theCUBE + NYSE Wired: AI for CFOs & AI Leaders in Silicon Valley | Gaurav Nemade, Inventive.ai

Clips
More from theCUBE + NYSE Wired: AI for CFOs & AI Leaders in Silicon Valley

Gaurav Nemade

Co-Founder

inventive.ai

play_circle_outline Focus on AI agents for sales teams and RFP workflow automation

play_circle_outline Importance of data quality and integration in building AI agents

play_circle_outline Comparison with Salesforce Agentforce and differentiation in approach and strategy

play_circle_outline Mastering RFP Responses: Engaging with Platform and Onboarding Process for Efficiency and Success

Info
Transcript

Gaurav Nemade, Inventive.ai

Gaurav Nemade

Co-Founder inventive.ai

Inventive AI is a company focused on building AI agents for sales teams, specifically targeting the RFP workflow to reduce time and effort by 60-70%. With over a dozen employees and part of Y Combinator's summer batch, they aim to democratize technology and focus on personalization and contextualization in sales workflows. They differentiate themselves from Salesforce Agentforce by building power drills for specific workflows and prioritizing deep integration and domain expertise over generic solutions.

Their core bets for the future are the complexi... Read more

explore Keep Exploring

What is the core focus of Inventive AI and how are they helping sales teams with their AI agents? add

What is the number one thing that helps agents create a knowledge hub and integrate important information for RFP workflows? add

What is the difference between Agentforce and the strategy of the company? add

What is the process like for getting started with the platform? add

bolt Powered by CUBE AI

Gaurav Nemade, Inventive.ai

search

>> Welcome back everyone to theCUBE's coverage special presentation here in Palo Alto Studio. This is where we have our West Coast point of presence. Of course, we have the NYSE show floor with theCUBE East and we bring all that open source content to you. Here, we're talking about the AI innovators in Silicon Valley, the subnet here in the valley where all the technology leaders are. And of course, we also were weaving in the AI for CFOs where business model transformation's happening. It's going to happen on a highly accelerated rate in the next two years specifically, and we've got a lot of great guests coming through, sharing their content with you. Gaurav Nemade is here, this co-founder of Inventive, another entrepreneur who's in the arena. Gaurav, thanks for coming on theCUBE. Appreciate your time and looking forward to chatting about your awesome opportunity around agents. Our favorite topic.

Gaurav Nemade

>> Yeah, excited to be here, John. Thanks a lot.

>> So obviously business model transformation teased up. That's for the CFO crossover, but the innovators in AI are leading the way, and CFOs and executives and businesses in the enterprise and startups, by the way too, are leveraging all this. So agents are going to play a real role with this data. And so there is no AI without data, making that data fast, making it low latency, addressable. You guys have a company that you started that's really in the sweet spot of where we see the next killer app, obviously search. We see RAG and search. Great, find what you're looking for. Vector embeds all good, but that's just the beginning. That's not going to go away, but the tasks become huge. So agents will be big. Talk about what you guys have built, are building what the venture is about, how you started, and we'll get into some of where you see agents going. So you have a great background with Google DeepMind and other areas. Let's explain what you guys do.

Gaurav Nemade

>> Yeah, for sure. Our company is called Inventive AI, and we are building AI agents for sales teams. So one of the core things that we have realized early on that agents are going to pick up really, really fast. And as you mentioned at my time at Google AI, what we realized the LLM technology is going to really accelerate over the next few years. So I quit Google early 2022 and started on this journey of figuring out how we can use agents to help the sales team. So the workflow that we are focusing on right now is helping the RFP teams in reducing their effort by 60, 70% in responding to these RFPs. And there are about $1.4 trillion worth of projects in the US alone that get allocated via RFP process and tendering process. So what we do is, we help them reduce their time and effort to respond to these 500 to 1000 page RFPs, and help them just save 70, 80% of their time overall on that.

>> Yeah, it's not your weekend that's blown. It's multiple weekends. How much time they spend on RFPs.

Gaurav Nemade

>> Right, yeah. And there are multiple people involved. It's not just a one person job. Usually you have teams of 20, 30 people responding to these RFPs. So we're very excited about that opportunity in terms of agents being able to automate some of those things. And we feel that we are just getting started with this whole agentic workflow. RFPs is just one workflow that we are going deep. We are going to be targeting additional workflows on the sales team and helping build a lot of power drills that can go and help them with these AI agents.

