Deon Nicholas is the Co-Founder & CEO of Forethought. Previously, Deon built products and infrastructure at Facebook, Palantir, Dropbox, and Pure Storage. He has ML publications and infrastructure patents, was a World Finalist on the ACM International Collegiate Programming Contest, and was named to Forbes 30 under 30. Originally from Canada, Deon enjoys spending time along with his wife and son, playing basketball, and reading as many books as he can get his hands on.

Forethought, the AI and Machine Learning platform for the enterprise, began with its deal with customer support. The corporate’s AI can learn from internal documents, email, chat and even old support tickets to mechanically resolve auto-route tickets accurately, and quickly surface essentially the most relevant institutional knowledge.

We sat down for our interview with Deon on the annual 2023 Upper Sure conference on AI that’s held in Edmonton, AB and hosted by Amii (Alberta Machine Intelligence Institute).

That is our second interview with Deon, in our first interview we focused on his past and the way he got his start in AI, on this interview we focused on his vision for the longer term.

Could you summarize what Forethought is?

At Forethought, we’re the generative AI for customer support company. We launched in 2018 at TechCrunch Disrupt, and since then we have grown to powering support for large-scale corporations like Instacart and others. And mainly, we offer an intelligence layer on top of their customer support ticket data, which then translates into things like chatbots, but hopefully smarter, right through to agent assistance, all leveraging nowadays, generative AI and things like that.

How way back did you make the switch to generative AI?

What’s interesting is even after we were founded, we now have been leveraging some type of generative AI for the reason that early days. Nothing I’d argue, as powerful as what’s available today. But for instance, GPT two was launched, I imagine in 2018, 2019, and open source, and there have been other models like T5. So, we had been leveraging various large language models, tinkering with those in addition to a few of our own internally that we have been constructing. But what’s modified, I believe, is that was a feature that gave you some benefits if you use sparingly. And now I believe during the last six months, what’s really modified is that it’s develop into this seminal shift in how business models are happening and the way the engines are happening. And for us, we actually needed to rethink our engine, I’d say, in the previous few months. And we launched SupportGPT in March of this 12 months, 2023. And that was leveraging large language models like OpenAI’s GPT, and really rethinking how we do the entire thing. And it isn’t actually necessarily a brand new product, but a brand new engine that powers all our products, which has led to a ton of improvements across the stack.

What’s the method for an organization that wishes to start using SupportGPT?

At the tip of the day, it’s embedded into one among our products. So what I’ll say is regardless of what, it all the time starts along with your data, and that is our differentiator. And the unique thing about Forethought with most corporations, you begin with a blank canvas, and you may have to hard-code these rules. For us, you begin by integrating. So, when you’re leveraging a well-liked help desk or a CRM like a Zendesk, a Salesforce service cloud, you would like to work with us, you enroll after which install our integration into your help center. And that kicks off indexing, that kicks off training, fine-tuning, and all those things, and builds the model and builds the engine. After which from there, you may configure, you may edit, and you may deploy one among our products. And our hottest product is Solve, which is an AI agent that may sit on a web site, almost take into consideration a chatbot, or can exist in email in really any form and begin mechanically conversing and responding to your customers, leveraging the SupportGPT engine like ChatGPT in your website, so to talk.

But then that automation only handles 50% of issues, so to talk. And so, what in regards to the issues that also must go to a human? Well, we also Triage, which is the name of our second product, route issues, tag them, be certain they get to the fitting agent in the fitting channel, the fitting time, so you may deploy that. After which Assist is an agent co-pilot, so internal facing GPT for the client support agents, after which ultimately, Discover, which is our most up-to-date, but in some ways, strongest product, that is in search of insights and making recommendations, also using generative AI to the business on what must be updated and what must be modified.

With a lot reliance on generative AI, is hallucinations a problem in any respect?

Yeah. I believe hallucination is one among the massive problems leveraging generative AI for many practical use cases, and I recorded a video on LinkedIn about this, which went just a little bit viral, on hallucination being one among the massive problems with generative AI. There are numerous cases where it isn’t a problem, like if you may have a human within the loop or it’s a really just creative use case, and you would like something recent like marketing, you are coming up with ad copy blogs and you are going to have any person edit at the tip of the day. That is really good that they are hallucinating just a little bit. It’s the shape of creativity. But in cases like finance or healthcare, or customer support where your health, your wealth, your liveliness, or something, or you would like an accurate answer, hallucination is big. It’s an enormous problem, and it’s an enormous limiter. So, in some ways, one among the cool things we realized and why we are able to do that and no one else can, was because we have been leveraging some type of generative AI for the past, call it five years.

