Press Release

Dive into the $1.4 Billion AI Deal with iLearningEngines – Insights from CEOs Harish Chidambaran and Matt Safaii on SPACInsider


Dive into the $1.4 Billion AI Deal with iLearningEngines – Insights from CEOs Harish Chidambaran and Matt Safaii on SPACInsider

Nick Clayton (00:03):

Hello and welcome to another SPACInsider podcast where we bring an independent eye in interviewing the targets of SPAC transactions and their SPAC partners. I’m Nick Clayton, and this week my colleague, Marilyn Hadad, and I speak with iLearning’s CEO, Harish Chidambaran, and Matt Safaii, Chairman and CEO of Arrowroot Acquisition Corp. The two announced a $1.4 billion deal in April. Harish explains how the company developed a profitable AI platform with very little outside funding through the years and how SPAC cash and public share capital can accelerate its growth, while Matt tells us how Arrowroot’s long relationship with iLearningEngines gives it confidence in how the company will be valued in the public markets. Take a listen.


Harish, AI has come into the discussion recently as something very new hitting the market, but you founded iLearningEngines back in 2010. As you know, this is something that is not entirely new. You’ve been applying AI processes in your work for years. Can you talk us through a little bit that journey and how your tools have progressed through the years?

Harish Chidambaran (01:20):

Reality, this has been a culmination of years of very hard work by a team that’s really demonstrated tremendous leadership and expertise in what we believe is in building one of the largest and one of the leading enterprise AI companies out there. We are an enterprise AI platform for learning, and work automation, and information intelligence. The real inspiration for me in building this company came from a personal tragedy that happened several years ago. My mom was diagnosed with stage four breast cancer and that was her first diagnosis. Both during her diagnosis and her treatment, what I found was despite all the knowledge that was available inside health systems, much of the information was very siloed and not readily available to patients. I really felt like health systems could do a better job in delivering vital content in real time and in context. Now, this is not a problem unique to health systems but almost all organizations.


As we’ve gone into a data-driven world, data’s in very siloed forms. It’s not readily available. A big challenge for every company is how do we extract the signal or the important information from all this data and deliver that to users. We really were built to help companies make better use of their organizational knowledge. My background was I started out in Silicon Valley at Sun Microsystems as a microprocessor architect, and we were really leveraging AI principles at the time to vastly improve computing power. That really underpinned my conclusion that AI was the key in achieving this vision for us. We set out to build one of the leading AI platforms really built on very proprietary algorithms and datasets. We invested over 400,000 hours of R&D. Now, you should go back to all these years ago, I think people understood AI to be this futuristic technology with great potential.


When we were looking to sell, what was very important for us was to determine what are the right use cases where we can sell AI to customers. Really, what we came up with our strategy was let’s figure out ways where we can use AI to power important functions inside organizations. We really started out with learning automation as the key area. How can companies use AI to really deliver better learning and help really improve very mission-critical outcomes? That was the starting point. Since then, we really started to add newer and newer functions expanding, so migrating from learning automation to work automation and so on. What, from our standpoint, some of the key things that we built in our platform was really very proprietary algorithms and we’re using very specialized datasets to build pre-trained models.


Today, we’re in 12 industry verticals ranging from education, healthcare, oil and gas, and so on. Really, each of these places, when we go in, it is really our software plus unique pre-trained data models that we use each of these verticals.. We think this is AI has just really gotten significantly more interesting and exciting to companies, so the future is pretty strong.

Nick Clayton (04:25):

Great. Matt, I’m just curious, based on the number of companies that Arrowroot Capital evaluates every year, how long have you been familiar with iLearningEngines and how do you winnow down your search to this company?

Matt Safaii (04:37):

We’ve been working with Harish and team, we’re in our second year now. We’re really culminating the partnership, and building that partnership and trust between the two, and learning about the business. We’ve seen it scale over time. It’s hard to find in this market. We’ve looked at a lot of businesses. Our overarching thesis was to find one to $3 billion business that was going to be the next $10 billion plus leader in its category and we stayed disciplined through that. We did have to kiss a lot of frogs. Process took a long time, but we’re happy where we are right now.

Marilyn Hadad (05:12):

Great. Could you walk us through what some of your typical customers are like and how do your services grow within their operations?

Harish Chidambaran (05:18):

We are in, like I said, in 12 industry verticals where they are ranging from healthcare, education, oil and gas. We’re selling to enterprises. That’s really what we do. We’re really selling to medium to large enterprises. That is our sweet spot in our go-to market.

