, ChatGPT Enterprise furnishes limitless access to advanced data analysis, previously referred to as Code Interpreter. This tool empowers both technical and non-technical teams to swiftly analyse information. The pricing includes complimentary credits to use their API if organisations opt for a personalised solution which includes a new shared chat templates foster collaboration and streamline common workflows.
It looks like OpenAI is making a shift from a focus of casual customers to more enterprise especially B2B customers since its launch last November. However, this is not new for OpenAI as their major financial partner is Microsoft and ChatGPT is available via Azure OpenAI Service, Microsoft’s popular cloud service. However, this is different as it is launched as a detached version.
Now, let’s look at Azure OpenAI Service. It grants access to OpenAI’s robust language models such as GPT-4, GPT-35-Turbo, and the Embeddings model series supporting natural language processing and generation. This can be accessed through a REST Application Programming Interface, Software Development Kits, and Studio User Interface. This service combines the power of OpenAI’s models with the security features of Microsoft Azure and an option for fine-tuning. It offers options like private networking, regional availability, and responsible AI content filtering (hate, sexual, violence, and self-harm from safe to high). They provide a new complementary course online called ‘Develop generative AI solutions with Azure OpenAI Service’ gives a kick-start to developers to jump to creating solutions with Azure OpenAI Service (responsible Generative AI module included) (Kasper 2023, Boyd 2023, Hawk 2023 & Microsoft Azure 2023).
Azure OpenAI Service emerges as an essential asset for enterprises aiming to leverage the vast capabilities of language models. Offering the provisioned throughput feature, this service empowers clients to establish their preferred usage thresholds using throughput units, which can then be allocated to the specific OpenAI model they select. Nonetheless, it’s important to note that the financial investment can be substantial and is contingent on factors including the model type, size, and usage frequency. The implementation of an efficient chargeback system yields multiple benefits, such as increased responsibility, clear and transparent cost allocation, and prudent resource utilization across the organization (Sinha 2023).
Azure API Management offers a comprehensive solution, including security, governance, Responsible AI features, and Generative AI apps, to address the challenges associated with adopting new technology (Kasper 2023, Boyd 2023, Hawk 2023 & Microsoft Azure 2023).
Microsoft Azure is focusing on helping to alleviate the apprehension caused AI by creating trustful technologies built responsibly and ethically while ensuring the uptake of new possibilities around Generative AI (Tezcan 2023). With the launch of ChatGPT Enterprise as a standalone version, it would be interesting to see how Microsoft’s customers react as they no longer need to subscribe to Azure to access ChatGPT.
I speculate that there might be a divide amongst the customers that use or subscribe to (1) a whole suite of Microsoft products and services (2) ChatGPT services with APIs. Those in the former category is more likely to stay since they are dependent on staying with Microsoft, while the latter category might stay or leave Microsoft. Both ChatGPT Enterprise and Azure OpenAI Services offer complimentary credits to use their APIs if a fully customised solution is opted for. The choice would depend on the business’s needs and goals of the organisation.
In summary, integrating ChatGPT into your organization’s data requires careful consideration. While a fully customisable solution can be obtained through OpenAI or Azure OpenAI services, the latter option offers the added benefit of Azure’s security measures. Would this cause a closer and more stronger relationship bond between OpenAI and Microsoft or would there be a rift developing as the market is competitive?
We would have to wait and watch. Stay tuned to Version 1’s blog!
About the Author:
Rosemary Josekutty Thomas is Senior Technical Researcher at the Version 1 AI Labs & Innovation.
