Building an AI Startup: What's Different in 2025?

Michael Krigsman: We are discussing AI startups today on CXOTalk, episode 871, in a conversation with Arvind Jain, the founder and CEO of Glean. He was a distinguished engineer at Google before starting and taking public cybersecurity company Rubrik. Glean has raised $600 million and is valued at almost five billion.

Arvind Jain: Glean is an enterprise AI company. Think of it as Google or ChatGPT, but inside your company. That's what we do.

Michael Krigsman: And you are, shall we say, born in AI or AI-native companies? You're the perfect person to help us understand what's happening with AI startups today.

Arvind Jain: Absolutely. We are, I think, the first company to bring transformer technology to the enterprise. It's a lot of great learnings, starting as a native generative AI company six years back.

Michael Krigsman: AI company for six years. That's well before the marketing hype surrounding ChatGPT and OpenAI.

Arvind Jain: That's true. The term generative AI did not exist at that time, but the core technology that powers generative AI, the transformers, they were there, and we were using them in early 2019 to understand enterprise data, knowledge, and information at a really deep level using AI and then making it all searchable for people inside a company.

Michael Krigsman: Can you describe the differences between AI and traditional technology startups?

Arvind Jain: Every company, first of all, whether you are a startup or you are a mature company, I think AI is becoming such a core fundamental tool that you've got to use in products that you build because that's the way to stay ahead. That's the way actually to build new, amazing things. I actually feel that's my view. It is probably all startups that you think of today; you can call them all AI startups.

However, maybe the other way to look into this is to see if some companies are actually building core foundational technologies. For example, they're building models and infrastructure to train models. And maybe that's one set of companies.

But then the vast majority of AI startups that are getting started today are companies that are thinking about solving a business problem. In this consumer problem, they feel that AI and the new reasoning and generation capabilities are going to play a big role in the product that they're going to build.

But like us living in the AI world, I've engaged with many startups, and I've not seen a single one that is not making AI a big part of its core tech stack.

Michael Krigsman: As AI has matured, there's become a greater clarity that AI, as you just described, must support the solving of some business problem for AI startups where the centerpiece is that AI technology. What are the differences from, again, traditional software companies where yes, they're solving a business problem, but the technology is very different, the underlying foundations?

Arvind Jain: AI technology moves very fast. It changes at a very rapid pace. When you think about product development, the new AI startups, they're going, they're able to, first of all, they're very lean. They're able to actually do a lot of things because software programming, building systems with AI, these things are becoming much easier than in the pre-AI world.

I think, from my vantage point, when you think about investment that is happening, most of the investment from venture capital actually is coming into companies that are making AI a big part of their story.

And I can compare more with, let's say that startups that were getting started two years back versus today. The key difference between them is thinking about this new model where you can truly build products which are a lot more powerful, a lot more capable than what you would've even thought of a couple years back.

But as such, I think from a company-building perspective, a startup is a startup. We are an AI company, and when I compare this with my previous startup, I would say that most of what we do doesn't change. We still have to think about the core business problem.

We have to figure out how we're going to actually solve that, how we're going to build great technology and the teams to then bring that technology to our customers. Personally, I have this feeling that yes, AI is becoming a big part of our technology stack, but fundamentally how we build and design companies is not changing, except for maybe one thing.

You do hear sometimes companies saying that now with AI you can actually have a one-person company that can generate a billion dollars. And you see a little bit of that in the new generation of companies. You are seeing some of these AI companies scaling up revenue at the pace that the previous generation of SaaS companies couldn't.

I read these examples all the time: a startup actually reaching $20 million in revenue two months after it's started or a company that's at a hundred million run rate within the first year. These things were not possible in many ways in the pre-AI world.

But given how fast you can build products and how different they can be compared to the current products in the market, it's allowing people to scale fundamentally at a different pace and level than the previous generation of startups.

Michael Krigsman: I just want everybody to know that you can ask your questions right now. On Twitter, there's a tweet chat, X, on X, I should say. There's a tweet chat, an X chat taking place. Use the hashtag CXOTalk. If you want to ask your questions, if you're watching on LinkedIn, just pop your questions into the chat.

Arvind, you raise a very interesting point just now how AI startups can generate so much revenue with relatively few people. What is it about this technology and the nature of the startups that are happening around this to enable this phenomenon?

