Only two months after its launch, ChatGPT became the fastest-growing consumer application ever with a 100 million monthly active users. Such widespread user adoption means AI isn't a flash in the pan, and many companies are now wondering how they can use AI to their benefit.
X-Team has several software engineers who specialize in AI. Rudolfs Heimanis is one of them. He works for an X-Team partner that's a large media organization, and his job is to help the organization implement AI in a way that makes sense for them.
In this interview, we discuss Rudolfs' involvement with AI, what benefits it provides, what risks there are, and how any company can get started with the technology:
Hello Rudolfs. You're an AI expert for X-Team, but AI has only recently gone mainstream. How long have you been involved with the technology and why did you decide to focus on it?
I’ve always been excited about the world of AI, but initially the complexity scared me away. I always thought that you had to know the mathematics behind it to really become an expert.
I was wrong.
When I experimented with the general purpose offerings from OpenAI, I couldn't believe how easy they were to use. ChatGPT showed me that you don't need to know complex math. There are so many layers of abstraction above the math that, sometimes, you only need ten lines of code to load a basic open-source model that solves your specific problem.
ChatGPT quickly became a valuable source of information for me. I learn by doing, not by watching. The interactive back-and-forth with ChatGPT allowed me to learn and experiment at my own pace. It's an incredible tool to learn and get better at your job.
Many companies are interested in AI, but don’t know how to use it effectively yet. How can companies use AI to become more profitable, efficient, cut costs, et cetera?
I’m trying to balance being pragmatic with being innovative, so whatever I write here is my subjective opinion. There are several ways companies could benefit from AI. Here are just a few:
Assistance. Anyone whose job consists of repetitive tasks can benefit from AI. For example, salespeople who currently manually research prospects could make good use of an AI tool that provides them with all available information about a prospect, as well as what features of their product could benefit the prospect. It would significantly speed up their work.
Automation. Verifying new customers in a Know Your Customer (KYC) process is a time-consuming process. Someone has to manually evaluate a customer's information to verify their identity and determine if they pose a risk for the company. A machine learning (ML) model can learn from real employees as they verify new customers to provide some sort of customer trust rating. This will help speed up the KYC process for a company while also showing them positive or negative trends for new customers.
Content creation. I don't believe that content should be entirely generated by large language models (LLMs). But there's definitely a role for LLMs in content creation, whether that's for ideation, the first draft, to find better ways to phrase things, or to better target a specific audience. So AI has its uses here too.
These sounds great, but can you give me a specific example of how AI can impact a company?
Let me give you a story. John Doe is an employee of Chad Sigma Beta Ltd (CSB). He hasn't been doing well. He's been feeling overwhelmed in his personal life and it's affecting his job performance. So he visits a personal therapy chatbot on his company's intranet. It tells him he should take some days off. In fact, the bot tells him he still has ten PTO days. It also provides him with a link to submit a PTO request.
When John submits the request, it gets approved immediately.
How? Behind the scenes, an AI solution responsible for HR flows at CSB automatically checked if John's colleague was going to take any days off during that time. She wasn't. Given her good performance of late, she was also likely able to shoulder John's work for a short while.
So the AI automatically set up a meeting between John and his colleague for a temporary project handover, and it approved John's PTO while also notifying HR and management. John's team leader then sent him an email wishing him a nice vacation while also asking his colleague if she's okay with the extra temporary workload.
All this is possible today. AI can dramatically improve workflows at any company.
And here's what all the examples I listed have in common: They don't put people out of their jobs. You can't ever rely entirely on an AI solution. That human touch is still important. AI shouldn't be thought of as something that will eliminate jobs.
Instead, it's a tool to enhance people's existing capabilities, to make them more productive, and to have them spend less time on tedious and repetitive tasks. It allows companies to deliver more value without spending more money.
What about the people who believe AI may not have a large impact on jobs right now, but think that it will as AI becomes better and better?
Think of it this way: Every generative AI solution on the market has been trained on content that humans have written. If you replace man-made content with automatically generated AI content, you'll not only have factual inaccuracies, but also a really limited worldview.
When civilization evolves, so do our problems. We can't rely on AI exclusively because we'd be solving new problems with old information. So human content is still a mandatory precondition for the growth of AI.
Many science-fiction stories are about robots controlling our cities and being our overlords, but even the best AI won't ever be able to outsmart a human. AI may be able to think faster, but that doesn't mean it'll think better.
So there will always be room for people, especially when it comes to crucial decisions that involve morality or that require solving really difficult problems. We can't outsource that to AI. We shouldn't either, because that's a scenario straight out of Terminator 😄.