>> When you were at Google DeepMind and you just recently left a couple of years ago, a year and a half ago or so, I mean, it's pretty historic if you look at the whole pedigree of Google's role and all this. When you came out of that, the market was booming. People realized the shift. I mean, they didn't talk about agents and that. That was just ChatGPT was the big, "Oh my God, the user experience is changing." That then trained everybody like, "Hey, this is the new user experience." I call it the best thing in tech that ever happened because I can stop explaining what AI is now because people just see it. And we call it the browser moment. People use different metaphors, but that really changed. Since then as you go into this thing, is it a productivity angle for you, knowing what was before the moment and then now, what's your view on how things have changed, what advancements have progressed? What's been the big needle moving moments besides the user experience?

Gaurav Nemade

>> Yeah, I think you absolutely nailed that, right John? I think it was the user experience. The innovation was not actually ChatGPT. The innovation was that you were able to hook up a GPT 3.5 model into a user interface that millions of people across the world could use. Because we already had models that were capable like 3.5 or GPT 3 or 2 for a very long time, at least two, three years before that. But as soon as the world got access through their keyboards, I think it just blew up. So that was super exciting, generally, and I completely agree with you, I think it was the user experience that got us here. But one of the things that I'm super excited about from now on is this whole idea of multi-agent systems. And that's where we are entering in. Right now we just have single agents that can go do booking of your flight or do XYZ things. We are entering into this era where you will have maybe a 100 agents, I will have a 100 agents, and then they'll be a 1,000 agents out there and your agents will talk to my agents and get things done along the way. So we are excited about how we can actually enable that within the sales teams. Sales people can have their agents and they can go into the finance team, HR software and stuff like that and do all of that by themselves, so that's a huge opportunity.

>> Have my agent call your agent for the next CUBE gig. But I think the key is that delegation trust, resiliency around the quality of agent interactions, the data problem, right. And so with this RFP angle you guys have, is that an entry strategy because it's an easy use case? I mean, hey, RFPs, what's the vision beyond RFPs? Because I can see the collaboration also a time killer too, waiting for someone to respond. You own getting that piece of it. Take us through where this goes from here with your venture, because there's so much more productivity opportunities with sales workflows because you got customer touches other data. What's the vision?

Gaurav Nemade

>> So I think one of the core things that we are driving is once we have access to the data in the organizations, RFP is just one vertical workflow that we are solving for. One of our core thesis or core understanding in this space is that a lot of these hyperscalers companies like Salesforce, they're going to build generic things that can go across different use cases and solve for, let's say, top 30, 40, 50% of the use cases. But our learning as a startup, how we are going to win this space is that there are these last 50% of these workflows that require deep integration, and that requires for solving for deep things.

>> Like what?

Gaurav Nemade

>> For example, RFP is a very deep workflow. You can build a simple agent by hooking up the data, getting the RFP and responding to it, but it just doesn't stop there. You have to collaborate with 20 other people. You have to have personalization and contextualization as part of your RFP and so on. Similarly, sales enablement is a big deal in itself. There is so much personalization that you can have in terms of coaching your sales people or creating the collateral for the sales team that they go out and pitch to different clients. So there are these five or six verticals that we are eyeing that will go and build power drills for, like I was saying-

>> Pre-fabricated workflow. Stands up the response, you can edit it, that kind of thing.

Gaurav Nemade

>> Yeah, but we want to go much deeper into that. I think we want to go and solve... One of the things we strongly feel is that technology is democratized. The large language models agents, this is going to be democratized. The companies that are going to win are the ones that are going to understand the user workflows end to end. So we are in this to solve for six or seven or 10 different workflows in the next couple of years. Instead of doing, let's say, 200 use cases that a generic agentic platform can do.

>> So rather than horizontal, you do enough horizontal with the workflows and then you're going to get more of the domain expertise-

Gaurav Nemade

>> Yes....

>> in the user base. So I'm the person in charge of the data center. That might be an AppSec review, part of the RFM, I'm making this up, but I own this piece that's more comprehensive than just, "We support the cloud."

Gaurav Nemade

>> Correct, yeah.

>> Are your S3 buckets secure? What's your two factor? I mean, these are in the weeds, details. Do you guys plan on servicing all that stuff? Is that how it would work? I'd have my own agent and I got to bring my agent to the party, so to speak. So is that how you see it?

