Person
2 episodes · 41 quoted claims
Praveen Akkiraju cites a Goldman Sachs study saying enterprises are using a year's worth of AI budget in three months.
“enterprises are blowing through their token budgets a a year's worth of AI budget in three months. This is notion of token maxing”
Praveen Akkiraju describes the end of 2024 as a seminal moment in model evolution, when a model introduced the concept of reasoning.
“the end of 2024 was, like, one of the most seminal moments in the e- in the evolution of models, right?”
Praveen Akkiraju says open source frameworks like OpenClaw helped show in early 2026 that models can do general work, not only coding.
“open source frameworks like Open Claw to be able to use models to actually show us that work can be done.”
Praveen Akkiraju describes large language models as stateless machines that compress the world's information into weights and do not remember prior work.
“you think of large language models as these kind of stateless, um, stateless machines, right? In the sense that they absorb all the information you give them”
Praveen Akkiraju cites Claude Code as an application with a well-defined harness of tools, memory and coding guardrails.
“an application like, um, you know, cloud code, right, has a very well-defined harness that gives us the gives it the appropriate tools”
Praveen Akkiraju says sophisticated enterprises have deployed over a thousand agents in production for various tasks.
“sophisticated enterprises today, uh, have deployed over a thousand agents in production for various different tasks, right?”
Praveen Akkiraju calls 2025 the year of agents because models advanced, including computer use and browser use agents.
“2025 was sort of the the the year of agents, and rightly so, because, uh, the models advanced.”
Praveen Akkiraju says coding agents are estimated to generate about 3 billion dollars in revenue.
“it's estimated about 3 billion in revenue right now from just coding agents”
Praveen Akkiraju says customer support agents generate close to 500 million dollars in revenue.
“customer support agents, I think, is close to about 500 million, uh, you know, in terms of revenue generated just for that for that sort of use case.”
Praveen Akkiraju says OpenClaw was the first open source personal agent framework that worked, because it could understand and remember work and feed context back to the model.
“open claw was the first kind of open source framework that actually started to work because, uh, it was able to bring, um, the the appropriate sort of ability to understand the work”
Praveen Akkiraju says developers use temporary sandboxes to run and check agent-written code against guardrails before pushing it to production.
“developers are spinning up these sort of ephemeral sandboxes to first kind of make sure that the the code is that the AI agent has written is sandboxed, executed”
Praveen Akkiraju says E2B, a company in the Insight Partners portfolio, helps developers add observability to agent execution through sandboxes.
“we have a company in this space called E two B that, uh, that that's been, you know, tremendously successful, kind of helping developers with this”
Praveen Akkiraju says that in multi-agent systems one agent's error can feed into another agent and compound.
“error is compound. So, if you have, uh, uh, a large language model, you know, you have a multi-agent architecture, and an agent makes an error and feeds it into another agent”
Praveen Akkiraju says Stampli, an Insight Partners portfolio company, makes accounts payable and procure-to-pay software.
“Stampli, which is one of our companies, uh, they do, uh, accounts payable that are procured to pay company, right?”
Praveen Akkiraju says Stampli's agentic step does invoice reconciliation and then a human approves, removing days of work.
“The AI does it for you, and then the human approves it.”
Praveen Akkiraju says Mozilla found 271 bugs in the latest Firefox using the model he calls MITOS, with only about three being high-severity CVEs.
“They found 271 bugs, uh, using MITOS, which is kind of amazing.”
Praveen Akkiraju says Tesla rebuilt its ERP systems in a previous cycle because existing products did not fit its vertically integrated car production.
“in a previous cycle, you know, companies like Tesla essentially rebuilt their ERPs because they felt like, um, the existing ERPs did not fit the kind of model”
Praveen Akkiraju says Claude Cowork reflected Anthropic's realization that its coding harness could also do non-coding work.
“Claude co-work essentially was a realization in Anthropic, uh, as well as other labs as well, that, um, hey, this coding harness that we built to write code can actually do work”
Praveen Akkiraju says enterprise budgets are shifting from per-seat SaaS pricing toward agentic tools, partly as excess seats are reconciled against usage.
“we clearly are seeing budget shifting in in enterprises, right, from traditional SaaS, um, you know, per seat pricing to, uh, to more sort of agentic tools.”
Praveen Akkiraju says enterprises lean toward buying standardized front-end agent tools such as customer support.