Arvind Jain: First, you can actually build products that are very different in their capabilities than things that are out there where when you think about, for example, software development, there are companies that are offering new development environments which allows developers to go five times faster than what they could before.

And there's such a big leap in the capabilities that these products bring, which creates that instant excitement in the market for them. These new coding tools, even for us as a native generative AI company, we see it every day.

People just want to use these tools. There is a, there's a big demand from the ground up. There's love for these new AI products. That is incredible. And that is what's allowing them to actually launch a product in the market, have a product-led growth motion, and there you go. People come in; they want to use these products. I think largely it's driven by these brand new amazing capabilities and software doing things that people didn't expect it could do. I think that's the biggest part, that's what creates that excitement and then instant demand.

But the other thing which is actually allowing these companies to have the success in a short duration of time is that software development itself has gotten very accelerated when you use these new tools. You can build products which are a lot more capable than current products that we have in the market, and you can do that in a short period of time. You can actually build something amazing in a month, in two months, something that used to take you a year or two years before. That's the other factor that is actually shrinking these cycles for these complaints.

Michael Krigsman: You have this combination of tremendous interest and market demand combined with significant developer productivity increases due to the tools. Is that a correct way of saying it?

Arvind Jain: That's right. Yes.

Michael Krigsman: We have an interesting question coming from LinkedIn, and this is from Santhosh Siruvole who says, who is actually using and paying for AI services?

Arvind Jain: AI has captured everybody's imagination. There is no enterprise in the world today that feels that, well, we can, we can look into AI a few years from now. Everybody knows that they have to act; they have to embrace AI today if you want to be, if you want to stay relevant. Starting from that high-level desire to actually bring AI into your enterprise, now what we are seeing is that there are two different ways of people are bringing AI into the company.

One is the CIO, and CIOs obviously hold the responsibility to bring the right set of technologies to the entire working population of your enterprise. They are taking the leadership in terms of bringing broadly applicable, useful products that are AI based to their enterprise.

And that's for us, for example, at Glean, they are our primary personas too because Glean, as a tool, it's a knowledge access tool. People go to Glean, they ask questions, we quickly answer those questions for them using all of the enterprise context and data and knowledge.

And that's a tool that's actually every knowledge worker, whether you are an engineer, a support person, somebody in HR or IT or sales, all of you have the need for that. All of us have questions. All of us have tasks that we think AI can do for us now. It's a broad tool, and CIOs are the ones who actually typically will purchase a companywide tool like that.

But then you also have every individual functional leader. If you're head of support, if you're the CTO and you're trying to actually make sure that your engineers, your developers are productive with AI. If you're a salesperson, the sales leader, and you want to make sure that you're using modern ways to prospect, to reach out to your customers, to actually have powerful engagements with them function by function, you're seeing every function leader in the enterprises looking into and evaluating AI tools and bringing them on board.

This is a very broad phenomena, industry-wide, vertical wide, geographical. I've been traveling a lot across the world. I don't see any company that is not paying attention to AI, any country that's not paying attention to AI either. This is, it is, everyone really is. The answer is it is bringing AI into their enterprises.

Michael Krigsman: Subscribe to our newsletter, join our community, go to CXOTalk.com, subscribe to the newsletter, check it out. It's following a traditional enterprise software purchase process in a way, because as you just described, you've got IT and the CIO, and at the same time, you have functional line of business leaders: HR, marketing, whatever it might be.

Both of these groups are looking at AI products, and I'm assuming this comes right back to what you were saying earlier, which is the core issue. It's not the technology, it's the business problem that's being solved.

Yes, we have a very interesting question from LinkedIn. Keep your questions coming. We have some questions on Twitter as well, and we're going to get to all of these.

Richa Varshney asks, how do your offerings address the AI ethics, jailbreak, prompt leakage, and AI transparency/interpretability risks? And I'll just mention that Isha is Senior Director of Risk Systems Development at Freddie Mac, and of course, financial services is keenly interested in these topics.

Arvind Jain: Glean is an enterprise AI solution. We work with the most prominent enterprises out there in different industry verticals. We work a lot with financial services, and these risks are all very real. I think with AI, it's a very powerful technology. And it's also, it is grounds for innovation and folks who are trying to actually create security problems.

You have to be very careful in terms of rolling the technologies in the right way, in a secure way, and also in a way which basically ensures that the technology is unbiased and it's doing stuff that you think is safe for your, in your context, in your enterprise.

