In the race to enable enterprises with agentic artificial intelligence (AI) assistants, the push to provide intuitive, human-like – yet unbiased and trustable – features is paramount. Styling itself as an agentic AI platform development specialist, Moveworks has now come forward with Quick GPT.
Moveworks’ latest agentic AI Assistant is a productivity tool for human workforces to automate tasks and find answers when questioning the company data repository.
Moveworks is specifically aligning the power of foundational Large Language Model (LLM) resources in a more quickly composable consumable format. This is hoped to be of particular use where enterprise users are exposed to sensitive corporate data. In other words, teams can “invoke” ChatGPT inside the chosen AI assistant where guardrails and other policy-based limiting controls may already exist.
You Asked the AI What?
In a world where an increasing number of knowledge workers actively use ChatGPT and other interactive agentic AI tools in the workplace, there is a clear need for security standardization. There have of course been many reports of enterprise employees – often senior knowledge workers – operating with a sort of blind faith in these new tools; many have entered sensitive data into the interface like code, financial projections, meeting minutes etc.
A large number of organizations reacted by banning the use of ChatGPT within the company entirely. Quick GPT is a feature designed to give these users a safe way to harness the power of LLMs without compromising enterprise security and requirements.
“Quick GPT helps organizations across all sizes and industries deploy AI while guaranteeing the protection of sensitive data and ensuring governance and compliance through industry-leading analytics and insights,” Varun Singh, president and co-founder of Moveworks said. “[We have] recognized the tension between employees’ growing demand for ChatGPT’s capabilities and the security requirements IT departments must uphold. Quick GPT empowers employees with the functionality they are used to leveraging every day while also providing the enterprise-level controls companies need.”
Singh says that Quick GPT unlocks a variety of productivity use cases for employees inside the AI assistant they already use today. For example, engineering teams can conduct code reviews to accelerate the speed of software development, sales teams can research target accounts and customer news for insights, and HR teams can generate first drafts of job descriptions or internal communication for improved efficiency.
Invoking ChatGPT
Sounding rather like a practice drawn from some kind of black magic, the invocation of ChatGPT via Quick GPT within the Moveworks AI assistant happens in two ways.
On one level, ChatGPT can be brought forward as an invokable plugin. When users want to intentionally invoke ChatGPT and access it, they are able to directly instruct Moveworks to use GPT-4o today. Other advanced models will be accessible as they become available, to generate a response.
NOTE: OpenAI’s GPT-4o is the company’s flagship multi-modal language model, launched in May 13, 2024. The lowercase “o” stands for omni to denote multi-modal modularity and malleability. It supersedes the 2023 release of GPT-4 Turbo.
Backstop Cushioning
Alternatively, at a second level, ChatGPT can be used as a “backstop” – a term used to define an AI assistant safety net function that kicks in when a primary AI assistant is unable to complete its job.
In this scenario, Quick GPT provides answers to questions when curated sources within an organization are not able to help. It prioritizes answers based on business-verified information before offering information from general knowledge contained within LLMs for instance.
“With the addition of Quick GPT, Moveworks continues to strengthen its all-in-one platform for employee support, helping them find answers, automate tasks, and be more productive at work,” enthused Singh and his team. “With Quick GPT, AI Assistant answers are reference-able and grounded in a company’s trusted knowledge or they can tap the power of frontier LLMs for general research, content creation, task automation and more.”
Work-Appropriateness
The company says that the key benefits of Quick GPT within the Moveworks AI Assistant include a unified enterprise AI assistant for a wide range of workflows. Quick GPT enables access to GPT capabilities in the same interface powering search and automation across enterprise apps. Built-in tools allow admins to set parameters, work-appropriateness filters and custom topic inclusion/exclusion to enable organizations to manage GPT access and experiences.
Rounding out the functionalities here, there are also built-in analytics for transparency and compliance. Admins are able to track query types and usage patterns, supporting internal compliance and offering visibility into how Quick GPT is being harnessed across the organization to accelerate work.
Quick GPT has a zero-day data retention policy with OpenAI. Quick GPT does not use employee inputs to train global generative AI models nor does OpenAI with their global machine learning models.