Marilyn Hadad (05:46):

With your existing operations already spread into India, the UAE, and Australia, how much does your software and its machine learning need to adjust for things like language across those diverse clients?

Harish Chidambaran (06:00):

From our standpoint, we are a global company with operations in all these areas. Our customer base is predominantly English operating customers. That doesn’t change the fact that we’re, but that’s where our platform supports multiple languages. Then the key to really understand what has really given us a pretty strong competitive mode from the beginning is more focuses on the enterprise, where the reason and why we focused on proprietary technologies and proprietary datasets is to give companies the capability to build their own institutional knowledge. How can they take their own institutional knowledge, make better use of it? To do so, and really what we’ve ended up becoming as a platform, is a company that can really help enterprises harness the full power of AI while managing the downside risks to it. I think that’s been a critical part to how we have built our product and grow.

Nick Clayton (06:57):

Great. Drilling down into that a little bit more, just among those core client verticals that you have, what are some of the examples of where you’re seeing some of the highest effectiveness of your tools in terms of cost reductions or however companies are managing or measuring rather how effective these tools are for them?

Harish Chidambaran (07:14):

From our standpoint, like I said, we are in a wide range of industry verticals. I’ll just give you an example of a case study just to illustrate the power of what we do. We had a shipping company in the oil and gas industry that was recovering from an accident. From their standpoint, they attribute that to employee errors. Their big thing was, “”How do I put a system in place to reduce employee errors?”” I think this is really where we, as a company, deployed our platform. At a scope, when we go into a company, we help them what we call create a knowledge cloud. We took all our platform, take all the content from within the company around safety, maintenance, different areas, and our AI platform can take a PDF document or a video and cover them into knowledge artifact by inserting assessment, quizzes, surveys, et cetera.


Using this platform, they were able to create knowledge artifacts and drive the consumption of those artifacts into the various workflows. We were able to really drive a consumption of company knowledge over 58 times. Some of the other things where there were a lot of leading indicators. These accidents don’t happen randomly . One of the key things that the company’s market is, for example, zero time loss due to incidents. That’s something that they sell. Those are the quantifiable metrics. We go into any business unit, we ask business unit owners to identify their top KPIs, and then use our knowledge cloud and our learning and work automation systems to improve those outcomes.


I think that’s really what is critical and why this naturally helps us in upselling, is we’re able to go business unit to business unit. We start out with business unit A, they’re taking their KPIs using our platform to improve those KPIs, and they’re looking like a state-of-the-art business unit. There’s a big competitive pressure for other business units to adopt the same thing.. The other part to this is most of our contract that we sign are multiyear contracts. We are really used in very mission-critical applications, and so I think that commitment from both sides is really critical to our customers.

Nick Clayton (09:21):

Yeah. You mentioned in your announcement materials that you plan to use this transaction in part to engage in some M&A. Are you looking to use those moves to further expand geographically or primarily to pick up talent and technology for the overall platform?

Harish Chidambaran (09:35):

From our standpoint today, we’re a business at scale. We are an enterprise AI business at scale. We have over 4 million users powered by a platform and over a thousand end customers on our platform that contributes to our pretty strong revenue base. We have gotten there on the back of a very strong organic growth strategy and we’re also supplementing or complimenting that with M&A. Part of what we really liked in this transaction with Arrowroot is their expertise in enterprise software and in M&A. We see a tremendous opportunity to acquire companies with the idea of bolting them on. I think I would say a significant portion of our acquisitions are going to be around acquisitions to acquire new customers. I think a smaller portion of our acquisitions is going to be around adding new products. We’re a company that takes R&D very seriously. We spend over 30% of our revenue on R&D. From our standpoint, our real focus here is going to be on revenue growth and then continue to strengthen our product.

Matt Safaii (10:35):

Yeah, it’s part of the reason we’re excited about the transaction and to bring our expertise here. As a growth equity firm, a private equity firm, we track 25,000 plus software companies globally and the platform of ILE really could strengthen or power a lot of these software and SaaS businesses in the background and make them better. The platform is so broad, it’s really… It even goes to we’re drinking from a fire hose in terms of which use cases should we be selling, sector selling to, and also what M&A we should be doing. I think that’s our challenge that we have. As a public company, we will be acquisitive and we’ll be fairly aggressive at it.

Nick Clayton (11:12):

Great. To that point, as far as I’ve been able to tell, it appears the company has developed itself through these years with very little outside capital to date. I mean, could you comment on that and just how has the company managed to do some of the things that’s gotten accomplished so far with that efficiency?