Also don't underestimate how many jobs AI will create. You may not be a woodcutter anymore, but you'll manage a smart machine that does much of the manual labor for you. The job doesn't go away, it just changes.
Still, that doesn't mean you should rest on our laurels. You still have to focus on educating yourself and learning new skills. If you don't make yourself irreplaceable, you're at risk of losing your job regardless of AI.
The main point is: Prepare for what may come, but don't live in fear of a future that's based on an unproven hypothesis.
That's the right attitude. Let's now talk about the risks for companies. What do companies have to be aware of when they’re working with AI? Are there scenarios where the risks outweigh the benefits?
I’ve already mentioned that companies should avoid relying completely on AI. No AI tool is perfect and it's unlikely one will ever be.
Still, companies should carefully review the privacy policies of their chosen third-party AI providers. Some providers will use company information to train their models. Some will retain data for a certain period of time for debugging purposes. All this is important to know if you want to comply with data privacy frameworks like GDPR.
As a general rule, if you're working with sensitive data, don't use the general purpose AI models that are available on the internet. Train your own model that's based on an existing and trustworthy open-source model. Or choose a provider that guarantees your data will remain in your control. For example, most solutions provided by Google's Vertex AI are safe to use.
Even still, make sure to double-check any claims a third-party provider makes. You can never be too careful. It's not that different from working with any third-party provider. There will always be some risk involved.
Still, there are so many great AI solutions available today that you'll always be able to find a replacement in case something goes wrong. Or even better, a solution tailored to your company or your specific problem.
You should also carefully choose the people you're working with for your AI solutions. Pick someone you can trust. Someone with proven expertise, who knows what they're doing, like X-Team 😉.
What would you say to companies who are reluctant to adopt AI?
It's normal to be cautious of new trends. My biggest advice is to test it before writing it off. AI used to be really expensive and only accessible to big corporations, but those days are long gone. It's become really affordable to experiment with AI and figure out if there's any value to it.
So you may as well try it out. A proof of concept costs almost nothing. Of course, don't except AI to suddenly solve your biggest business problems. It's no magic wand. Give it a little problem first. Get familiar with what it can do. Step by step, expand its use cases.
Every company can benefit from AI. If you want to get started, here are a few things you need to figure out. Of course, this is only one way of doing it:
Get really clear about what you're selling. For example, if you're a software company, you're not selling software. Perhaps you're selling experiences. Perhaps you're selling the time your software saves your clients. Perhaps you're selling the cost it cuts. Perhaps you're selling the money it makes. Understand the benefits your product or service offer your clients.
Talk to the employees of different departments about their workdays. Ask them how they currently do things and how they feel about the way those things are done. Provide them with a safe environment, so they open up and are genuine with their responses.
Talking to your employees, chances are that you've noticed a few things that could be done better. Perhaps some departments have many repetitive tasks. Perhaps some departments are slow because their requirements are unclear. Create some kind of diagram about a workflow that could be improved, like so:
Imagine a solution for the workflow you want to improve. For example, what if you trained an internal chatbot that answered the above question for you in one second? Like so:
That looks like an improvement, but you'll only be sure of that if you benchmark your current processes. How much time does it really take your employees to do this? How often do they do this? You don't need perfect numbers, but you need some sort of baseline to compare your solution to.
The natural next step is to find a company that has experience solving identical or similar problems with AI. Tell them about your problem and, together, define as small an MVP as possible. Minimize the scope, because you want to test the validity of your idea before you spend lots of money building out a full solution.
If I may, that's the benefit of partnering with X-Team. My colleagues and I always start with a pilot project. We ask ourselves, what's the most amount of value we can provide our partner for the least amount of work? What's the best and most cost-effective solution to solve the partner's problem? That's where we start.
Once the pilot project is up and running, it's time to validate your idea. How does the solution compare against the benchmark? How much faster is it? Is your team satisfied with the solution? Was it worth the expense?
If the pilot project had a positive impact on your company, it's time to build. Scale the solution, add useful functionality, optimize it. Make sure the solution grows with your business metrics and your goals.
Modern software engineering is iterative. There's no need to assign a giant budget to the development of one big thing. Watch your solution grow week by week, month by month, year by year, decade by decade, okay I'm going to stop, you get the point. There's tremendous value in AI and I encourage you to at least experiment with it.
If you enjoyed this article and are interested in how your company could benefit from AI, don't hesitate to contact X-Team. We have several software engineers like Rudolfs who are deeply experienced with AI, and who would love to build awesome software to help you reach your business goals.