All of our models have been focused on correctness from the get-go, understanding the policies, understanding the workflows that when any person asks for a refund, if it’s inside 30 days, you may issue the refund. If it isn’t, you may’t. And all of that stuff that we might already built. And when it got here time to leverage these more modern large language models, we found that the humanization of the big language models, plus the actual information and correctness that we could provide from the fourth-up models was an ideal match. And you then can leverage all of that correctness through prompt engineering, through fine-tuning, which actually brings the hallucination problem down. It isn’t eliminated completely, but it surely gets minimized to the purpose where it’s extremely effective. And our eventual goal, obviously, is, have less hallucinations than humans would have errors. So long as your accuracy rate is on par with human accuracy, you then’re in a extremely strong spot.

We last interviewed you in 2021. What have you ever learned since about being an entrepreneur?

Oh, my goodness. The entire world has modified. From an entrepreneurship perspective, just a few things have happened. In late 2021, shortly after our interview, we raised our series C. We raised 65 million series C from Steadfast Capital, NEA, in addition to luminaries like Gwyneth Paltrow, Baron Davis, and Robert Downey Jr. In some ways, we saw this ton of pleasure around our vision, and I suppose these were individuals who saw generative AI before. It was cool in some ways. That was an exciting and crazy time for us to only grow and pour fuel on the fireplace of something that was working. But what was interesting is that shortly after, mid-2022, so six months later, recession. The entire world blows up, so to talk, right? And in order that taught me rather a lot because where some businesses were spending, they cut spending, and the entire business model of even being an entrepreneur modified from grow in any respect costs.

Hey, you may have this giant war chest of capital, burn it and grow to get to your next round. To attend, suddenly everyone, including our VC, everybody’s VCs, and board stars, get up and said, well, hey, money’s not free, and it is advisable to be constructing either a profitable business, or it doesn’t necessarily must be profitable, but it surely must be efficient. And it isn’t about growing in any respect costs, but it surely’s about growing at reasonable efficiency. And you continue to obviously need to grow maximum given the reasonable efficiency, however the model had almost been flip turned overnight because there is not any guarantee you are going to have access to capital for one more one, two years. If we slip right into a recession, possibly three. Buyers are tightening up their purse string, so to talk. And so, you actually need to deal with efficiency and in addition driving efficiency in your customers.

In some ways, that forced us to get rather a lot more focused on, I mean, we were all the time delivering value in some ways since it’s just the industry we’re in. If we are able to show you how to help your agents produce more, solve more questions in a given time, you are going to avoid wasting a variety of costs and you are going to have a greater customer experience, and you are going to have more attention. All this stuff need to occur, but we now have to get tighter on our pitches, tighter on our messaging, and on our product focus with a view to share that. After which internally, ourselves, we needed to deal with efficiency. It wasn’t just, hey, you may have this big war chest, and you may all the time raise extra money, and yada, yada, yada. No, it was like, let’s determine how we will construct a business. What if we are able to never fundraise again?

What if it is going to take us years to fundraise before the VC market opens up? Well, that is effective. We’ve got to be constructing a really efficient business in order that when it comes time for a series D, we now have all of the metrics that success looks like, at the least on this recent world, by way of what’s being measured. I believe that was big, and layoffs and every part, and just the entire world shifted. And I believe it has been a really tough 12 months. 2022 was a troublesome 12 months, after which 2023, generative AI is hot again. And so, there’s ups and downs on this.

So, anyway, I’m talking rather a lot there. But yeah, I believe all of that teaches you to be focused and to be resilient. That is the opposite thing that is essential is just, it is a 10-year journey, when you’re successful. For those who’re not, it’s early, and so do not forget that through the ups and downs, you bought to take it in strides and that you simply’re all the time constructing towards that eventual vision.

What are the largest challenges that you simply face attempting to construct chatbots for customer support and other use cases?

Yeah, biggest challenges. I’ll go in chronological order. The primary was technical, do the models work? And in order that’s why we began even back in 2017, leveraging a variety of these modern, what we call, natural language understanding and natural language generation engines. And just ensuring that you simply’re all the time staying on the forefront of research, but then, at the identical time, ensuring you are delivering value. Because one among the things we realized is after we began asking our customers like, Hey, why? Within the early days, they told us that this problem is big. For those who can construct this for us, we’ll pay for it. The issue is chatbots have never worked.