“a lot of, like, front end customer facing tools where the workflow is perhaps a little bit more standardized, right? Um, you, you are, y- you know, um, enterprises are leaning towards a buy kind of decision.”
Praveen Akkiraju says models from Opus 4.5 to 4.7 and GPT 5.5 are becoming more efficient in token use.
“they are getting much, much more efficient in terms of token use, and I think that's important.”
Praveen Akkiraju says Insight Partners' pricing team has considered pricing agents as a fractional full-time employee, though this is still experimental.
“perhaps you think of this as a fractional FTE right um or a head fractional headcount”
Praveen Akkiraju says OpenAI had just launched Operator, a consumer-focused AI agent.
“Uh, OpenAI just launched, uh, their consumer agent called Operator, uh, very interesting.”
Praveen Akkiraju says Claude Sonnet 3.5 was a step-function jump in how accurately large language models generate code.
“the Claude Sonnet 3.5, that was, like, I think, a step function jump in the ability of large language models to accurately generate code”
Praveen Akkiraju says Salesforce launched its own agentic offering, called Agentforce.
“Salesforce, uh, la you know, launch their own sort of, uh, agentic, um, agentic workforce, agent force, as they call it”
Praveen Akkiraju says Microsoft offers Copilots and that OpenAI launched the consumer-oriented Operator agent.
“Microsoft has Copilots. We just saw OpenAI launch operator, which is more consumer-oriented, uh, agent”
Praveen Akkiraju says Insight Partners is an investor in CrewAI, a popular open-source framework for building AI agents.
“you know, we are investors in a company called CrewAI, for example. Um, and I'm sure anybody who's talked about AI agents probably knows about CrewAI.”
Praveen Akkiraju names LangChain as another framework used to build AI agents, alongside CrewAI.
“There are others, like LangChain, uh, which are also do something similar.”
Insight Partners' market map includes a category called the agent computer interface, covering the tools that let AI agents use computers.
“Uh, there's this new sort of bucket we, uh, you know, we call sort of the agent computer interface, uh, which is the ability for AI agents to use, uh, computer”
Praveen Akkiraju says Cursor seems by far the most popular AI coding tool among developers.
“I mean, Cursor, by far, seems to be the most popular among developers today”
Praveen Akkiraju says Devin works differently from Cursor by firing off a bunch of agents that execute the user's plan.
“If you take something like Devin has a different sort of way it it engages, you know, it fires off a bunch of agents that, you know, execute your plan”
Praveen Akkiraju says Samaya is building AI agents for finance workflows, including a finance-analyst-style agent.
“Um, there's this really interesting company called Samaya, for example, that's building doing some amazing work, uh, focusing on building agents for the finance workflow”
Praveen Akkiraju describes a reflection loop in which an agent's output is tested against another language model trained on the correct output.
“once the output of the agent comes out, you test it against another, uh, large language model, which is trained with the right output”
Praveen Akkiraju says large language models are trained heavily on math and coding because skills there transfer to broader reasoning tasks.
“there's a reason why a lot of large language models are fundamentally trained on math and on coding tasks, because there's a lot of transfer learning.”
Praveen Akkiraju says what changed in this generation of AI is that powerful models can be accessed through an API call.
“What changed, I think, in this generation is that today, really powerful, complex, uh, AI models are available on the other end of an API call, right?”
Praveen Akkiraju credits the release of ChatGPT, because it was so accessible, with unleashing corporate adoption of large language models.
“And it's largely, I think, the ChatGPT moment, um, which unleashed this because, you know, it was so accessible.”
Praveen Akkiraju says most large corporations have rolled out some form of coding copilot, though not yet at scale.
“Most of them have rolled out, for example, uh, some form of a coding copilot.”
Praveen Akkiraju says agent pricing may combine a platform fee with a charge based on how often the agent is run.
“So you might, you know, pay a platform fee, like for the agentic thing, and then maybe you hire an agent.”
Praveen Akkiraju predicts outcome-based pricing will become the norm where the value of an agent's output can be attributed accurately.
“So, based on that, if you can, then outcome-based pricing is is essentially where we're eventually going to go to, right?”
Praveen Akkiraju describes a security operations center as the nerve center of how most companies run security.
“uh, a security ops center is essentially the nerve center of how most companies run their security operations.”