AllVoices embraces AI, using ChatGPT across multiple departments

But some HR experts still have privacy and security concerns.
article cover

Francis Scialabba

· 5 min read

It all started with a tweet about removing a peanut-butter sandwich from a VCR. As the tweet catapulted across the web, the nonsensical musing caught Remi Silva’s attention. In it, a screengrab of a paragraph written in King James biblical verse provided actionable advice on, well, removing a peanut-butter sandwich from a VCR.

It was written by ChatGPT.

Spellbound by the machine-generated paragraph, Silva, the VP of marketing at the HR management platform AllVoices, was compelled to dig further into the AI alongside a colleague. “We both [dove] into GPT and started playing around with it and started talking about how does this apply to our jobs?” Silva told HR Brew.

That was last December, one month after the public release of ChatGPT, the generative AI developed by Microsoft-backed OpenAI. Flash-forward to this April, and AllVoices has leaned into the technology, partly automating several facets of its day-to-day content-generating operations, AllVoices employees told HR Brew. While viral hype swirls around ChatGPT and at least one study suggesting generative AI can increase productivity (especially for “novice and low-skilled workers”), some HR leaders have expressed caution, citing concerns about inaccuracy and bias. Still, AllVoices remains invested in using the tool.

Making the case. Before AllVoices started using ChatGPT, Silva had to convince the company’s top brass of the usefulness of Generative AI. “Everyone was very skeptical because AI is a buzzword,” Silva explained. But then the company’s head of product, Jeff Harris, used the tool to write a description of a new product. “Writing user stories and requirements…is usually not the most exciting thing in the world to do for product,” Harris told HR Brew. So, he suggested, “‘Let’s see if we can give ChatGPT some details on a project and see if [it] can write a clear user story and requirements, and [it] did a really, really good job.” Now, Harris’s whole team is “heavily using ChatGPT,” he said.

Across product, marketing, and sales, AllVoices has adopted the tool to save time, according to Silva, particularly when it comes to written content. “I always find writing copy should be simple. But it just takes so long to find variations of certain words and write it down,” he said. The sales department is using it to devise boilerplate outreach templates, then individually tailoring them for recipients. It has made the organization more productive, Silva argued, and also allowed AllVoices to pare down budgets by relying on the AI instead of freelance writers to produce blogs. Now, using ChatGPT to help write the blogs, his marketing team gets them done “within a couple hours, and it’s essentially free.”

Roll your own GPT. Instead of the vanilla version of ChatGPT, the company’s people operations manager, Jenna Birkmire, uses a generative AI based on OpenAI’s GPT-3.5 and developed internally at AllVoices, designed specifically to help with workplace issues. The tool, called Vera (virtual employee relations assistant) behaves similarly to ChatGPT.

Quick-to-read HR news & insights

From recruiting and retention to company culture and the latest in HR tech, HR Brew delivers up-to-date industry news and tips to help HR pros stay nimble in today’s fast-changing business environment.

Birkmire said it can be difficult to comb through the internet to read up on the ins and outs of various employment statutes and how they differ. Vera has allowed her to cut out the most laborious aspect of that process, she said. “I’m an employer in California, but I have an employee in Ohio. What are the laws that we need to follow when they go on maternity leave?” she said, citing a question she has posed to the AI.

Two months in, Birkmire said she occasionally turns to Vera as a kind of compliance dictionary or search engine. “We get cases from employees or reports from employees, and being able to read those and copy and paste it into [Vera] and be like, ‘How would you handle this from a legal perspective?’ is really awesome.” Birkmire said she only consults Vera about 10% of the time and always double-checks answers by referencing state government websites, online resources provided by SHRM, and her HR advisory group. Anything that Birkmire gleans from Vera that eventually makes its way into AllVoices’ company handbook “goes through our counsel,” she said.

Asha Palmer, SVP of compliance solutions at Skillsoft, told HR Brew that such a strategy poses potential risks, because so far, “There’s no accountability of accuracy or transparency of data set for these generative AI tools.”

Privacy and data concerns. Concerns over data privacy and intellectual property are high on the list of ChatGPT fears for HR leaders. A Gartner poll of 250 HR leaders across a variety of industries found that 59% were primarily concerned about data privacy when using the tool. “When we ask the tool to answer questions or perform tasks, we may inadvertently hand over sensitive information and put it in the public domain,” Uri Gal, a professor of business information systems at the University of Sydney, wrote in February.

Many HR leaders have drafted guidelines for usage of the tool at work, while other big employers such as JPMorgan Chase, Samsung, and Verizon have banned it. AllVoices has not yet provided usage guidelines, Silva said

For Bryan Dyer, the founder and chief career officer at career coach HireJourne, generative AI is an exciting prospect, but users should “be aware that it has limitations and can create risks,” he told HR Brew via email.—SB

Quick-to-read HR news & insights

From recruiting and retention to company culture and the latest in HR tech, HR Brew delivers up-to-date industry news and tips to help HR pros stay nimble in today’s fast-changing business environment.