How to Stop Worrying and Start Using AI at Work

As someone who spends most of their time thinking about the stock market, I didn’t expect AI to take over so many of my conversations, but here we are. Whether it’s smarter chatbots, tools that automate creative work, or Silicon Valley companies pushing the boundaries of what’s possible, AI is everywhere. And yet, a lot of people still haven’t found a practical way to plug it into their own lives. These breakthroughs aren’t just tech demos; they’re real tools designed to help people like you and me, to help make our lives easier, our work faster, and maybe even a little more fun. But despite all these benefits, one thing I hear often is fear: fear that AI is coming for people’s jobs. And to be fair, some roles will be replaced. But the bigger picture is this: AI is creating more opportunities than it’s eliminating. For most people, it won’t take your job, it’ll help you do it better, faster, and with less friction.

If you're into chess or data science like I am, you’ve probably heard that classic line from former world champion Garry Kasparov: “No human is better than a machine, and no machine is better than a human with a machine.” He said this after famously losing to IBM’s Deep Blue in the late '90s, which was a turning point that made him rethink the relationship between humans and machines. That quote hits home and it’s a reminder that the smartest move isn’t to fight the tech, it’s to team up with it. 

At Golden Hello, a regulated wealth management and technology firm focused on the retail investor, we use more than just AI tools for finance, we also rely on tools that help improve key processes, enabling us to scale at lower cost, increase efficiency, and reduce risks and human error.  I’d like to share a handful of tools you should keep in mind to help eliminate time-consuming or repetitive processes, but before that I would like to remind you that with any technology, it’s important to ensure that the tools you use comply with relevant laws, industry standards, and data privacy regulations in your region.

First, let’s talk about custom GPTs, short for Generative Pre-trained Transformers. You’ve probably heard of those AI chatbots that can answer questions, draft emails, or explain complex topics in plain English. But here’s what most people don’t realize: many of these tools now let you build your own version, customized to your exact needs. For example, you could create one that lives within your internal tools, trained to answer team questions about company policies, or workflows, saving hours of back-and-forth and keeping everyone on the same page. You can also set up a personal assistant that preps your weekly internal reports or generates slide content, exactly how you like it. Tell it once and it will remember, no need to repeat yourself. It’s kind of like giving superpowers to a very smart, very reliable intern who never forgets a thing and is available 24/7. 

Now, if you sprinkle in a bit of code and the model has a well-documented API, you can unlock some truly powerful possibilities here. A few years ago, I built a script for personal use that pulled headlines from a company’s daily news and used an NLP library to analyze the market sentiment around it. It would break down each headline, score the words, and determine if the tone that day was positive or negative for a particular stock. Fast forward to today, and these models don’t just stop at headlines, they can process entire articles, annual reports, or even dense documents. You can use them to quickly scan and summarize pages of financial jargon, pulling out the key takeaways in seconds. What used to take hours of reading or some pretty scrappy automation can now happen almost instantly and with far better results.

Some people take it a bit further with AI Agents, which are pretty similar to custom GPTs in that they are meant to serve specific tasks but these are less supervised; however, these are a little more complex so I would not place them in the “easy” to use section, but if you really put work into it you can create amazing automated entities that can be reactive as opposed to having to act based on a prompt or a command. You can think of these as “proactive custom GPTs” which take the initiative, make decisions, and complete tasks autonomously as opposed to responding to a prompt. That said, not all AI Agents require you to build them from scratch. There are already powerful agents out there doing real work, like meeting assistants that automatically join your calls, take detailed notes, generate summaries, extract action items, and send follow-ups. These are AI Agents in action: operating independently, reacting to what’s happening, and making your life a whole lot easier – and there is no prompt required.

Some people are taking AI Agents even further by building what’s known as “Embodied AI Agents,” or AIs with a physical presence. When you connect an agent to a robot, drone, or any kind of machinery, it becomes more than just code on a screen. It can move, see, touch, and interact with the physical world. These embodied agents can pick up objects, navigate spaces, and respond to real-world environments in real time, opening the door to some truly sci-fi-level possibilities.

Another way to leverage AI is through simple machine learning models, especially if your company already has clean, table-based customer data. These are often predictive models that are like smart guessers: they look at patterns in past data to make informed predictions about what’s likely to happen next. But machine learning isn’t just about forecasting; it can also be used to classify information, detect anomalies, recommend products, or even personalize user experiences, all by learning from the data you already have. Take the hospitality industry, for example. If you run a hotel, you could use non-personal data like age, gender, booking habits, and room preferences to predict how much a guest might spend on extras like spa treatments or room service. Or, if you run an app-based business, the same approach can help you anticipate subscriber churn before it happens. The real magic here is pattern recognition. These models can uncover trends and correlations that would be impossible for a human to spot manually, giving you a serious edge in decision-making.

Not every use case requires building a custom GPT or programming an agent from scratch. In fact, in many fields, especially content creation and multimedia, the tools are already built and ready to go. If you work in marketing, advertising, or any kind of creative field, there’s a growing ecosystem of AI-powered platforms that can streamline everything from writing copy to editing videos. You don’t need to code, just plug in your ideas and let the tools handle the heavy lifting.

Now, let’s talk about a few challenges, starting with one I’ve personally run into (and heard from others as well): math! Most of the LLMs we commonly use, especially the free versions, aren’t always connected to a calculator or equipped with one by default. I’m not entirely sure why (probably something to do with cost, speed, or design choices), but the key takeaway is this: Before you ask an AI to crunch numbers, make sure it has access to a built-in calculator or a tool for actual computation. Otherwise, it might look confident, but the answer can be way off. These LLM models are great at explaining how to solve a problem, but not so great at doing the math themselves because they don’t actually “calculate.” They predict the next word in a sequence based on patterns in text. And while that’s powerful for language, it’s not ideal for structured, step-by-step logic like math requires.

Another area where there’s still room for improvement, not from personal experience but rather conversations with my lawyer friends, is, of course, law. While current AI tools can be useful for basic legal tasks, like summarizing documents or drafting templates, they often struggle with more complex or jurisdiction-specific materials. Legal language is nuanced, and rules can vary dramatically depending on the region, so one-size-fits-all models just don’t cut it (yet). That said, if you work in the legal space, you can train a custom model tailored to your specific practice area or jurisdiction. It takes some setup, but the payoff can be huge in terms of speed and consistency.

Ultimately, my idea is that AI isn’t perfect, but it’s improving at an impressive pace, and staying curious now puts you miles ahead. This isn’t about chasing the next shiny tool or following the tech hype, it’s about recognizing a moment, one where the way we work, create, and think is fundamentally shifting. And in that moment, the smartest move isn’t to wait, it’s to engage! 

You don’t need to be a coder or an engineer. In a lot of cases, all it takes is curiosity, a bit of common sense, and the willingness to explore. And yes, it might feel overwhelming at times. But think back to the early days of the internet (if you’re old enough), when logging on felt foreign, building websites was reserved for a select few, and businesses weren’t even sure if they needed an online presence. Fast forward to today, and we can’t imagine life without it.

AI is shaping up to be just as transformative – many call it the core driver of the Fourth Industrial Revolution. And like every major shift before it, those who adapt early will have the biggest advantage. At its core, AI isn’t about replacing people, it’s about empowering them – empowering us. Of course, this, like the rest of the article, reflects my personal opinion and perspective. It should not be taken as professional advice and is best evaluated in the context of your own needs and circumstances.