B (3:53)

Amazing. Thank you so much. And I appreciate that coming from an engineering perspective, you obviously, as you mentioned, engineers do some amazing creative things, right? But as you said, they could be lazy, so they want to automate a lot of things. And, and I agree engineers typically look for problems that can be solved by machines. But what I really took away from this is finding where people can excel in the areas whether they're passionate about it or not. But again, building relationships is one I think that humans really connect with and will always do better than machines. If you can eliminate those monotonous tasks which typically drain people of their creative energy, I think you're onto something. And I think we can totally leverage this agentic era that we're in now. So thank you.

A (4:35)

Fahad.

B (4:35)

Noodle Seed operates at the intersection of automation and intelligence. What are some common misconceptions businesses have when it comes to implementing AI powered agents?

A (4:45)

Some of the common misconceptions I would say, is that people believe that you can Simply take an AI like ChatGPT, just paste some question into it and it'll give you the answer. In reality, it's a lot more nuanced and complex. A simple question to answer. Chatbot is great for consumer use cases, but for business processes, you typically need to map out each and every step in the business process and then look at each individual step, whether it can or should be automated using large language model or any intelligent software versus can it be done deterministically with either code or simple rules? And the overall end to end system, what we call agentic system, is going to be composed of some individual steps being performed by AI and then many Steps along the business process will still be traditional rule based or code based executing processes. So that's one of the common misconception that I find. And people kind of take AI because it's a buzzword in industry. They sort of think that you're just going to sprinkle some AI on top of a business and all of a sudden it's all profitable and all singing dancing solutions. It's not quite as simple as that. I do believe that there is a lot of room for people to apply their creativity using these newly available tools, but still be able to automate a lot of stuff and understand your customers business processes a lot more deeply along the way. You might even find opportunities where you eliminate entire processes now because you have these intelligent agents rather than having to send files to different people or data to different people. So that's one of the things that drives Noodle C and that's how we go and engage with our customers. We try and find opportunities to sometimes even eliminate processes rather than automating them and using the intelligence with where it's actually useful. Rather than applying it end to end, we break it down and only apply it at the steps where it makes sense.

B (6:56)

Thank you so much, I appreciate that. And different perspective obviously on that. You're right. Talk to a lot of people here on the show and outside the show that have that misconception that they can simply say, oh gosh, ChatGPT is so awesome. You can give it a prompt and it'll solve all the world's problems. It's just not that simple. You and I both know being in business a long time and leading technology teams, business processes are a lot more complex than that than than just to wave that magic wand and hope that generative AI will fix everything. There's a lot to unpack in a business process and to do business automation. A lot goes into that. One thing I highlighted is sometimes you need to eliminate a process in some cases versus just automating a process. I think that was very insightful. So thank you. Fahad, the third question I have for you. How do you define agents as a service and what are some real world use cases that businesses can start using today or in the near future?

A (7:47)

So one of the things that comes with every new paradigm shift is that you have to go back and rethink some of the fundamentals. Something that a lot of business leaders out there are saying is that traditional software as a service is dead. I believe to some extent that is true because the traditional software as a service model that charged businesses or their customers on a per user basis may not apply in the case of agents or AI agents. In the age of agents, a single business user can control an army of AI agents doing millions of tasks all at once. So you're not bottlenecked on the number of business users. Now, how do you price such a solution to your customers? Typically when I go and talk to customers, what they're interested in is the outcome, not the process, not the number of tokens, not the compute, not the gigabytes or whatever internal technical metrics that we have for our, our cost to the solution provider, but the outcome that they are going to receive. So as an example, let's say if we build an agent that can do one task and save a customer $100, we would like to be able to price it at 10 to $20 so that we can share in the benefits of the savings that we're providing to our customers. And that task is being done by an agent, not a business user or not like a subscriber model. So when we go and say agents as a service, that really just means that AI will do the work on your behalf, help you achieve an outcome, and we will share in the benefits of that outcome rather than just charge you on the number of users or tokens or whatever else that the cost metrics. And then those kind of business models, they typically don't care about the outcome your customers achieve. They just care about the number of users or seats that they can sell because they can sell 100 seats. While a customer achieves no outcome, that's no good for anyone. It's good for the vendor, but not good for the customer. In case of agents, it's more aligned with the outcomes a customer achieves.

B (10:05)

Thank you, that's very helpful. At the end of the day, you want a long term relationship with the customer. So I like how you broke that down. You know, you started out saying during every paradigm shift, you need to rethink these fundamentals. And it's absolutely true. Especially in this age of AI, you know, customers are interested in the outcome or the solution, not how many agents are being used or how it's being performed. I like your model agents as a service where you're both mutually sharing the benefits of the outcome. And it's not about charging the customer for, like you said, a hundred agents or whatever that is just so it's one sided. I really like the partnership and the vision that you share in your model. Fahad, the last question of the day, what are some of the biggest pain points you're seeing in enterprise AI adoption today, and how is Noodle Seed specifically solving for those?

A (10:52)

So some of the biggest pain points, and I'm sure these are different for different industries and companies and markets, but just in the markets where we operate, the customers that we interact with, some of the biggest challenges are really just the adoption of AI by the leadership of the company. I read in one of the reports that a very good indicator for how well an organization adapts AI is whether the chief executive is using ChatGPT or not. And this is, you can say it could be a correlation rather than a causation, but when the leadership team is aware of some of the fundamental level capabilities of AI, then they will push the next layer in an organization and they will push the next layer in the organization to start using more and more of these tools to improve the efficiency of a business. When you look at Bottoms up, some of the challenges are people have these checklists that are updated checklists, requirements that every single technology or solution that they adopt has to fulfill whenever there's a new kind of technology. A very good example could be, let's say you're moving from horse carriages to automobiles. Now, if you have a checklist of when you procure a horse or a horse carriage, that checklist is just not going to apply to an automobile. Same thing applies when you're moving from deterministic or rule based systems to AI based systems. You have to have different criteria by which you measure the suitability and the benefits of such systems. So as an example, an AI system is statistical, it's not deterministic. So if you have a checklist of whether it answers the exact correct value every single time as a percentage, that may not be a good metric for you to adopt AI. What might be a better metric is how do you adapt when it fails and where do you apply it in terms of the different levels of impact that a decision? Let's say an AI can make a decision automatically and once in a while it gets it wrong and costs 10 extra. Maybe that's a good use case for adopting AI rather than just making sure it's correct 100% of the time versus let's say if there's a life and death decision or a financial decision that can impact someone's life there. You don't want it to get it wrong once every while and ruin someone's life. So how you apply it, in what business context should drive the criteria with which enterprises adopt AI solutions into their business processes?

B (13:50)

Thank you really appreciate that. Back to the question, you know, what are the biggest pain points you see? And in your space, what you're seeing is that adoption by the leadership within the organization, which I absolutely agree with you. Luckily, I've got some people that are very somewhat tech savvy in the leadership team that are embracing this, which is amazing. And believe it or not, CEO is doing just that, using Copilot and ChatGPT in his everyday life. I like how you highlighted it. It's important that you ensure you're applying the correct criteria to a business process, especially when leveraging AI. And again, it's how it's applied in the business context. So I thought that was very insightful for us and our audience. Fahad, it was such a pleasure having you on the show today and I look forward to speaking with you real soon.

A (14:33)

Thank you very much. Pleasure to be here.

B (14:35)

Bye for now.