Gaurav Nemade

>> Yeah, so I think the way it's going to work is at the end of the day, there will be, we already have a couple of different agents within the RFP workflow. So we have a main agent that goes and figures out which of these agents it needs to work with, and then that agent takes that action. So I think one of the bottlenecks in this agent world right now is having good orchestrators. By orchestrators, I mean you have a parent agent and they have a multiple set of smaller agents, for example. So which agent needs to be called for the right job and in what order, so that you are not tampering the efficiency, so I think that orchestrator layer I think is missing right now. So we have built something in-house to handle some of those things. But over a period of time, I would be excited for the research community to get in that zone and build that strong orchestrator.

>> Yeah, I love the multi-agent thing. Definitely the sub-agents we call them, but I think that's happening. Our research has definitely showing that. How big are you guys right now? Talk about the scope of the venture. Where are you at and the progress? Give us some of the highlights.

Gaurav Nemade

>> Yeah, so we are a seed stage company. We're part of the Y Combinators last year's summer batch, and we are over a dozen people at this point of time. So yeah, growing steadily and yeah, we're seeing a lot of customer traction at this point of time, so super excited about that.

>> So talk about the data challenge with the venture, because a lot of customers sometimes don't even know what they have. So do you apply first wave of figuring out what's in the landscape of the customer and then rolling it out? Take us through the engagement of a customer. Say I mentioned, "Hey, I want to be a customer. Got me sold. I would want the agents, what do I do?"

Gaurav Nemade

>> So the number one thing is we help our agents create something called a knowledge hub, where we work with them to understand for this particular RFP workflow, what are the most important bits of information that you have that we can work with essentially? So the information can be anything in form of their previous RFPs that they might have done. A lot of times they just like re-circle content within RFPs, but we also go in and integrate with their systems from, let's say, SharePoint, Google drives, confluence, notions and so on. Our approach right now is that we want to keep the data as high quality as possible, so that they start seeing immediate returns. Because if there is all the data from the organization and the retrieval is slightly hard for the retrieval generation systems, then it becomes challenging. So our approach right now is that we just try to get as much high quality information as possible. And then along that we are also building integrations with all of their systems from Jira to Slack and stuff like that, and getting that information in the knowledge Hub, and set that up really, really well. So this is step one, and this is the most important step in the whole process. After they are done with that, I think our agents are pretty well set in terms of we can help you do competitor research, we can help you do generation getting into first draft and things like that. So then it becomes much easier. And of course we have some fine-tuned models, so we would take these, evaluate how it's working, and then tweak the quality along the way.

>> Talk about external integrations. Do you connect into Slack or... I mean, we have news people always use our news. Do you see Perplexity? That's a great engine. Do you see yourself peering with other sources or in that knowledge hub?

Gaurav Nemade

>> So we have customers who are asking for that right now. So far, what we have been focusing on is the enterprise data itself. We don't go outside of the-

>> So you're nailing the first use case, which is get that knowledge hub up and running.

Gaurav Nemade

>> Correct.

>> Highly reliable.

Gaurav Nemade

>> Right, yeah.

>> Accurate.

Gaurav Nemade

>> Yeah, that's the number one goal. And then what we are expanding into now, is we are looking at some of these public type of infrastructures where we can also get some public information in there. So if we don't have, let's say, a response to a section in the RFP, then can we go in and get some information from the-

>> Dun & Bradstreet or whatever, however.

Gaurav Nemade

>> Yeah.

>> All right. Talk about Salesforce because they just rebranded Agentforce. Obviously they're not known for the best user experience, but they have the system of record. They've always had that collection. They would remind me of Cisco in the early days. They would buy a bunch of companies, but they all didn't work together. They think they can bring it all together. Obviously Salesforce, sales people are in there, but I haven't met one salesperson that loves it, but their boss tells them they have to use it. They're trying to change that. How are you guys going to compete against Salesforce? Are they a partner? Are they a competitor? What are you guys doing different? Because a lot of people do have Salesforce, so that's where that data entry's in. So can I get an agent to integrate Salesforce or do I go with you guys? Talk about that dynamic.

