The State of Work with Jimmy Lowery Jr. (Mid-Atlantic)

0:09 — Jimmy: Hey, I'm Jimmy Lowery Jr., and I am the Founder and CEO of Simplora. We provide a unified meeting workspace for modern teams.

0:20 — Anthony: All right. Well, Jimmy, we've talked a lot about artificial intelligence, but I do want to delineate, because I think we often conflate these terms between automation and AI. Oftentimes AI can help you set up some automation, but the value you're really getting is automation, not AI. You don't need to think about repetitive tasks — you just need them done for you. In the TriNet State of Work Report, one of the things we found in your region, the Mid-Atlantic, is that employees will rarely use AI for something they haven't done before, but they're okay with having AI applied to things they have to do all the time — but really, that's just automation. In your company, employee-wise, do you guys even delineate? Do you just allow them to conflate because it doesn't really matter to get into the semantics? Or are you guys die-hards about vocabulary and you're going to correct someone if they say it wrong?

1:12 — Jimmy: Yes. I do think it's important to delineate between automation and AI. The main reason being that AI is a specific set of functions that has to happen using some sort of model. That model is usually trained on millions or billions of data points, and that AI is generally used for generative purposes — actually generating a specific outcome. So if you are automating something, you are essentially speeding up the process to get from point A to point B, which you could effectively say, "Okay, AI is helping us do that to some extent." If I want it to generate a design, well, AI is generating that design, and it's kind of skipping from point A all the way to point B without the need for any designer. But automation would be having some process in place that collapses the workflow.

For Simplora, the way we think about it is: we have generative AI, and then we have agentic AI. Generative AI is using those large language models to help us produce certain outcomes. Sometimes that's a matter of explaining a complex concept while you're on a meeting. Sometimes that's the matter of retrieving a definition to a specific word that was just mentioned. But then there's agentic intelligence, which is, "Hey, I actually need to retrieve a specific document from our connected applications to then surface that in real time." So you're actually allowing the agent to act autonomously, to make certain actions or take certain actions that otherwise wouldn't have been possible by just automating it per se.

When it comes to automation, what we are doing is we have pre-meeting preparation, which means before every meeting we equip our users with advanced insights about who they're meeting with, context about previous meetings they were in, any sort of action items that are still pending, any connected data from their systems that's going to help with that upcoming meeting. We actually use automation and AI within those briefs so that we can effectively automate certain processes that the individual might have needed to take — such as going to LinkedIn and looking at that individual's information on their LinkedIn. We've created the pipes. We're using APIs and automated workflows to aggregate some of that data. But then once we get that data, we now apply a layer of AI to help analyze that data and parse out insights that could be valuable as you're going into that meeting. So it's a matter of combining automation and AI, but I think it's important to delineate between the two so that you can think about them in that way.

5:06 — Anthony: The TriNet State of Work Report shows that the Mid-Atlantic data shows 36% of employees say they work too many hours. We believe AI is going to help make them more efficient, which we believe would make them work less hours. But ironically, is this efficiency going to make them feel compelled to work more because they can do more with all these tools they have? Are we making burnout worse?

5:34 — Jimmy: Great question. I am actually published on the topic of burnout. So to address the initial question and then to discuss burnout in correlation: if you are using AI, using automations to save time, the hope is that the time being saved is then being utilized for higher-leverage tasks or projects or work that needs to be done. That's the ideal outcome. Let's automate the busy work, or let's use AI to manage the busy work or the work that is capable of being automated, and let's use the manual time, personal time, to focus on higher-leverage strategic work. If we can do that, great. But a lot of times what you may end up seeing is almost a sort of detour towards other types of work that aren't as high of leverage — or the lack of work altogether, because that sort of busy work you're used to doing is now just being done. So why even do anything now at all when I can just scroll on LinkedIn or something like this, right? The hope is that it automates the busy work so that you can focus on higher-leverage work, but sometimes that's not always the outcome.