I'll give you some examples of problems that you have to solve on the security front. One of the key things with AI is that if you're going to make it useful for your enterprise, for your organization, you have to take these language models that are built using the public web data.

They don't really know much about your enterprise and your information, your data. Somehow you have to figure out how you're going to actually connect that enterprise context with the power of these language models and do it in a safe and secure way. For example, if you train the model and you connected all of your enterprise data and knowledge with that model, and now anybody can go inside your company and ask questions, it's going to leak a lot of sensitive data to people who should not have actually have access to that information because enterprise data and knowledge is very, it's private in nature.

Like a given document, only a few people may have access to it inside the company. You can't take content wholesale in your enterprise and just train or build any AI system with it. Any AI system that you build, it has to understand who's using that particular system or software.

And how do you make sure that AI only uses information that this person is entitled to, is allowed to use inside the company and create those safe and secure AI experiences? That's one of the problems that we solve in Glean. Any AI usage, any agent that you build on Glean, you have to use that as an employee and you have to be signed in and then they'll actually ensure that whatever we do for you is done with knowledge and information that you could access too.

Michael Krigsman: I would like you to join the CXOTalk community. Go to CXOTalk.com and sign up for our mailing list so we can notify you about upcoming shows because we have amazing discussions like this. Let's go to Twitter and to Arsalan Khan, who says, do you think AI will become a simple plug and will become simply plug and play even for non-technical employees in the enterprise? The broader ease of use for the rest of us.

Arvind Jain: Absolutely. I'm both hopeful and confident that's what's going to happen. I think the true power of AI will be realized when it's so easy to use. You can actually do things with AI as a business user. You should not be required to understand how to code, how to build systems, be an engineer.

For example, let's say you are an employee in the legal department and you review contracts every day. It's a process that takes you a lot of time, and you should be able to just ask AI, and say that look, this is how I review a contract. This is a process that I follow.

And you should be able to just say that to an AI system that AI should then goes and identifies that particular business process for you, automates it for you. AI has to truly become accessible like that. With our agent platform, that's what we do.

We are actually really thinking about how to make this technology super accessible. Make, make sure that it doesn't matter who you are, you could be in HR, in finance and legal and you don't need to know anything about AI, how it works.

What are language models? That's not what you need to worry about. You need to just work with a smart, you, like working with AI should be like working with a really smart person who you could actually go and get some work to. You just tell them how this work needs to happen.

And then AI just makes it happen for you. That's the model that you're going to see. Any, like the agent, AI agent platforms that are going to succeed in the market are going to be of that nature. You have to elevate the capabilities of these systems. You have to make them more accessible to non-technical users.

And the other thing that I also add to it to go one more step beyond that, people are not seeking help from AI as much as you would expect. We all have habits, we do things the way we do, or there's a lot of inertia in terms of thinking about should I do this task differently?

People don't think that way. You don't have time often most of the times. You don't even have time to think about should I change the way I work. And it's not only that AI has to be easy and you need to be able to summon it and get it to do work for you.

AI also has to follow you and it has to come to you and say that, look, hey, I'm observing that you're doing this work every day and spending two hours trying to get this work done and I can help you with it. It has to come to you, and that is when you really see people will embrace the technology at a large scale.

Michael Krigsman: I'm looking forward to that day happening 'cause I can tell you I use large language models every day and I use multiple language models and different modes, research, non-research and so forth. And so much depends on the model you're using. It seems the time of day, the phases of the moon, how you construct your prompt, and then the whole thing's just a pain in the butt.

Arvind Jain: Yeah, it is hard. It's not easy today. And then by the way, you are using what I would say probably one of the most accessible tools. You're just talking to a system like in natural language. Yes, you have a few knobs to select here and there, but it's gonna get easier and that's what our job is.

At Glean, one of the things that we do is, we are not a LLM company. We're not building and training these foundation models, but we are, our goal is to see that you as a business user, how do we make sure that all this innovation that's happening in the industry, how do we make it more accessible to you? And make things easy, as seamless as again, the working model for me with AI is AI is like a smart human that's been in your company from day one.

They know everything about your company. They know all the people. They've read all the documents all the, they've been part of every single meeting, and they're ready to now help you 24/7. Just ask them what you want and they do it for you. That's the right model for when AI is going to deliver true value in enterprise.