Gaurav Nemade

>> Yeah, for sure. So Agentforce is definitely in the area of where we are operating, but I think our strategy is very different and the thesis we are operating with is very different. Number one is, we see that Salesforce Agentforce is more like a toolbox, versus we are building a power drill for specific workflows. They have a toolbox that can solve for let's say 40, 50% of sales use cases, but they cannot solve for the rest of those 40, 50% of the use cases that require a lot of workflow integrations, collaboration, and a bunch of other things. So that's that. We are building a power drill, let's say, for RFP workflows right now, and we'll expand the same analogy for other verticals, as I was telling earlier. That's one. The second is we've realized that by the way, we started by building a horizontal company first. One and a half years, two years ago when I left Google and we were looking at this space, we started building multiple agents and we realized quickly that it's a workflow game and not, we can't do this and compete with Hyperscalers or Salesforce and stuff like that. So that's why we decided to go with this approach. And the second core insight we got was about personalization and contextualization. The last decade was about spray and pray approach. What SaaS enabled you to do was if you can send 10 emails in a day, it enabled you to send 10,000 emails in a day. What the next five, 10 years is going to be about personalization and contextualization. So you can do all of that, but at a personalized scale. So it's hard to do personalization if you are at a generic level, contextualization-

>> It's all vanilla. Yeah, dear recipient, we've seen those on those email spam. That's just bad. So you're saying get down to the personal level when the account details are intimate. For example, I want to send an email to a customer, it's not a generic, "Hey, we're having... Here's our newsletter." I mean, that's an over the top example, but that's what you're getting at, personal.

Gaurav Nemade

>> Yep, yep, yep. So in the RFPs for example, it's all about personalization and contextualization, and a lot of our clients are coming... Some of them use ChatGPT for obvious reasons, and then they are like, "Hey-"

>> You read my emails, make it sound more professional or whatever. We've seen that. But to me, I like this workflow angle because process in the enterprise is where the data is. That's where the expertise are. That's where the domain users are that have expertise. Do you have the power drill? You're going to create your own power drills, but I'm the customer, I want my own power drill. So can I use your power drill? Can I make my own? Because I have process, I know my workflow and I know the data, I know I want to play around. Is that something on your radar?

Gaurav Nemade

>> Yeah. So we see a lot of customer in this dilemma right now and with Gen AI, right. Should we build this internally or should we buy a solution outside? If I understand correctly, that's what you're asking. We have customers all across the board. We have customers who have come to us and they were like, "You know what? We are going to build this internally." And we have customers who are building it internally, they realize it's super hard and then they came and started working with us.

>> Yeah, but if I'm a tool, so if you're a power tool, so you're doing use cases like RFP, that's something that's clear, not instant. But I might have other workflow things that are important to me, so are you enabling me to use your tool to build my own tool with your stuff, versus build versus buy. So there's build versus buy. I get that. I'm a big believer in build is harder now with integrated systems architecture because black boxes emerge, technical debt happens, people could leave. There's a lot of problems with trying to do something in-house, if it's plumbing or if it's an abstraction layer that's feeding the core. So I agree with that, but the model we're seeing is that enablement. Now I know my workflows, I have an RFP, but they're not always that big. Can I customize it? That's the question.

Gaurav Nemade

>> I see. I think we don't have that functionality right now and I don't think they're thinking about that anytime soon. But the way we would do it maybe eventually is through integrations. We would want to have integrated in your workflow. So we have simple integrations right now. You can start a project from whatever your CRM system is like, a customer is requesting something, you can click a button and that project gets created and invented. So we have things like that right now, but maybe in the future we would go and integrate more within your systems and provide a-

>> Well, it's still early days. I mean, we always say in the hype of agentic systems is super high right now, and the reality is it's hitting, low-hanging fruit. You're starting to see some use cases that are obvious and then we're going to see a progression we think where everyone sees more data available than the agents will get smarter. I'm not sure you would agree with that, but do you?

Gaurav Nemade

>> Yeah, so I think you hit on a really good point, which is the challenge right now is how these different tools and systems would interact. Part of the agent workflow, the important bit is that the agents have to go in and have integrations with the tools so that they can take actions. And that's where ChatGPT plugins I feel failed. ChatGPT has a great interface to go in and act on, let's say, 200 tools they were integrated with. But these APIs were fickle. The protocol was non-existent at that point of time. So I feel that Anthropic has taken a good first step in that direction with model context protocol. So that open sourcing, if that picks up, that could become an HTTP of this world essentially. And that could really enable a lot of companies like us to go and integrate with different tools.