Even within the early days, like 2017 for us, before we had officially launched, we’re like, oh, well then should we even be on this market? What is going on on? But you then dig deeper. You already know there is a problem, a clear-cut problem, however the solutions have not worked. After which once I kept poking and asking, I’d be like, why? Why not? After which I noticed that we, just, either by accident or due to our backgrounds, we were approaching it in a really different way. And though people had heard the word, AI, had heard the word chatbots, had heard this stuff, what we were doing was not what that they had been doing. That they had been approaching every part from a choice tree perspective. You hard code rules. If I see the word refund, go and issue refund, but when any person says the phrase, I just want my a reimbursement, it has no idea how one can handle it, since it didn’t say the magic keyword.

It was all decision trees, it was all rules, and these were just no code builders that made it conversational, but they weren’t AI. And I used to be like, oh, that is what you have been doing? And I didn’t even realize that is how all chatbots are built. And I used to be like, but that is not what we would like to do. We would like to do that. They usually’re like, really? You possibly can try this? And I believe the at the beginning thing was just recognizing there’s the nuance that though AI as a buzzword has existed for probably 10 years, this contemporary type of AI, what we’re now seeing is feasible with generative AI. And when you backtrack just a few years, what was beginning to develop into possible was not actually being applied to this space. And I believe that was big. After which there’s a variety of fall on effects from that.

There’s a variety of noise available in the market because we go and we’re like, guys, we’re doing something different. After which they’re like, Hey, are you sure? This other company said that they had AI too. And we needed to bang our heads against the wall rather a lot to only prove our worth. Once we got right into a POC, once we got into their system and showed the way it was working, how you may learn, then it was game over. And we start, as you may see from all of those logos and all the shoppers, but it surely type of felt like a slow slog to prove our worth every step of the way in which. I believe that is probably been the largest, interesting thing on this space, particular, one which’s been noisy, but with bad solutions.

It’s type of obvious when you consider it that call trees won’t work. It just wasn’t really obvious on the time.

It wasn’t really obvious on the time. And I believe it is also that was the very best you could possibly do, right?

That is true too.

People conflate conversational customer support with AI. And the 2, there is a large overlap of where it’s, like ChatGPT is a great conversational example, but GPT-3, the engine powering it existed before, and that was the AI. And also you needed that to make the conversational higher. Previously, what happened was we wanted conversational, we wanted AI, we could do conversational, however the only strategy to try this was to script the conversations. It’s an enormous IVR (Interactive voice response), or a phone tree, in chat form. But then people began slapping the word AI on, and so we were getting conversational, we weren’t getting AI. I believe that was where the breakdown happened, not that it wasn’t obvious, but just that it wasn’t even possible, or people didn’t even know it was possible, or that these are two various things.

You’ve described yourself as an AI optimist in a recent podcast. Why are you so optimistic about AI?

Well, just a few things. First, just broadly, I believe it’s a brand new business model or platform shift is going on. The identical way computing in and of itself became a thing. Computers going from giant mainframes to the private computing. The going from the web, the web being a platform shift, after which cloud going from desktop to cloud being that recent business model then to mobile, and all this stuff. Every decade or so, there’s this recent shift and these shifts bring about recent business models. And in truth, trillions of dollars price of latest businesses in of themselves, Salesforce, for instance, a 200 billion dollar juggernaut, and all they did was they recognized this shift to the cloud. They probably would claim they brought the shift to the cloud, but they recognized the shift to the cloud, they usually took databases, your system of record, and took it from on-premise, Oracle, Siebel, whatever it was to the cloud.

After which they made a killing off of that and constructing a system of record for salespeople, IPO, billion-dollar business. And, by the way in which, then they did it for service, service cloud, marketing cloud, boom, boom, boom, $200 billion business. And now I ask myself, what would’ve happened if that Salesforce were in-built our time? In 10 years later, whatever it’s, I do not know the way old they are actually, but in 2030. The identical thing goes to occur. And I imagine, which is why I’m working on this company, that it’s these AI first corporations, like a Forethought that may develop into the subsequent Salesforces, that may create all this chance, trillions of dollars’ price of opportunity. I believe that that is going to mean great business, but it surely’s also going to be amazing for purchasers, for consumers. When phones got here around, you now have a supercomputer in your pocket.