Michael Krigsman: Greg Walters says, Glean looks to be the glue or connector between disparate databases. Do you see a future where this function is no longer needed?

Arvind Jain: Enterprise environments are so complex. Today, any large enterprise, they have thousands of systems, data repositories, databases, unstructured data repositories, documents and the like.

The true magic of AI happens when you have access to all of that information across all of these different systems. Maybe I think things are definitely going to get simpler in the sense that AI is going to get smarter and smarter in terms of being able to connect with all of those different systems over time.

And a lot of things that we do today at Glean, which we have to actually do a lot of hard work for to actually connect with these different enterprise systems, like we do expect it to get easier over time. Yes, I think that's the, if the answer is that the complexities are not going to go away, then I think AI is not doing its job.

Michael Krigsman: Dave Brace on LinkedIn asks, is it likely that non-deterministic agentic systems can honestly prove that they are truly trustworthy in the enterprise and that business leaders can trust products like Glean to make thousands of business decisions every day?

Arvind Jain: AI is non-deterministic. It can also be wrong. It can hallucinate. To some degree humans are also like that. If you have questions and you go and ask somebody sometimes, they don't have the full context, they're going to give you an answer, and it may be wrong, or it may be incomplete.

And I think that's just a fundamental thing to remember that these AI systems are actually more like humans and less like machines that they used to. And now you have to figure out how do I use this technology? This is not perfect. This can, it is going to make mistakes.

How do I trust it with my mission-critical processes where precision is actually, absolutely required? And there are a few different strategies that I would say that as an enterprise you can think about. There's a lot of work where you don't need precision, work where you need creativity, and that's where this technology is already really well suited to do.

But then when you think about tasks that require precision, AI can be actually used in a few different ways. One is that, let's say that there is a business process. Now you're going to identify that business process. You're going to automate it. You're going to ask AI to understand that business process and come up with a plan, with the workflow to actually execute that business process from now on.

And when you use AI in this fashion, you will go and ask AI to actually build that agent for you, and it may make mistakes. You should go, you should be there, go and supervise it, go and ask it to tweak its work, go manually fix, edit it.

And ultimately in that collaboration with AI, you actually encode and build that agent, that workflow. But now this workflow is deterministic. You put some investment in it. AI helped you build this workflow very quickly, but you are totally in control.

You are monitoring it. You've gotten into a, gotten it into a place where now, as I said, it's deterministic and now this business process can actually run and it is okay. You don't need full automation. You can actually work with AI and spend that initial time, like an hour or two.

But now you're going to have this automation for years because, and it's no longer non-deterministic. You have to think about it like the technology allows you to do great things and you don't have to rely on AI to actually make complex decisions behind the scenes. You can actually work with it that initially and build that six systems with it as well.

Michael Krigsman: You're describing essentially the role of that phrase we often hear, the human in the loop. Yes, and what's the appropriate relationship between the person and the AI system that at this stage of development is a tool rather than, I'm looking at LinkedIn and Greg Walters says, the true magic of AI replacing old standard applications.

Arvind Jain: Let's take some examples. We initially were talking about a legal person that reviews contracts. Now if you get a hundred-page agreement, customer agreement, it's going to take you a week or two weeks to actually go and review this and make sure that all the terms and conditions meet your enterprise's requirements and you have to redline this document.

And you can say that I don't trust AI to do this for me. This is pretty sensitive stuff. But get AI to do the first version of the redlining, and it's going to do a great job. If you just tell it how you do it, you tell that to AI, it's going to get 90% of the way there.

And if it is 90% of the way there and now you can actually fine tune that and finish that work with your context and knowhow. That two-week task, now it's actually a one-day task for you. That's big. You don't have to, you don't have to actually aim for a hundred percent automation and remove yourself from the task completely.

There's, we can, I'm very happy if I get 90%, it's a big impact to the business.

Michael Krigsman: This is from Gursharan S. on LinkedIn, who says which agents are driving the most value in large enterprise functions. Even better if you have examples in an industrial AI context, and I should mention that Gerran is an AI product manager in metals mining. Which agents are driving the most value in enterprise functions?

Arvind Jain: Top use cases for AI today in the enterprise are the following three. Number one is general knowledge, access and assistance. You are a knowledge worker. You may be in healthcare or financial services or industrial sector.