>> I mean, I think the opportunity for you guys is to continue to take the leverage from the knowledge of your AI background and watch the space because the whole sales vertical is up for grabs potentially. I mean, but you could land somewhere, right? I mean, the old land expand, it was a SaaS concept, but when you land in the new data world, if you have data, I mean, you can easily hit adjacencies.

Gaurav Nemade

>> Yep.

>> I'm sure that's on your radar too.

Gaurav Nemade

>> Yes, absolutely.

>> All right, so talk about your role in the industry right now. What's the bet? If you had to say, "Our bet for Inventive is..." What? What's the big bet? What are you betting the venture on? What's the big trend, tech, use case, market force? What's the bet?

Gaurav Nemade

>> I think, like I mentioned earlier, I think those are the two bets that we have. One is, sales workflows are much more complicated to be able to handle by generic things like generic agents that some of these hyperscalers or Salesforce is doing. So that's our bet that we are going to be a workflow first company within sales, and that is going to propel us forward and put us head to head against some of these bigger players like Salesforce in the market. And the second is, contextualization and personalization is hard. And it might seem easy with everything going on generally, but it's super hard. I've been in this space for the last four or five years trying to understand this and it's not easy at all. So I think those are the two big bets that we are taking with Inventive and we are hoping that that will play out and get us a-

>> Yeah, I mean, scaling personalization is extremely difficult. Most people think it's personal because they just use ChatGPT, but that's a one-off.

Gaurav Nemade

>> Correct.

>> That's manual.

Gaurav Nemade

>> And in the personalization, in the proposal space, the personalization is even more challenging. You have different industries for which you could be doing RFPs, there are different countries, languages, needs are different. So I can't have a tailored response for everything. I need to understand what the context is before I can put a 200-page proposal.

>> So talk about that then. I know we are running on time here, but I want to get this out there because that's a great point. Some companies don't know that they actually could go after RFPs. That's a revenue opportunity. That's a net new. That would be a CFO conversation.

Gaurav Nemade

>> Right.

>> I mean, theCUBE, let's just say theCUBE has no idea. We've got video. Maybe there's an RFP out there for 10,000 video. I don't know. How do I know it's out there? So if there was RFPs, we'd have to then stand that up. Would we work with you guys on that? How would that work? I just made that up but that's...

Gaurav Nemade

>> Yeah, yeah. First of all, the market is huge. Like I mentioned earlier, at 1.4 trillion worth of projects get allocated by RFPs in the US alone. Leave UK, Australia globally, right?

>> All contract government for example.

Gaurav Nemade

>> It's huge, government private sectors, like anything from DMV trying to buy a cybersecurity software, to government trying to buy defense stuff and so on. So it's a huge, huge market. We are not in the space right now of helping you find the RFPs. You're in the space. Once you have made your mind that you want to bid on this RFP, that's where we come in. We could expand into this space in the future, but right now-

>> How would I work with you guys? Let's say I saw an IFP out there, or RFP out there. How do I engage? How would someone engage with the platform?

Gaurav Nemade

>> Yeah, so just come, we'll give you a demo on how the platform works, essentially the step one is, we help you just understand what knowledge you have, and can it be used to generate RFP responses essentially. So the step one would be we just have a conversation, understand what your knowledge base looks like, figure out what is the best way to move forward from there.

>> Is that a service or is that technology?

Gaurav Nemade

>> Technology. And for the onboarding, we have operations team and service enabled within that.

>> And how long does it take to get it up and running?

Gaurav Nemade

>> It could take anywhere from half a day to seven days, depending on how complex-

>> How much they have.

Gaurav Nemade

>> Yeah, how complex that ecosystem is in terms of data.

>> I got it. What's next for you guys? What's the key milestones? What are you trying to knock down? Put a plug in for the company.

Gaurav Nemade

>> Yeah, we are super excited about the opportunity right now. I think we have a really good product at hand. We have a lot of customers coming in. Literally it's so gratifying. We had a call this morning, people were just thanking us like, "Hey, thanks for building this product."

>> My favor.

Gaurav Nemade

>> No more working on Saturdays and Sundays because of this. So excited about the opportunity and if people have RFPs that they're working on, please do talk to us and we'd be more than happy to help.

>> Congratulations. Looking forward to following up and seeing how you guys are doing. Thanks for coming on.

Gaurav Nemade

>> Thank you.

>> All right, this is theCUBE plus the NYSE Wired community. I'm John Furrier, host of theCUBE. Thanks for watching.