Now you are going to have an excellent brain in your pocket with GPT. The following time it is advisable to have a haircut, you only have your thing, your AI go and seek advice from the AI of the barbershop, and you may book a haircut. And hopefully, those are all powered by Forethought, effective. But what I mean? That is just a very different way of doing things. I believe economic models are going to shift. I also think that is an equalizer since it’s the type of AI innovation or the type of innovation that is straightforward to adopt. Before, you needed to have a PhD in AI to even know how one can do these items. And now, my mom is using ChatGPT, and individuals are beginning to get exposed to AI, it’s opening it up, more individuals are probably going to need to learn how one can code, more individuals are going to be prompt engineers, whatever it’s. And I believe it’s actually an equalizer, whether you’re an authority or not. And that is what I believe I’m most optimistic about is just that economic opportunity.

What are your predictions for AGI? Do you think that we’ll see it in our lifetime?

Yeah. First, I’ll start by saying no one knows. This might be true.

Anyone who claims to know is lying. But yeah, I believe we’re seeing enough advances. There may very well be AGI that exists today. Who’s to say that GPT and that class of models aren’t roughly as intelligent because the human brain? I do not know, possibly. It’s probably not. The answer is probably not, but a human brain has what? A trillion neurons, GPT’s getting up there. And so, what happens when this thing has a trillion neurons and it’s trained with the fitting data, and our brain’s evolved as a consequence of evolution, but in some senses, reinforcement learning is evolution on steroids for machines.

And all of those techniques exist right away. And it just doesn’t seem that farfetched to assume. And we’d need recent techniques just like the transformer and the recurrent neural network, were all advances that, in hindsight, seemed obvious, but on the time or simply before that, were not possible. And we couldn’t even do language before we had RNs, after which transformers completely killed the RN, after which now we now have generative pre-trained transformers and GPT. And so there is perhaps recent research to get there, but it surely’s now not far-fetched since the stuff is that good. And I believe we’re either already there in some senses. But we’re either already, we have already got all of the ingredients, and we just need to seek out the fitting configuration, or we’ll probably be there in the subsequent 30 to 50 years.

And I do not know when you asked me this or not, but what do I give it some thought, is I believe there are some risks. The very best analogy to this was any person, I believe it was Andrej Karpathy, Tesla’s AI person, described it because it’s like nuclear energy. It’s about as powerful as nuclear energy. Unlimited capability with fusion, whatever, to create unlimited energy and permit us to do things we never thought, travel through space, because we’ll have enough energy to make the trips. There’s going to be so many things which can be possible with nuclear energy, but you furthermore may have the dark side of those things, that are nuclear bombs, which I do not know if I’m allowed to discuss in these type of interviews, but you furthermore may have these items, or nuclear weapons, I’ll say, right? And that massive capability for destruction.

And the very best we have discovered as a society on how one can deal with that’s just everyone has it, and everybody mutually just put in regulations to be certain that all of us use these items accurately. I believe there’s an equivalent amount of power there, it starts on the digital sphere. So, you almost have less fear originally, because unless something like that gets access to the nuclear weapons, then, but my point though is, when you can create the guardrails and the safeguards as AI is advancing, which I believe we’re doing as a society, you then’re actually in a reasonably great place. Having AI that may detect fake news which is as powerful as GPT to detect GPT generated fake news is super essential. I believe I will be, again, optimist about it, but that is to not say we should not be concurrently developing regardless of the technology is to make it secure.

You like to read, what are some good books that you simply’d recommend?

Probably the most recent book I read was ‘The Advantage’ by Patrick Lencioni, which teaches about how one can construct a healthy organization, every part from the way you do meetings to executive team alignment, to communicating mission and values, and truly constructing a healthy organization. And it is a culmination of all his other books, like Death by Meeting and whatever. I’d say though, the Advantage might be essentially the most recent one I’ve read. On AI, I have not read, sarcastically, I have not read any AI books recently. I read a number of the AI papers, but I believe there is a book called ‘Superhuman’, which was highly advisable to me, that I want to go and browse. After which lastly, I’ll say, yet another non-fiction, ‘Good to Great’. My favorite management book of all time. Fiction, the Expanse. It’s now a TV series.

I’ve seen that TV series.

You’ve got seen it, right? I believe it’s on Amazon Prime. I adore it since it does such a great job of probably not science fiction, but science realism. I mean, there’s a variety of science fiction in it. But they take a variety of really subtle concepts, like what would occur in a world where we had space travel, and the way would politics evolve? And that is how humans would probably behave. For those who notice, it’s extremely, very, very subtle. However the phones have a sophisticated type of AGI in-built, and it isn’t highlighted. It isn’t even the central point. And that is why I prefer it. That is why I’m mentioning it here. But you will notice they’ll just say something to their phone, and it goes off and does it, after which a few hours later or whatever, in movie time, it’ll have done the duty. And it’s extremely subtle. You simply see people just randomly talking on their phone as in the event that they’re talking to an individual, after which they simply carry on, after which it comes back with some more and talks to them.