And you have questions, you have questions that you need answers to. You have information that you need to do your tasks and you use AI to actually, help you with that. Tools like ChatGPT or tools like Glean, inside your company that are just general purpose.

They're not meant for a specific use case. They're basically knowledge tools. They help you, they make knowledge accessible from the world, from their enterprise to you so that you can move faster with your tasks. That's the number one use case for AI today.

And it's not surprising because this whole revolution was catalyzed by ChatGPT, and that's the application that comes to everybody's mind when they think about AI. The second use case is, I would say for software development which is around code generation, like developers, tasks, like how you build systems, develop technology that's fundamentally changing with AI. That's a very powerful use case.

A lot of correction there. And then the third one I would say is, is around service. When I say service, folks who are actually taking questions, taking complaints, tasks from their customers or their internal employees.

These are customer service teams, internal IT help desk, HR help desk folks who are servicing other people's requests and demands taking your company's data knowledge and automating a lot of that interaction with AI is, that's the big use case.

Functionally, those are the three top applications, and I think across industry verticals, that's what we are seeing. I'll use some examples on customer service. We have these large telcos that have 50,000 or even a hundred thousand customer care agents day in, day out.

They are getting questions from their customers that they need to quickly answer. And if you make those transactions twice as fast those are hundreds of millions of dollars of savings for those teams. That's a very powerful use case for AI today.

For Glean, we already talked about software development. We have large enterprises in industrial sector, in retail, where you are fundamentally changing how how you use AI to build systems faster, to test them to review, code to troubleshoot.

These are some of the use cases as well for across industry.

Michael Krigsman: This is again from Arsalan Khan on Twitter. He says, AI might be able to automate standard operating procedures, but what about the institutional knowledge that resides in people's heads? And he says, isn't this significant job security issue?

Arvind Jain: Today AI, even with Glean, we are able to actually tap into a lot of their institutional knowledge to then create these powerful ChatGPT-like experiences where you can come and ask questions and you can answer that using that institutional knowledge. When you think about institutional knowledge, it's actually present in a few different form factors.

You have documents inside your company where people write stuff. You have ticketing systems. You have databases, CRM systems. There's a lot of wealth of knowledge inside each one of these different systems.

But then there's also a lot of knowledge and communication tools like email or Slack. And there's an increasingly enterprises are changing behavior so that they can capture more and more of that institutional knowledge. One example of that is that, if two people are going to actually talk and have a meeting, record that meeting, or at least capture the summary of that what happened in that meeting and make it available to AI so that it can be tapped into in the future.

I think that's, these are all the things that you can do today. We capture all these forms of institutional knowledge to then actually make that knowledge work for you as an individual and help you. And is driving that.

People are also more motivated to actually capture this information more in our company. For example, we now record every non-one-on-one meeting. Let's say, a confidential meeting between the employee and a manager, we won't. But if it's about a technical discussion, it's about getting some work done, we'll typically record those meetings so that all of that data is available to AI in the future to help us.

But now coming back to your second question on does it actually create an issue with job security? We, I believe that as an individual, the strategy for you to make sure that you stay relevant is, go and learn AI. These tools are amazing.

And don't think of AI as something that's going to take a job away from you. I don't think AI is that powerful. I don't think it's going to do it for most people, but you'll certainly lose the edge against somebody else who knows how to use these AI tools and work faster and better than you.

I think what as individuals, what we need to do is like this. It's been always like that. When new technologies come, folks who embrace them, who are quickly try to learn then are going, who come out ahead.

Michael Krigsman: I think this is a very important point, and I certainly give people that same advice that you must be learning how to use AI to make yourself faster, better, more efficient, and there are so many jobs that will get displaced. If you're listening and you're not doing that, well, I'm assuming if you're listening to CXOTalk, you are doing that, but tell the folks you work with who maybe are not so far on the cutting edge. Give them that advice. It's good advice.

Let's come back to startups and we have a question from Twitter directly on startups, and this is from Gus Bekdash who says, Arvind, what advice do you have for people forming AI ventures, many make the mistake of choosing problems that the platforms will take over making their companies irrelevant.

Arvind Jain: I actually don't believe, first of all, that if you start a company, if you start to solve a problem that you are going to fail because of somebody else, because of for example, an incumbent or a large company, actually solving that problem and solving them before you, you got to have a firm belief like as an entrepreneur that you can solve the problem faster than a large company.

