Breaking My Resistance to AI
Preserving the Human in HR Ops
My pen-and-paper impulse is strong and enduring. My gut reaction to considering AI in HR/Ops is a blaring “NO!”. I may be more traditional than most, but when I think of HR I think of faces, conversations, solutions and my notebook. I don’t think of robots. My ignorance is showing here, as is my stubborn mentality that as one of the few not in production at Invoke, I don’t really need to be utilizing a ton of tech to do my job.
I’ve realized, however, that I am merely 33 years old, it’s 2023, and most of my peers aren’t relying on Google Sheets for everything. Thanks to two of my coworkers who put together the initial presentation on embracing AI for our recent annual offsite, I got schooled and heard how it could assist me in my daily work.
When we talked about tools that assist in recruiting and external-facing communications, I voiced why I’ve resisted using many tech/AI tools up until now: I want to be in control of what info is out there attributed to my name.
The response of the team was empathetic (because they’re humans, not robots), and they encouraged me to think about all that I control, what I like, and in contrast, which are the most major pain points in my job (a blend of HR and Operations). If I can identify the most painful and tedious parts, I can start to think about how I can make my life easier while also feeling like I’m still a human doing my job.
Can AI Actually Help?
If I had to choose my biggest pain point — and this likely contributes to my tech resistance — I’d say it’s the time and effort required to sync data between the software we use. For example, our accounting software does not properly sync with our banking portal, so I need to spend time tapping my keyboard and clicking my mouse to make them both happy.
But, to the point of this blog: I was resolved to try out a new tool. My colleagues recommended Sintra, a resource for ChatGPT prompts. It’s supposed to help scale the learning curve by taking very human-centric, basic language, and transforming it into a prompt that will ultimately get a better result from ChatGPT. You may be picturing me using a rock and chisel to record our confidential information, but I have used ChatGPT before as a way to ensure that I’m not just relying on my memory for things that matter and could have legal consequences for our employees.
The goal in my head was clear: Use Sintra to find an AI prompt to assist in ensuring our HR policies are comprehensive of current legislation, best practices, and verbiage that covers us in context-dependent situations. This could work for a swathe of policies, but I figured I’d test it out on a traditional and well-known process first — just to make sure I can trust the robot.
My search in Sintra: Sick Leave
Instantly, Sintra came up with a thorough recommendation based on my sparse input. It gave me information on what I can use (no surprise, ChatGPT), and why ChatGPT is qualified to help draft and create forms within legal compliance.
Sintra’s results: ChatGPT can assist you in managing leaves of absence policy by providing you with accurate and up-to-date information on company policies, legal requirements, and best practices. ChatGPT can help you draft policy documents, create leave request forms, and answer common employee questions related to leaves of absence. ChatGPT can also suggest ways to handle complex situations and help you ensure compliance with legal requirements.
I appreciated the placeholder text where I could modify the copy based on nuances. This is often the tedious part of policy updating: Finding where the nuances or one-off situations fit into the stricter context of the standard policy. It’s nice that Sintra recognizes its limitations and accommodates them accordingly.
I also like how Sintra is basically an intermediate training tool that helps me understand how to interface with something that will end up maybe mitigating if not entirely eliminating, the tedious parts of my job. In a way, it’s a little ironic having a robot tell me how to speak to another robot. It’s like an HR reverse uno card.
This structure expedites the learning curve and helps build more trust to reinforce that Sintra understands what I’m going for and how to optimize my interaction with ChatGPT.
My next action: Open ChatGPT, paste the prompt and input my variable information.
ChatGPT’s response: Quite comprehensive and with 9 long bullet points, but prefaced with this:
I’ll admit that this opening statement by ChatGPT stopped me from reading the rest of the answer. It immediately made me feel uninterested in the results below. If this all-seeing robot admits it may not have the most updated information, how can I feel confident in the following response?
Now, I realize this isn’t Sintra’s fault. So, in the spirit of being reasonable (like a human), I appreciate Sintra’s original prompt, even if I don’t love ChatGPT’s result. I like that using Sintra gives me a detailed prompt that should give me the most comprehensive answer. When writing policies, there’s nothing worse than publishing it to your employees and realizing a day later that you should’ve included another section.
Looking Forward
I’m open to using Sintra. It takes time to train the tool to recognize what I want, but given the first attempt being pretty comprehensive from only two words (Sick Leave), the impact was clear when ChatGPT gave me a super detailed bullet point list. The only barrier now is the paywall — Sintra has one free prompt with more behind a paywall.
For now, I’ll stick to the free version and widen my AI horizons slowly. I have Sintra-ready prompts to update our HR policies, and I’ll continue to test prompts and responses to compare with what I know, as I build my trust and break down my resistance to AI.
–
Next in our AI Powering Your Product Design Series
Our 6-part blog series casts a look at the year of generative AI, through Invoke’ team members’ stories — their insights into how they’ve found using the suggested tools and how they’ve been introducing AI in their workflows.
Dec 19 — AI Text-Generation Tools, Today and Tomorrow
Our previous posts in this series:
- Optimizing Project Management with AI Tools by Terence Chan
- AI Powering Your Product Team by Martin Ambrozic
- AI-Powered Product Design: The Pros and Cons by Carmen Perry
- Redefining Software Development Efficiency by Luke Date