Regarding burnout, what usually happens — it's not necessarily a matter of working too many hours. That is more or less a factor that plays into what can cause burnout, but burnout typically actually starts more at an identity level. It's how one equates themselves to the work they do, how they view themselves within the work they do. Once your identity becomes intertwined with the work you do and the outcomes you produce, the expectation you place on yourself is much, much higher — meaning that you rise and fall with the wins and the losses in your workplace. So if you are working so many hours, it's not usually a question of, "Oh, you have that much work to do." It's usually a question of, "Oh, I'm using this work as a bridge to become someone who is accomplished, or someone who is recognized, or someone who is great in this field." So it ties back to identity economics — what is this work helping me to achieve or to do? If you're hopefully saving some time by automating some work or using AI to take care of work, you'd like to think you can then use that time in a more strategic way. But whether or not that's going to reduce burnout is unlikely if your identity is too closely intertwined with the work you're doing.

9:06 — Anthony: Let's go through it. 50-person company. What is your advice? You're the consultant. I'm hiring you. I'm the CEO. Jimmy, what do I do here? Do I get voice bots? Do I get some chatbots? Do I hire someone? Do I have my own employees do it and just give them licenses and let them go free? Or do we need to hire an outside dude to come teach us all stuff? What's your advice for me?

9:25 — Jimmy: At a 50-person company, I'm likely going to recommend to just get started. What that's usually going to look like is to probably get some sort of subscription to something like ChatGPT or Claude — something that you can equip your team or your employees with to be able to start interacting with AI in a safe and guarded way. The real challenge that happens with using AI in the workplace, or at least the pushback that generally exists, is one of privacy. If you purchase the right plans, or you're working within the enterprise structure at, say, OpenAI or Claude, whoever — you're going to have the privacy built into it. So then the question just becomes, "Hey, how can we start giving our employees these tools so that they can start to use them?" The employees are then going to figure out how to best use those tools. That's something I would say: just get started. Buy a tool, give it to your employees, allow your employees to use it. That tool may look different depending on the type of business you're in, but I think a large language model — one of these general models from one of these established companies — is going to be the easiest way to get started.

However, if you're at, let's say, a 500-person company, a lot more logistics, a lot more moving parts, bigger departments, more teams, more employees, of course. You're going to want to take more of a strategic approach. I would still say you need to basically think through, "Okay, what are some of the areas in your business that would benefit from having some version of generative AI?" That's going to include areas like email. What about video calls? What about communication tools that we're using? How can we start using AI to improve those workflows or make the use of those more efficient or more effective? Almost kind of thinking through some of these gaps. There's also obviously conversations that would need to be had per department. So we'd look at engineering and say, "Okay, how can we make the engineers' jobs easier and make them more effective?" Maybe that means adopting an AI-native IDE, something like Cursor. Maybe that means equipping them with ChatGPT or OpenAI's solution Codex. Maybe it's something akin to that. How do we essentially give our engineering team a solution to help them? Then we're going to look at marketing — what are some AI-based marketing solutions? Breaking it down department by department, figuring out where you can introduce different tools or different solutions.

12:46 — Anthony: All right, 5,000-person company. Now I'm international, and I'm publicly traded. So I have legal, I have investor relations, I have brand safety, and most importantly, I have stockholders to answer to. How, if any, does your advice change?

13:02 — Jimmy: It's a multiplier on top of what the 500-person company is doing, effectively, in the sense of the strategy. It does need to be somewhat top-down at that point. You want to encourage bottom-up adoption, but it has to be within the confines of what is going to be safe and secure and ideally scalable. You don't want to make any rash decisions or have it be like the Wild Wild West. So you're going to want to say, "Hey, as a company, we are adopting this AI solution for this purpose." If you need AI for that specific purpose, let's say it's something like meetings, this is what we use for that. If there are other solutions that you would prefer to use, you'd have to submit those through IT for approval, etc., etc. But still wanting to give your team and your employees something that they can use. It's just a matter of how top-down is that approach going to be, or how bottom-up is it going to be. For a 50-person team, the bottom-up approach can work. For a 5,000-person team, the top-down approach is likely going to be more useful there.

14:34 — Anthony: Excellent. Well, thank you for joining us for TriNet and the State of Work Report. You can learn more if you visit trinet.com, and we encourage you to dive into some details and give it a quick read-through. Thank you for joining us today, Jimmy. Have a great day.