A large company always has lots and lots of things to actually worry about. They're structurally not designed in a way that they can move faster then you as a really, as a nimble, early startup. I don't actually agree with that premise of people making that mistake.

Most of the time startups fail because people give up, because you lose confidence in your own idea, you don't have enough conviction, because if there's a real problem while you have the right to go and solve it and you will be, and competition is not something that's going to matter.

With that said, I will add what's the right strategy for you to choose as a founder? Pick, I always like to pick problems which are, first of all, they're obvious. If you go and talk to five people, they'll not argue with you that, hey, is this a problem or not?

I think if they're all, if you talk to first five people and they're, and you don't have that clarity from them, then there's something wrong and you got to work a little bit more on your idea. Get to that level where whoever you talk to actually agree with you that yes, you're solving a real important problem.

And also I like to work on problems which have broader impact, big problems that a lot of people are going to have because that's going to actually create more opportunities for you. Even if there are a few other companies that also solve the same problem, where there's a large market that you can tap into and you'll have your own success, along with them.

And then over time, if you do the best that you're going to win, like over everybody else. And then the last thing I would add is, don't come up with an idea where you feel like all I need to do is use AI and it's going to actually solve this particular task for me.

Think about if it is easy to build, if it is super easy to build, then everybody else can also build it. And then you're not really adding value. AI should be no more then one of the tools in your toolkit to solve that problem, but make sure that there's something substantial that you're building.

Michael Krigsman: There are so many companies that are building wrappers around the models. What about consolidation among these types of companies as the models advanced their capabilities?

Arvind Jain: If you are a startup and you are a thin dropper or the core capabilities of an LLM, you will be, you'll be irrelevant quite soon. You have to, in the mindset, I'll tell you what we do at Glean. We know that, the way our product works is that we actually work with all the LLMs that are out there in the market, whether it's LLMs from OpenAI or Anthropic or Google or Meta.

And all, so many elements from open source. We work with all of them. And our model for what we do is we are going to use all the capabilities that these LLMs provide to us, and we actually bring those capabilities to our customers.

But we'll also be ensuring that we're building a very deep technology stack on top of that platform. And as the LLMs advance, as they actually add some of those capabilities that we built ourselves, that actually throw our, we have to throw what we built that the LLM providers can do already for us and actually keep going up the value chain. And that's the model that we choose.

Figure out your space and maximally use the innovation that's happening in the industry, but then build a significant layer on top of that so that you can make that technology accessible to your customers.

Michael Krigsman: Given the importance of the foundation models to your business, how do you manage the fact that AI is evolving at such a rapid pace and how do you balance the stability of your business and product direction while these underlying capabilities are just shifting all the time?

Arvind Jain: We have no choice. You have to take advantage of the rapid innovation that is happening in the industry, otherwise you're going to be left behind and you have to fundamentally change your execution model.

We have to, when we started Glean, we very understood at a very, at a fundamental level what technologies we had available from open source and cloud and then you get to build your roadmap and you build a one-year roadmap. And that's not no longer how things work today.

Now technology changes on a monthly basis. And we had to fundamentally change your architecture. One of the things we change as an example, because we don't have a annual or a quarterly planning process, we switch to a monthly planning process in terms of how we are going to actually build our technology.

'Cause every month there's new things to look at and you have to quickly adapt. The other thing that we also added is this concept of figure out what you're going to throw every month in technology that you've built. This is a new fundamental way of building tech startups, is that you would become obsolete if you actually hold on to technology that you build for multiple years because all of that technology that you built two years back is most likely obsolete at this moment.

And you have to constantly think of this execution model where not only are you thinking about new things to build, you're also very actively thinking about things that you need to actually throw away and actually leverage the innovation from the industry to replace that.

Michael Krigsman: We have another question from Twitter, from Arsalan Khan again, who says, since most AI and data is created in English or Chinese, do you think startups should also focus on non-English or Chinese? And does this create a digital divide?

Arvind Jain: First, the content is created in many different languages. Yes, English is dominant in some ways, but there is plenty of content, plenty of systems, in different languages and I think what AI makes it possible today is it actually helps you build products that work, that are, that are global nature.

It's much easier today to build a product that you can actually bring to customers in all parts of the world. You can localize your products much easier with AI. You can make it work in Japanese and in Korean and Hindi and all the different languages at a much faster pace.

