Fireside Chat: The Future of Agentic AI

Today we have a very special guest, Abhijit Mitra. He's the CEO of Outreach. Welcome, Abhijit.

Thanks, Gerard, and thanks for having me here.

Tell me what we should know about you that's not on your LinkedIn profile.

I like creating stuff and I like creating software, I like creating businesses. That's kind of what I've done my entire career and that's what I'm doing here at Outreach. I am super excited about what technology can do in the service of people. I truly believe in that. A huge impact of that is with AI. Generative AI is new and we are all trying to figure out what to do with it. And I truly want to make an impact to people with AI.

Let's dive right in. We're going to look at what Outreach has accomplished. Increase win rates by 26%, improve forecasting accuracy by 43% and shorten sales cycles by 19%. How do you achieve those outcomes?

We use a combination of agent, take AI, assistive AI, predictive AI to automate workflows across all revenue teams. So Outreach, we have an AI revenue workflow platform and what that does is all the way from forecasting to deal management to coaching to prospecting, engagement to retention and expansion. Not just new business, but also post sales where you go and retain your existing customers. Customers and expand your business with them. All of those workflows are powered by our platform. You know, we have all the data that is being gathered through all the interaction that they're doing within Outreach. That data powers our automation engine. Then that triggers the AI agents which are orchestrating workflows. Not everything is done by agents. There's a lot of humans also involved in this. So we actually take specific tasks and automate them through Agent and we are using the technology ourselves internally and for all our customers as well.

When it comes to creating a point of view for the salesperson of the customer situation that they're calling on, how does that work? How do you get this information curated and then make sure it's accurate and then deliver it in real time?

I use Outreach forecasting and I use my Pipeline summary dashboard. And today's the 1st of August that we are recording this. Yesterday was the end of my quarter. My AI prediction has been accurate every quarter. The way it works is that because the AI actually is looking at all the work and helping the sellers do their work day to day. That's why it knows which deal is in what situation, how frequently are there deals being made? Have you had engaged the stakeholders, the right stakeholders in your deals or not. And that's how that prediction is so accurate, you know, so basically from the forecasting, it's tied to that really. The deal has a health score that's generated by AI. And again, that health score is based on what you're actually doing with the deal. You can actually go and ask your AI assistant what is actually happening in the deal. Will the deal close? Do you have the right buyers identified? Do you have the right pairing points identified? It's going to summarize all of that for you. And the reason it can't do that is because of our conversational AI. What that does is that when you're talking to a customer, it's actually giving you guidance in real time. It's actually able to detect intent in less than 10 milliseconds as we're having the conversation. And you use that to make sure you're being coached in real time, not post call, not like after the call has been over and you've done the damage that you could have done, but in the call along with post call as well as post call intelligence that gives, you know, pointers to your managers to coach you better. So all of this data is actually feeding into the forecast intelligence that I talked about. Now your question was, how does this AI help me craft the right messages so that I know what to actually say to my customers besides the real time question that I talked about? So what we introduced to the market is revenue agents. And what these agents do is they essentially work side by side with the reps. They're giving instructions, hey, go find these customers in this segments, these kind of people that those customers were interested in this kind of products. Look at what's happening for them, them, what news information, what signals you can gather, look at what data we have about them, what have we sold them, what have we not sold them? Then tell me when to target them with what content. And by the way, go ahead and do that on my behalf. And if they say they're interested in talking to me, then put me in the discussion with them. Since the AI agents are actually doing a lot of this grunt work, that's why the conversations are so much more meaningful, and that's why the AI is so much more impactful and so much more predictive.

How do you position outreach as a solution that is more comprehensive and gives you better results faster?

We say that, look, you can put a solution together from different vendors, or you can come to outreach and have one system. Having the data in one system actually makes a big difference because Then it's all interconnected. Your workflows are talking to each other and your agents are so much more effective because they can talk to other agents within the outreach ecosystem or even outside of the outreach ecosystem, as well as work hand in hand within the humans. So this is what we tell people that really the differentiation of outreach is that it's an agentic AI platform. And because it's a platform, you can deploy it across your entire revenue workflow all the way from the time that you first contact your customer to how you sign them up and then how you support them and how you retain them, how you expand them with customer success as well. So that is truly the differentiation of outreach.

So you are having agents that execute customer success tasks as well. Right.