Harish Chidambaran (11:27):

We have raised less than $2 million in equity. We’re an incredibly capital-efficient company. We’ve been a high growth software company, but a profitable high growth software company. We focused on profits and becoming profitable well before it became fashionable to become profitable. Really, on the back of this profitable high growth, we’ve been able to also get access to a lot of debt and venture debt. We’ve had some pretty strong venture debt partner. It really on a back of less than $2 million in equity, we’ve been able to drive our growth so far. I think it’s really a testament to the product that we built, our go-to market strategy, and really the demand that we have been seeing in AI well before all the recent hype that we’ve seen.

Matt Safaii (12:11):

I would add it’s testament to the team and Harish, or for being disciplined through a time when everybody was getting drunk over the excess of capital available, really that put a lot of companies in bad situations and created serious misalignments. Harish and team never did that. For us, looking at this as a long-term investment just as a fiduciary of capital, let’s say, we feel quite confident with the team.

Marilyn Hadad (12:34):

Great.. So for those that are perhaps not as familiar with the term in the software space, can you just explain what that is and why it’s so important?

Matt Safaii (12:46):

The way you run the business and it’s the outcome of how you’ve run the business, but the Rule of 40 is the growth percentage of the business top line and then the percentage of profit margins. If it equates to 40% plus, it’s a Rule of 40 company. It sounds simple, but when you look at it from the delivery of product and the cogs, the gross margins, how much it costs to get new customers, say just going down the opex line, it just a Rule of 40 companies are these top tier businesses. It goes back to the efficiency and fiduciary of capital that Harish and team have done and the discipline they have.

Marilyn Hadad (13:24):

Given all the recent chatter around AI and potentially increasing regulations on it, is there anything out there that might perhaps impact the way that your programs operate in terms of how they source information?

Harish Chidambaran (13:38):

Yeah. No, absolutely. I think as an AI platform for the enterprise, we’ve always taken the role of supervision and having an AI that not only allows a platform that not only allows companies to leverage the power of AI but also manage any downside risks. We truly believe we are a disruptor in the space largely behind our proprietary technology and our specialized datasets, and most importantly, a strong supervision of our AI outputs.. This is how we have always operated, because we just didn’t ever think that selling an AI platform that is uncontrollable is a good business strategy. We think that these regulations are really a must have, especially in the enterprise world. From our standpoint, we feel like this has just been a very strong validation of our strategy.

Matt Safaii (14:47):

I would just add it was a big part of our [inaudible 00:14:50] investing committee process or conversations at the board level at SPAC around AI. When the whole ChatGPT generative AI forever happened was we didn’t want to target a business that was trying to figure out its use case and figure out its ROI to the customer. A lot of these businesses, which are getting high values, are burning tons of cash. I think they’re meandering around still trying to find their way and their cadence per se, where iLearning had figured that out. It’s proven from its growth quarter over quarter and the profitability and metrics quarter over quarter.


At the same time, we were also looking at the regulatory risk of it all. I think there still is deep regulatory risk in a lot of these AI plays, but I think that’s taken away here by iLearning by using their proprietary datasets and using the data within the organization to have the machine learn over time, including the activities of the “”experts”” at the enterprise to help learn the machine, which Harish could get into. We thought those were both taking away and there’s frankly not many companies in the AI space that have both of those boxes checked.

Harish Chidambaran (15:53):

Right, yeah. I just want to add one thing to our [inaudible 00:15:56], that I think the discipline that we use in building or focusing on not just revenue group but also profit, I think, has naturally extended into our approach to AI in itself. We see lot of players in the AI space and this has generally been an area of focus. If your AI is a perfect system, the market is infinite, but if your AI is even 90% right, the market is zero. That’s how most people have looked at it. For us, that discipline to be focus [inaudible 00:16:24] always meant identifying opportunities where companies can use AI.


If you go back to our track, we’ve been selling AI systems for over five years. Really, during those days, we had to pick areas inside an organization that are mission-critical, but also something that they are more likely to invest in. I think that is really where our first focus was on the area of learning because that’s an area where you can really improve the way organizations can perform learning functions and then slowly adding more work functions. I think really all of these things, I feel, are interrelated, being able to have a very supervised approach to AI, making sure that you can focus on good business discipline. I think sometimes we tend to think of them as two independent things where I think they’re already related.