This is one of the core capability of AI. But then in terms of biases and the digital divide, it is true that the, and it's been true forever, even on the internet, even pre-AI, when you go and Google and research for information, there's always that bias that creeps in because English dominates as the source of knowledge in the world, right?

And that, that's a good one. That's a problem. I don't have a good answer on how to resolve that. As AI becomes more and more capable, how do you make sure that it's taking everybody's point of view, taking all the knowledge that's out there?

I think one thing I should add is that we have more capability today with AI to process even non-digitized content in a much, much easier way. Hopefully, there's some silver lining on that side: You can actually use data, knowledge, and content in different languages more than ever before with AI.

Michael Krigsman: Let's talk about startups now, from the perspective of enterprise buyers, which is a very important part of course of any startup's lifecycle. If you're an enterprise startup, do you have advice or a framework that enterprise buyers can use for evaluating AI startups, especially given the fact that, as you pointed out, every startup these days is using AI and the hype is so intense? Very often it's hard to sort through the claims to find out what's real and what's not.

And I'll just mention one, one thing here that I remember traditional software companies, ERP vendors and other enterprise software companies. And I have to say this situation, 20 years ago was no different from that standpoint. Now, software companies make outland just claims. And what should IT buyers and line of business buyers do about it?

Arvind Jain: This is one of the toughest problems for an enterprise buyer today. I think the AI industry has done a bit of disservice making bold claims, but then not being able to follow through. It's very easy to create really amazing demos and visuals for what AI can do for you. And enterprise buyers have, they've realized that as they try these products out, that the claims often are, were much larger than the reality.

I would say maybe let me take a step backwards and then first talk about what should we even do? I think there should be a plan for how you're going to roll out AI inside your enterprise. And I feel centralizing that, I'm having a core AI strategy for enterprise, first of all is the right way to start.

Think about all the things you'd like to do this year with AI. What are the different areas that you would like to actually see AI make an impact? You can pick a few departments, you can say that for our engineering teams, for our customer service team, like these are the top two or three priorities, where we want AI to actually make an impact.

First, build that roadmap and build that together in your enterprise. The CIO or if you've just, if you've chosen to have a chief AI officer type of role in your enterprise, let them, give them the charter of making sure they're working with all the different functional teams and you come up with a desired roadmap for AI for you for this year.

You start there. Now in terms of the number of vendors, it's not just Michael as you said, it's not just the startups, it's actually every existing software company is also an AI company. They all have AI things to sell to you. And I think you have to make some decisions there.

What's the right strategy for you? We feel that no com you have to be, you have to control like how many AI products you going to actually bring in? It's very hard to evaluate. It's not easy, by the way it's just the time that you have to spend.

Trying to evaluate every single AI product is enormous. You like setting those systems up, getting them up and running in your environment, then testing them and you companies have gone through this exercise where they spent six months, and just to find out that, that this thing didn't work.

One of the right strategies, two, my suggestions I have two pieces of advice: number one, like basically work with fewer number, don't have too many POCs start with a few so that you can do justice to them. And second instead of relying on demos and getting excited by a presentation, you can have a safer strategy, which is go and for every vendor that you work with, see if they have proof points, go and make them, connect you with customers that they were able to create success for.

I think that is it's a lot more helpful, instead of evaluating like actually have those conversations with your peers in the industry and see where they're achieving success. If you tried four things, one thing that's successful, share that story with others and they will share with you.

And that's probably the way to scale. And maybe this is a plug for Glean, but the way we think about AI is that fundamentally all AI in your enterprise is about making, working with some data that's in your enterprise, using the reasoning and intelligence powers of the language models, and then after that, doing some work, which you're again going to save.

In your enterprise systems, that work is going to get recorded and saved in enterprise systems. Fundamentally when you think about all AI use cases, they're about working with your enterprise information and then, making and applying AI on it to do some, to make some magic happen.

And we chose a different strategy with Glean, which is that why don't we actually build a system that's connected to all of our enterprise data? That's what Glean is. And now we are giving you this platform where you can go and build like many of these agents, many of these applications yourself.

With this horizontal strategy, you can do a much better job of security governance and not have to buy many different products. It's more cost-effective and allows you to get more value through one product.

