1Mby1M Virtual Accelerator AI Investor Forum: With Yash Hemaraj, General Partner at BGV (Part 3)

Sramana Mitra: Lyzer sounds very interesting. So how does this company go to market? It sounds like it’s a horizontal technology. So presumably selling to the CIO.

Yash Hemaraj: CIO or business leaders. They have a unique approach – they build pre-existing agents for certain domains where the data architecture and workflow are well understood. This enables business users to start using it as they have enterprise-grade features which integrate within the existing system.

Sramana Mitra: They have a platform, but they have themselves built applications on top of that and selling to the business users for those applications. Is that right?

Yash Hemaraj: Yes. If you want to build an agent, you need to have a mixture of expert agents underneath that agent. For example, let’s take venture capital itself. If you want to build an investment thesis for a particular area, you need to have sub-agents – one which looks at competition, one that looks at time, one that looks at financials, and one that builds the investment memo. They all need to interact with each other. So the framework is in this mixture of multiple agents.

However, in order for enterprises to catalyze their activities, they have pre-built agents in multiple domains. They can take one of those and start giving inputs to make things work. That is where the LLMs have some amazing capabilities to break down the procedures in your existing standard operating procedure documents into steps that could be executed. So a real enterprise system will have a combination of all of these technologies. That’s how Lyzer is addressing this in the agentic software space.

Sramana Mitra: So in the case of Lyzer, where in the enterprise are they getting initial traction?

Yash Hemaraj: Business users.

Sramana Mitra: What business users, like what firm?

Yash Hemaraj: For example, they already have agents for marketing operations. So, when marketing teams go to market, they’re not just approaching marketing ops saying, ‘Hey, we have this AI technology that you need to use.’ They’ve already researched the process and can say, ‘Okay, here are the steps you’re currently doing, and here’s an agent that can automate five out of the eight tasks you’re already handling.’

Similarly, they have an SDR (Sales Development Representative) revenue agent called Jaszon. When you enter the revenue functions—like the Chief Revenue Officer or business operations—you can immediately start utilizing this agentic software that automates hundreds of tasks that SDRs and BDRs (Business Development Representatives) typically handle. That’s how they go to market.

In any enterprise, there are teams focused on data quality, data integration, and other areas. Let’s take an example from revenue operations. You need to integrate external data, such as information on a public company if they are your prospective customer. This includes S-1 filings, earnings calls, and relevant news events. You also need to integrate your internal data—how big is this account? Who is involved in the buyer’s committee? What are the previous engagements with this account? That’s the customer context.

Additionally, internal exchanges between an Account Executive (AE) and an SDR might involve unstructured conversations, such as Slack messages or emails discussing high intent levels from the client. To realize real enterprise benefits, you need to incorporate all this data while ensuring the right privacy, protection, and governance layers are in place.

This is where tools like DAX provide role-based access control for data management. Companies like Stack Identity help with identity governance, ensuring that each agent accesses only the information they’re authorized to. As this process unfolds, other departments get involved. The CFO gets involved in data governance, the CIO gets involved in data architecture, especially for large enterprise sales.

You start engaging with decision-makers, and the architecture of Lyzer is built to scale into other use cases within the enterprise. You may begin by working with business users, but as you move forward, you’ll engage other stakeholders. To succeed in this multi-stakeholder, committee-based sales process, you need a solid architecture that addresses the concerns of each group.

If you don’t do this, you might get interest, but conversion rates will be low. This is true for many platforms. When we conduct evaluations, we look for founders with deep domain expertise who understand architecture and the challenges they’ll face. We look for innovation—what have they done to overcome the hurdles they will face? Because enterprise sales is increasingly multi-stakeholder, you need mechanisms in place to address all those perspectives.

This segment is part 3 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: With Yash Hemaraj, General Partner at BGV
1 2 3 4 5 6