AI agent

“The synthetic human, the best thing which synthetic humans can do, which the AI agents cannot do is they can fail. They can fail at doing a task, which is a reflection of how a real human would fail in using a product. An AI agent will always figure out, or th…”

“Those metrics are not going to change because all the business drivers are still the same, right? It's only the enablers and the variables to do that is changing. So we should still focus on the actual value.”

“SaaS is becoming more of an execution layer, not an intelligent layer. Now the intelligent layer is moving more towards the AI agents, where it is more autonomous in nature. Where, ah, where, when employees interact with it, they don't interact with the SaaS a…”

“In a world where the cost of the activity comes to close to zero, because an agent is doing the activity, the infrastructure to enable the activity will all have to change. So I think that's what's underway today. And that is going to cause us to reinvent the …”

“Agentifying workflows are getting very simple. So the workflow is not, not a mode. Being embedded in a workflow is not a mode because switching costs are getting lower.”

“So we have built an AI agent also. So the AI agent works in two ways. One, it can speak with consumer AI agents. So let's say you are using an AI agent and you trust the AI agent to make all the bookings. It can actually go and talk to our hotel's AI agent and…”

“I strongly believe AI is going to gonna have a huge impact on how travel is purchased , ah, especially with consumer AI agents, and a lot of business models will get disrupted.”

“Plus, I think the future of agents would be that agents will create their own agents on demand, which means, you know you know, if an agent thinks that we need some tasks to be done, it'll, it'll think about Thinking and they'll probably create their own.”

“Agent, agent with the powers to execute is autonomous. Which means, ah, as an example, just to complete the story, is that if I want to say I want to book a ticket, an agent is deciding on my behalf which seat to pick up, then that's autonomous. But if the age…”

“Because the number of SaaS apps will go down. The number of agents will go up.”

“To get to that, and that might not make sense for everything, but there are many situations where that makes sense. To get to that nirvana, we need context graphs.”

“Because context is a function of pairwise relationships. When I ask you a question how things are going, your wife asks you the question how things are going, and your CFO asks you the question how things are going, you have very different answers to the same …”

“How can you have a agent running at scale if you don't have the evals or AI observability in place, like otherwise, You are totally flying blind. And especially when the system is non-deterministic, which can do, you know, you have no idea whether it's perform…”

“Observability you need it when you have something in production, your agent is in production, and enterprises, like the large enterprises where the agents are in production more than, you know, like the, if you look at the Silicon Valley companies, they all ar…”

“I mean, look, I'll tell you, Europe is pretty slow. Siddhartha Ahluwalia: Yeah. Sudheesh Nair: But the US, absolutely.”

“So, you know, for example, support is, is, is an interesting vertical, right? Which is a lot of stuff happening within support where, you know, the ticket deflection and automatic resolution of tickets, these are things that whether somebody builds it themselv…”

“I think that is very early days. I don't think anybody has really tackled this holistically yet, but this is I think important.”

“Every, every use case in security also is getting, you know, rethought, like, how can I build a, you know maybe an agent, right, that can do this automatically versus relying on how it was being done before. So every workflow is being, being reinvented and eve…”

“You can define those as deterministic functions, you know, implemented in code, or you can use LLM as judges, but then you can evaluate like, how did each span perform? And that can give you a fairly good way to zero in on problematic areas of your agents.”

“I think the really interesting inflection we've seen over the last three or four months is that the models have gotten good enough that agents can actually run the entire workflow, right? They're not additive to individuals. They are they're the ones that own …”

“AI agents and giving them rich data to feed off of that's high quality is going to be, is starting to be more of a mainstream problem, but will be a mainstream problem two years from now.”

“At some point, I bet you, 12 months, 18 months, two years from now, the customer will come and say, Shashank, this is very dumb. I'm unhappy. Why? Because you've already done the whole resolution plan. Why do I have to come in and manually hit approve? Just do…”

“So at some point, you're going to start seeing front page headlines in the Wall Street Journal saying, this healthcare company or insurance company rejected one million claims, which it shouldn't have because the AI agent went rogue, or this credit card compan…”

“Basically the agents create V-one of the work. Humans are verifying and iterating. Humans never create V-one.”

“I think it's the first time in human history that AI has become the face of the product. A user is directly talking to an AI and working with an AI. So every output of an AI system is mission critical. And with agents doing actions and making critical decision…”

“I think if you don't have a strong technical founder on the team, it's very difficult to build these because as you know, pushing these agents into production is, is far different than building a POC. The non-technical person can build the POC and get it to lo…”

“Now everyone is 2.2. We've been on the journey for a bit now. Right. So everybody is now observe 2.2. We took the older product, right. That we built completely replatformized it. It's now all agents and whatever new builds all agents anyway, sellers are all a…”

“Number two is I think the concept of agents where you don't really need A clean API per se. You can make do with unstructured data and take inputs and shoot out outputs that are processable, human readable, but also machine readable. And that to me is going to…”

“This is the context that any agent, actually human or AI agent needs to be able to solve this problem for the customer. Knowledge, expertise, and norms. So if you had to build this in a business, you need to be able to bring together first the map of the busin…”

“Actually I, I'd say 50% of our end users are now agents and the rest 50% are humans.”

“Investing in the Orchestration layer or the model layer, like that's the easier problem to solve. Building an agent is easy. How do you actually make it accurate? How do you actually manage your context? That's going to be the hard problem.”