Matt Safaii (17:11):

Sorry to pile on even more here, but for everything Harish said, Arrowroot doesn’t perceive iLearning. Obviously, it’s of scale at profitable as venture capital risk, where I think there are tons of AI companies that are worth billions of dollars that have VC risk, have binary outcome risk here. That’s something at the SPAC that we did not want to take on and that was from the start.

Marilyn Hadad (17:33):

Right. Are your AI models primarily tech-based or data-based? Do you have any image-based tools as well?

Harish Chidambaran (17:43):

Without getting too specific, I think what I would say is the kind of data that we use to train our AI models are highly specialized. There’s a lot of data out there in terms of how people travel, spend, entertain, et cetera, but the kind of dataset we use are how people engage with their learning content or their work-related function. For example, our safety folks inside an enterprise engaging with safety related content in their work functions, how are folks in the finance world engaging with their kind of content at work, or if you’re in the education vertical, how are high schoolers engaging with math, science, or literature content. Very specialized datasets around really how people engage with various functions. The content that we had in what our system does is we can support content from our PDF video images, et cetera, but the core of this is also those data around how people engage.

Matt Safaii (18:42):

It goes to the add-on acquisition strategy and potentially acquiring other software companies, software and services companies that have a lot of interactions. Not just content but interactions with outcomes because we could utilize our technology as the backbone and pinning for AI solutions absorbing all of the interactions and content that other software companies might absorb.

Harish Chidambaran (19:05):

Just to close the loop on this, we think these specialized datasets are a very strong competitive mode. This is an area of interest for not only startup but also very large companies. I think one of the key things is the kind of dataset that we have been acquiring are highly specialized. These are not very readily available and we think this gives us very strong competitive modes going forward.

Nick Clayton (19:29):

Well, great. Well, I think that moved us naturally too. We’ve talked about what you built so far, but going back to the here and now moment of it and the timing, what were some of the things that made you decide that now was the right time for iLearning to take this step, both to go public and to do it through a SPAC?

Harish Chidambaran (19:49):

From our standpoint, we feel like we are an enterprise technology company at scale. From our standpoint, we feel like we have very strong competitor to advantages. We think the market right now is very aligned with our strategy with tremendous opportunities ahead, so we felt like going public was the next natural step. Because we had raised smaller amounts of equity, we felt that having Arrowroot, and SPAC, and the broader backing of the private equity company behind us would be very helpful for us post going public. Their strong understanding and enterprise software and track-regarded M&A, making sure that we have the right systems in place, we think those are all really critical for us.


Going public was a very natural step in our evolution, but we also want to make sure that when we are going public, we’re not making any missteps, that we’re doing it really right, that we have all the right pieces, both operationally but also strategically from an advisory standpoint. For us, this was a very natural decision to make working with the Arrowroot team. We’ve been working closely together now for over a year, and so we really got to see how well we can work together.

Nick Clayton (20:58):

Great. I’m interested in your valuation process as well, Matt. There’s plenty of even more diversified software companies that are not pure play AI and maybe don’t have some of those same dynamics out there, trading higher, sometimes several times higher. What did you see as the advantage for shooting at this particular price point and how it’s going to sit in the market?

Matt Safaii (21:22):

Yeah. I think the key theme overall when looking at the valuation was a successful transaction and long-term view of it on both from Harish and iLearning and from the Arrowroot standpoint. We wanted this to be successful and we believed… Really Harish and team believed that being a public company from a long-term perspective, was the right move. As we look at the valuation, we did an analysis recently of public software companies, not just AI businesses, with the metrics that iLearning has, and there’s something around 2,400 software companies out there. If you whittle it down in terms of our business, it’s growing over 30%. It’s got the 115% plus net retention and it’s profitable. If you whittle all of it down, how many companies are trading below four times, the answer is zero.


We’re pretty confident that this is a good transaction, a good deal. Frankly, I think if iLearning was owned by a private equity firm or VC firm, they’d be looking at this way more transactional, which I think a lot of SPACs do in these and aren’t looking what’s best for the company from a long-term perspective. Harish, he’s the controlling, he’s the founder, and he has the control to do that and take that long-term approach. We think it’s a great deal for all those reasons and believe it’s going to be a great public company.

Harish Chidambaran (22:41):

Well, that’s really well said, Matt. From my perspective, going public, this is a very long-term decision and it’s not a decision that we think is short-term or transactional. How do we structure something that will provide success to the company over the longer term is really the key and to our investors. We felt like it was really important for us to price this thing in a way where people who are coming in here will also feel like this was a very fair trade and these are people who will continue to be people we can go to in the future as a public company.


