Internal Gen AI Assistants Gain Popularity Among Companies, Including JPMorgan and Walmart

• JPMorgan Chase and Walmart, which previously limited employees' access to generative AI, are now permitting its use.
• Building a guardrail-equipped internal solution helps deal with privacy concerns.

JPMorgan Chase CEO Jamie Dimon, after eighteen months of restricting employee use of generative artificial intelligence solutions like ChatGPT, introduced an AI assistant developed by the company. The solution, named LLM Suite, is built on the technology of ChatGPT's creator, OpenAI. The service has already been released to 60,000 employees for tasks such as writing reports and composing emails.

The trend of restricting employees' use of gen AI to building a guardrail-equipped internal solution is becoming increasingly common among companies of all sizes as they attempt to harness the power of the technology. In fact, over a quarter (27%) of organizations have temporarily banned public gen AI applications, and the majority of businesses have imposed limitations on the solutions employees can use and how they can be used, according to Cisco's latest Data Privacy Benchmark report.

A recent report from Veritas, an enterprise data management company, highlights the main concerns about an employee's use of gen AI, including the leakage of sensitive information, hallucinations, and compliance with industry regulations.

The outputs that AI platforms generate are not randomly generated but are based on user inputs that are stored and used to improve generative AI capabilities. This can pose a risk to sensitive information about a company or its customers, leading many organizations to implement bans or restrictions until they can better manage the technology.

Walmart's approach to employee AI

David Glick, senior vice president of Enterprise Business Services at Walmart, stated that they have been proactive in establishing guidelines for their AI usage. These principles encompassed listening to associates' needs for assistance, simplifying complex information, aiding corporate employees in utilizing the ERP system, and automating specific tasks.

Glick is focusing on smaller, more precise projects, such as Walmart's use of gen AI to enhance the search and support capabilities of its Benefits Help Desk team.

Keeping data from going 'external'

Ensono, a managed service provider, has changed its approach to employee access to gen AI tools. Tim Beerman, the company's chief technology officer, has decided to restrict employee use of large language models like ChatGPT to protect sensitive data. However, Beerman still wants to utilize the "vast amounts of unstructured data about our company to provide value to our associates without letting this stuff go external," he stated.

Ensono's internal AI assistant was rolled out to all 3,500 associates during the summer, utilizing GPT-4o as its foundation while maintaining the ability to incorporate different language models depending on the type of data being used. "Our associates now have a unified interface to these tools," Beerman stated.

Ensono is currently developing small language models for specific use cases, such as root cause analysis and RFP responses. The key to success is model flexibility, as Beerman emphasized that what is true today may not be true 12 months from now.

The internal AI platform of Varyence and its associated startups automatically prompts data classification when someone creates data such as emails and Word documents. The built-in guardrails of the platform ensure the security of non-public data, while also providing an opportunity for employees to experiment.

Gen AI is highly sought after by everyone, as they want to explore its potential to enhance their department's efficiency, according to Hishmeh. However, banning its use is only a temporary solution, and companies of all sizes must eventually provide safe solutions for employees to work with.