AI is having its moment. After years of working behind the scenes, tech titans from Microsoft to Salesforce are rolling out new above-the-hood machine learning tools — such as ChatGPT-like generative AI — amid what Salesforce co-CEO Marc Benioff just called “one of the most profound technological shifts” the world has experienced. And it’s just the beginning.
We spoke with Rodrigo Liang, CEO of Silicon Valley-based enterprise AI platform company SambaNova Systems, about the present and future of AI in the workplace. His view? AI is so big that within the next couple of years, we can expect a tech revolution as transformative as the advent of the Internet — and it may happen even faster.
How would you describe the current state of AI in the workplace? AI exists in all our lives already. It’s baked into our phones, our day-to-day computer applications like our email spam filters, online malware detectors and so on. It’s not quite as visible as some of the more recent developments like generative AI, but I would say that likely 100% of companies are already using AI technology under the hood, it’s just that many of them don’t know it.
Now, if you look at how many businesses are really thinking more explicitly about introducing AI into their company — saying, for example, “I want to bring in a generative model to replace workflow A to workflow B”— I would say north of 50%, maybe 70% of all midsized to large businesses are thinking about this. That’s because they saw what happened with the Internet. Companies that said “oh, this is nothing” like the Circuit Cities of the world, they’re gone. So, many people realize that with this transformative technology coming, if you don’t have a plan, you may be gone. But in terms of which companies have gone beyond just thinking about it and have actually started deploying the technology, there are very few.
What are some examples of “above the hood” AI that is starting to be implemented in the workplace? ChatGPT is one, which allows you to generate incredibly well constructed sentences and paragraphs with just a few prompts. It’s a great example of what AI can do, and yet it’s still a very small example of its potential.
Beyond that, we’re starting to see people say “I want that [AI capability], but for my company.” Mostly very straightforward services, like something that can interpret and manage infinite Slack channels, for example. Or they’re looking for chat and Q&A capabilities, something that allows them to generate a response that’s coherent. So those types of things are being deployed very quickly, and I think people are finding it very valuable.
Another example is image recognition. Our phones already use it for facial recognition, but we’re starting to see in different workplaces that you don’t need a badge anymore, because a machine scans your face and lets you in.
What advice would you give to businesses looking to implement more AI technology? Have a strategy. Find a partner who you can trust and who understands the technology broadly and can help guide you on the journey of transforming your business. It will be a systematic transition, as every aspect of your business and every minute of your day becomes AI-enabled to significantly increase productivity. Similar to the internet transformation, but perhaps a little faster.
Since the technology is changing so quickly, I recommend hiring a team or service that can buffer you from having to constantly adapt. It’s not so different from the models we already use, like phone or cable or Internet service. The average person doesn’t really know the technology, they just sign up for a service. Then the experts providing the service update the software and hardware and suddenly it’s just better.
What potential pitfalls should businesses be aware of when integrating AI? If you look at ChatGPT and what it’s done in the consumer and enterprise space, you can see that the required level of accuracy is significantly higher. You can’t be wrong, and you have to be able to point to how you arrived at whatever conclusion. So, if your results aren’t referenceable, if you can’t reference exactly why you’re getting certain answers, the technology becomes much harder to use in an enterprise environment. Shipping times, for example. If you’re shipping a product and it says it’ll take eight days, you’ll want to know how you arrived at that timeline.
What do you think is the future of AI in businesses? How will it affect employment? We’re seeing the fastest industrial revolution the world’s ever seen, and we’re going to see this technology become pervasive in every industry, in every function of our every day. It’s just a matter of figuring out how to get started and how to have a plan that allows businesses to, frankly, survive this transition in the next three to five years.
We’re still in the very, very nascent stages of AI being used in the workplace. But at some point in the next 12 to 24 months, every part of your work day — checking emails, responding to emails, creating content, going to meetings, summarizing meetings, etc. — will all be AI. It’s just going to suddenly become standard. In the same way that 20 years ago, none of your workday activities involved the Internet, and now they all do. Zoom calls will automatically generate action items and a meeting summarization. You’ll see entire webpages being created with just a few words. That technology already exists today, it’s just still the early days of deploying it in the workplace.