The revenue agent is able to execute a lot of use cases. So one use case is in the retention and expansion space. Let's say a software product. You have sold a software product. And then typically when we sell, we track usage of our product to see if the customers are using the product and they're getting value from it or not. If the usage is decent, then we want to sell you the next product. It's a land and expand strategy. Right. For most of us. So that's when then the revenue agent can figure out which customers having decent usage, what products have they bought, which products have they not bought. There's a product and purchase section in outreach and based on that, come up with optimal messaging and optimal timing to help you position the next product upsell too. That's the expansion flow. On the other hand, if the usage is going down and you know that this is a potential contraction risk because the licenses are underutilized, then the revenue agent can actually reach out to those customers and say, hey, you know what, have you tried these best practices? Oh, by the way, customer success manager, you need to now reach out to this customer. This is the time for you to have a conversation because the usage is not looking so great. So these are examples through which we help account managers and customer success managers also with outreach.

So how do you orchestrate all those agents and teach them how to play from the same sheet of music and optimize everything?

It's a combination of data and it's a combination of how the agents are triggered and then how are they orchestrated. So if you are not, you know, using the right data and giving the right instructions to the agents, they're going to go haywire, basically. So that's a very important function that revenue operations teams has to do to make sure that you have a right strategy as to where and which use case you want to deploy, what is the data and what's the input and what the outcome you want to get to. Once this is defined, then the tuning is relatively simple and easy to configure. That's really not the pain point. The pain point. The main issue is really around figuring out your strategy and what you want to do with this platform. There are use cases, for example. So for example, let me give you an idea of how it works. So interaction data is something that you get within outreach anyway and that's very fresh and clean data because what you and I are talking about is coming from directly from the customer. Then it has CRM data which is synced into outreach and that is always not very clean because you know, people put a lot of stuff in CRM and that's not, you know, the same customer appears a hundred times the same people appear and sometimes the phone number is not accurate and you don't know even like. So there's a lot of, you know, data cleansing. That's usually a problem in CRM system. We don't solve that problem, but we do manage to enrich that with the interaction that you have within outreach. Then you have other first party data that you have on your landscape that we actually help to bring into outreach through our Snowflake and databricks connectors or through open APIs even further. There are like third party data that you can get through our data vendors like Zooming for sales center like Florium. They bring in external data signals or external data about your customers prospects into the outreach platform. All of this gets aggregated into our data cloud which we give to all customers at no additional cost. And then that actually is where we make sense out of the data. So there is some amount of data harmonization that's already happening in outreach based on which you can trigger your agents to take actions. Even if after all of this, if the data is not so accurate, one of the things I hear from customers is that look, Abhijit, can I ask my agents to execute actions on this data and see what sticks. Because what sticks is you know, if a customer responds back saying I'm actually interested in that personalized message because the timing is right, because you have the right thing for me, then actually even that is actually worth the while because it saves a time, ton of work because all the junk data that you had in your system, okay, fine, they won't get a response. That's okay. It didn't waste human time we wasted some AI time. That's okay. So this is the way in which even if the data is not, say, accurate and clean, it's still helpful for the AI agents to act on that, because it cleans it out for you. It filters out the noise from the signal for you.

What are you doing to win in this conversion dynamics in the market so that you win that consolidation battle?

My message that I'm saying in this video is what I want all my customers and prospects to hear and internalize because, frankly speaking, a lot of what we do is not so well known in the market today. And that's because we are so good at one thing. That happens for every company who's so good at one thing that everything else they do is not so well known. So I had customers up on stage at Unleashed talking about how they use the full platform. I had SAP, I had Siemens, I had Amplitude. I had a bunch of other customers go up on stage and talk about their outcomes when they deploy the platform. And some of them talked about how the deal size has gone by 50%. Some of them talked about how they have consolidated multiple systems and outreach. They don't need all these other systems. And we're just talking about that before. And they've saved money, they've solved cost. Some of them talked about how they have standardized the sales methodology, their sales, their core sales process, on outreach across, you know, global deployments across 180 countries. Actually, some of them talked about how outreach is the heart of everything they do in go to market. These are the stories that our customers are talking about us now that I want our other customers and our prospects to hear about. Because in the end, it's their success, you know, that is our success. And so we get very excited when our customers talk about what they have achieved with the outreach platform. And the more and more customers see those kind of outcomes, the more successful we will be.