And I believe that is how AGI goes to operate in our world. Everyone’s going to have their very own personal Einstein or C3PO, or R2-D2, whether it’s on a phone or eventually in robotics. They usually’re just going to do tasks and answer questions for you and figure things out. And they are going to be as smart as a Einstein. And that is it. If every human on the planet had a friend named Albert Einstein, AGI, intelligence as well, or higher than, humans, you then’d just have the opportunity to do more stuff. And everybody all the time defaults, okay, well, what if all those guys became malevolent? Well, we now have human level mental creatures, eight billion of them. And I mean, humans suck in so some ways, but it surely’s also beautiful that we have been capable of create society and civilization that we feature on.

It’s interesting how opposite it’s of Star Trek where they make a show of using technology. 

We do not discuss phones, but 100 years ago, when you told any person you had a phone and you could possibly seek advice from any person in Qatar instantaneously, there’s so many things now which can be literal magic to individuals who would’ve been 100 years ago.

Like even how a Kindle would have been magic 100 years ago. It is a book that is all the time different depending on what you would like.

Depending on what you would like. It’s literally prefer it’s straight out of Harry Potter, right? Anyway, I believe that is how I’m once I take into consideration these technologies. You’re taking the analogy of anything that rapidly modified, and there are lots. Electricity was powerful, the printing press, after which it’s crazy and massive for some time, after which the subsequent generation just takes it with no consideration and it’s just in-built. Imagine a world where AGI is just taken with no consideration and just built into every part we do.

You lately tweeted, “Measure success, not by what number of stuff you start, but what number of stuff you finish.” What does the finish line look in your company?

What’s the tip game? Going back to the concept of latest economic business models, one among the things that I find super invigorating, it’s like when people say Google has no mote. In truth, Google said that it leaked or something, and everybody had a panic. I mean, it’s true. I believe they haven’t any mote. But additionally, OpenAI probably doesn’t have a mote. No one has a mote. But more importantly, the present business models of today, there are going to be many more stories in the subsequent coming months around business X has no mote. It’s being disrupted by AI. And that is going to occur in customer support. It should occur to assist desks, it is going to occur to CRMs, it is going to occur to serps, it is going to occur to every part.

I believe essentially the most exciting part about that’s there are a variety of corporations who are only jumping on the bandwagon right away, but this technology is as much a sustaining technology because it is a disruptive, and what I mean by that’s there are going to be individuals who adapt faster, but additionally individuals who have an inexpensive amount of scale can actually speed up at a better derivative than the individuals who are usually not yet at scale.

And sometimes that will not bear true. There is perhaps some smaller corporations, and it is going to be far and wide, but it surely’s not necessarily a given. And so, you may have these different strata of corporations, and there is perhaps the oldest ones who’re going to be too slow to adapt and get toppled over. You would possibly have the smallest corporations who move faster, and you may have somewhere in between. In order that thought that we may very well be the type of company that makes each touchpoint between people and businesses faster and more intelligent. Today, that is in customer support because that is essentially the most often interaction we now have. You might have an issue; you ask a matter. We’re helping power that. Today, we process over 100 million support tickets a 12 months, and we’re just getting began. Eventually, that is going to be a billion, 10 billion, trillion. And that is in customer support alone.

I believe AI, and hopefully through Forethought, you may apply this same technology to marketing. Why do you simply ever reach out to an organization when you may have a matter? What about when you would like to find out about a brand new product? What about when you may have a service that you simply’re all for? What about sales? What about you reach out to businesses that you simply work for? You might have a matter about IT? You might have a matter about HR? Each touchpoint might be transformed through AI. And I believe we now have the chance, genuinely have the chance, to be an element of that story, to be that company or one among those corporations that brings about this future, and that is the tip game, that is our mission. And ultimately, it’s about unlocking human potential. Ultimately, that is going to make every part faster, every part more efficient, give people time, energy, and a spotlight back in order that they can deal with spending time with their family members, whatever it’s. Humanity is leveling up.

And I believe we now have that chance. I believe we are able to try this. And I believe from a business perspective, that may very well be a multi-billion, if not tens or tons of of billions dollar price of company. It may very well be the identical way, same because the Salesforce story. What if Salesforce were in-built 2023?