I think the decision that we make other company having raised less than 2 million in equity and are looking at purely a very much a long-term horizon is obviously very different than decisions that a company would make that’s been backed by institutional investors who deployed hundreds of millions of dollars and are in that liquid asset for five, 10 years. I think this is just a very different strategic decision that we have to make. Going public for us is very different than going public for many of those companies.

Marilyn Hadad (23:40):

You’ve also set out to raise at least a hundred million dollars in capital for this deal, some of which will come from Arrowroot’s trust. How are those funding efforts going?

Matt Safaii (23:50):

Yeah. It is going well. We’re engaged. I mean, there’s various ways to raise the capital. I’ll just leave it out. We’re highly confident that we’re going to get something done.

Nick Clayton (24:14):

Great. That’s actually what I wanted to ask about next, Harish, is that given the profitability, have you guys thought about how you plan on allocating that capital, given the fact that once you’re a public company, some of these M&A transactions could be done with shares? Just how would you like to ideally structure some of those transactions once you get running with it?

Harish Chidambaran (24:35):

Right. I think from our standpoint, I think you’re absolutely right, Nick. One of the big drivers to go public is really we feel like it can really facilitate a strong M&A strategy here. I think on the back of a public stock currency plus cash, the promise of cash per stock, I think would be a really good sweet spot. We’re in 12 industry verticals, so it gives us an opportunity here to make acquisitions in vertical-specific areas. It could be in products, software companies. There are lots of companies in the five to 25 million revenue range that are really looking for homes, a lot of companies that have good product technologies too. I think it’s a combination, but we think being public and having a combination of stock plus some cash will allow us to get maximum value out of these transactions. That’s why we feel like raising this capital is really good for us.

Matt Safaii (25:23):

I would add that it goes to being public and having public shares plus having some cash on the balance sheet. This company has been run so efficiently and it’s grown so fast that we are helping… Arrowroot is helping Harish and team with recruiting and hiring. We have an internal recruiting team for our portfolio and we’ve already hired some quite pedigreed professionals and experienced professionals. That’s going to continue, so I’m sure you’ll see more of that. And that’s just a function of being larger. Frankly, we’re going to have that currency and that cash to keep going and make sure that the right team is around Harish.

Marilyn Hadad (25:57):

Great. I’m curious to hear what’s the most exciting thing that you see coming now in your space in terms of the new capabilities of your technology?

Harish Chidambaran (26:06):

I’ve always been in the tech world and I think our focus has always been on making sure that we have really cutting edge products. We believe we have one of the deepest product and technology roadmaps. We’ve always really believed in the power of AI and what it can do for enterprises, both in terms of enhancing the capabilities of their human capital and their business performance. There’s a whole slew of products. I think, for us, what really excites us is building new technologies and making sure that those technologies have a great home in the market. I think, for us, we’ve been in 12 verticals so far, and the idea of adding new capabilities within these verticals and also add new verticals is really exciting. The impact, the more you see… At a very high level, I say it where AI can really make a difference is we think there’s a tremendous opportunity for AI to do to traditional brick and mortar industries what going online did for retailers.


Having AI power education institutions, AI power health systems, those are really exciting things. We think there’s also a great opportunity here to bring in capabilities around AR-VR into making sure that that can be used to train employees better. I think there’s a whole slew of exciting areas. Generative AI is another very exciting area that’s out there, but I think, for us, as a company at scale, what we really enjoy is make sure that we have cutting edge products and those products have a great home. I think that’s what has really excited us.

Matt Safaii (27:38):

Yeah. I would add that as we look at it, this is a broad platform. When cloud started, there was the ServiceNows and the Workdays of the world really creating the cloud infrastructure. Maybe people didn’t recognize in the beginning, but over time, that’s what occurred. Arrowroot looks at iLearning as really building the AI infrastructure, and there’s only a couple players really doing that. It’s very exciting, but it’s also challenge. What I said prior, it’s like drinking from a fire hose and what do you do, what use cases, what do you acquire? Because it is so broad and there’s so much inbound activity. I think that is a challenge. It’s a blessing and a curse.

Harish Chidambaran (28:18):

To summarize what having Matt said, I think we are at a point in the AI industry. Really, if you think about the market, we think the sheer excitement the AI industry is going to create a new class of companies much like what cloud computing and SaaS did a few years ago. I think while AI products and other things can be exciting, we do think that the opportunity for us as a company that’s ahead is even more exciting for us.