Michael Krigsman: Alright, we're almost out of time. Arvind, I'm going to ask you a bunch of questions, some from me, I'm from the audience listening, and I'll ask you to respond very quickly from an enterprise perspective, the build versus buy decision. How should enterprises make that choice? Very quickly, please.

Arvind Jain: It's a build plus buy. You have to make sure that you get as much turnkey technology as you can, but you have to also remember that it's not going to be enough to add true value. You will have to build on top of those systems.

Michael Krigsman: From Twitter. A big hurdle for AI and LLM platforms is that they are not integrated with workflows. How do you see the market evolving?

Arvind Jain: That's exactly right. And I think the you have to build that layer. One of the key techniques for that has been RAG so you take all of enterprise data, knowledge systems, workflows, and you actually build the, this, retrieval system and an action system on top of all of the enterprise data and systems. And then you connect it with AI that's exactly the problem that we solve between.

Michael Krigsman: Another one from Twitter who, the CTO, CFO, COO, Chief Digital Officer, who creates AI strategy and should the Chief AI officer report directly to the CEO?

Arvind Jain: That's a good idea. If you actually have a Chief AI officer, having them report to the CIO or the CEO could be a good strategy because this is a company-wide effort.

If you don't associate it with a function, that's actually better in my opinion. And, but otherwise I think it's not, don't be rigid like whoever is motivated, who's excited and and has the capability, let them drive the efforts initially.

And like you can always figure out how to reorganize later in a more scalable way.

Michael Krigsman: Here we have another question from Twitter, and this is about pain. If AI startups are getting more lean, are older AI startups at a disadvantage? And as these older startups become more lean because they are more efficient using AI, what is the implication for employment and their workforce?

Arvind Jain: On the first question, there is indeed a first mover disadvantage in AI because when technology changes so fast and you have some of that legacy code base in your systems, it makes, it makes you a little bit less agile. That's part of you know how it always is as a new startup you always have the agility advantage.

And as a startup that's a little bit older, work hard, work hard on modernizing your systems, your stack so that you can, you don't fall behind for those reasons. And then in terms of employment, I think the I'll just say one thing.

I have not seen people losing employment because of AI. We work with so many large enterprises, we are bringing so much automation and efficiency for them, but every enterprise that we work with, they actually are concerned with both bottom line and top line.

And when they get some wins, nobody's giving up their team members they're not actually, they're just actually thinking about I can do more. And companies are actually building products at a faster pace. They're getting more things done.

I'm not super worried I think as like I will come back to the point that just stay relevant. That's the thing that matters for you as an individual. If you learn how to use AI, you know you'll have no problems in the future.

Michael Krigsman: You are advocating maintaining intellectual curiosity. I'm not trying to put words in your mouth because as AI drives efficiency gains, if you're intellectually current on what's going on, you can adapt and work around the changes that are taking place.

Arvind Jain: Yeah, and companies are really hungry. They're really hungry for AI experts, for AI talent. Become one, I think, it's going to be good for you.

Michael Krigsman: What advice do you have for entrepreneurs considering starting an AI startup?

Arvind Jain: Number one conviction. Stay, stick with your idea. Don't give up. And number two if you are, if you want to succeed be ready to work very hard. And and and then last thing I would say is that, a company is all about people that build it with. Focus on like having, having a great founding team and spend a lot of time, trying to get the right people that employees, in your organization. And when you get the right people they will do the, they'll build great products, they'll make you succeed.

Michael Krigsman: Final thoughts or advice for CIOs who are making AI investment decisions very quickly, please.

Arvind Jain: Focus on security focus on the centralizing, like your AI software stack. And I think the last thing is, that I would say is that try to get small wins. This is my advice. Don't like, don't like create projects where, if it doesn't deliver 50% success, it's a failure. Force every team, every function in your team, in your organization, to pick one or two wins for AI every quarter and see how that goes.

Michael Krigsman: Alright, and with that, we are out of time. A huge thank you to Arvind Jain from Glean Arvind. Thank you so much for taking your time to be with us today.

Arvind Jain: Thank you so much. It was a lot of fun.

Michael Krigsman: And thanks to everybody who watched, and especially you folks who asked amazing questions. You guys are truly awesome. Before you go, subscribe to our newsletter, join our community, go to CXOTalk.com, subscribe to the newsletter, check it out, and we'll see you again next time. Everybody. We have awesome shows coming up, talking about AI. See you later.