Yeah, what you're really saying is that the CEOs need to understand that if AI is not part of their business strategy, they're headed towards obsolescence.

CEOs mostly understand that already today. I think the majority of the market really understand it. But I think the, the lack of clarity is actually how. How am I going to get there? What I'm saying is that we can help them with that how. The problem, you know, Gerard, is that everybody says you can help you with the how. And that's where customer successes stand out. So if my customer is saying, hey, I have done this with outreach, you should probably do the same. Those are the stories that we want to focus on because that actually drives credibility with my customers, my prospects on what outreach can do for.

So there's also a generation gap in terms of adoption because people under 35, they have no problem adopting AI solutions. But half over 55, age 55, they don't. What can leaders do to address that generational divide? And do you have a plan for your own organization to bring the older generation, you know, to lead them into the future?

Some personal anecdotes is that my mom uses ChatGPT every day and she's way over 55.

You have raised her well.

Thank you. So I think the reality is that there are people across the all spectrum who are actually using a lot more generative AI than we think they are. That's what I meant. That's what I wanted to say. So even though the tooling may not be used as heavily, there is shadow it, personal shadow it sprawling across every company. And so I think the solution to that in my mind is for the companies to adopt AI fundamentally as the tool of choice so that you don't have all these shadow it, personalized shadow. It's very dangerous actually, because you don't know what data is going out there to each other. That's not very good. So that I think is going to be really my response to your question saying that the way to address this is to make sure that you are taking a very aggressive stance on rolling out a AI platform across the entire company and also vet out that platform. So outreach. For example, we are ISO certified. There's ISO certification of 42, 2001. I don't know if you know about that, Gerard, on AI governance. And we are certified on that. One of the very few companies, in fact the only company in Revtech certified on that. So once you have a platform like that, even now as a company, as an enterprise company, I would feel more comfortable in rolling that out to all my employees so that you don't have to resort to things outside of what the corporate systems are there for you. If you do that, then I think that option across all spectrum of all people will go up substantially. We are seeing that already happen. Because here's the thing, when you roll out AI, especially agent, it's not people are forced to do something that they are not doing before. It's the other way around. It's like people are not doing something they were doing before because the AI is doing it for them. So AI adoption doesn't mean that I have to go for Something in AI now it's like, wait, wait, hold on one second. AI has actually figured out a bunch of stuff for you. This is the stuff that you anyway like doing. So go do more of that. That's the paradigm here. So it doesn't matter if you're a 35 or 55, we all love to do things. For example, if I'm a seller, I want to have the strategic sales conversation with you to enclose a deal. That's what I want to do. If the AI is going and doing all of the grunt work, researching who I should be talking to, what I should be talking to them, crafting my email messages for you, I would welcome that. That is the paradigm shift that I'm talking about.

Is there a. Does outreach have a tool like an AI ROI analysis checklist?

We do show to our customers what the AI is doing for you.

Okay.

We also want to make sure customers are keeping track of how many credits are using for us. Because the whole idea here is that the more you use, the more benefit you get, the more you want to buy from us. So that's all part of the tooling and we provide that. Besides that, we also have our forward deployment engineers and these are people who will be working with you hand in hand to make sure that the best practices of identifying the use case is getting your data together, figuring out like the rollout strategy and all of that. Actually we are hand in hand working with you. And beside that, we also have our customer success motions post customer success motions who also use out of the AI agents.

Like I mentioned before, maybe there's an idea that you can create at the outreach to instead of the salesperson of the month, you have the AI champion of the month and publish the success stories.

Absolutely. We do have now amplify awards that we give out to our customers who are championing certain causes. Let's say we did that unleash as well and we will continue to do that going forward and we can roll in what you just suggested into that.

Actually, can you paint a picture what the high performing sales organizations are going to look like three years from now.

Are going to do a lot more work autonomously through AI agents and that is definitely going to happen. They're going to be a lot more strategic in their how they approach their customers. For example, the basic, you know, things around hey, I have a product. Are you interested in this product? Hey, what's your pain point? Let me solve that for you. Those are anyway done by the AI agent. For example, if I'm in telecom and I'm looking at fixed wireless or I'm looking at fiber network utilization. Now you may think that, oh, Abhijit, why are you talking about that? You are essentially at top of the funnel revenue orchestration workflow engine. What's that got to do with fiber network endpoint and network utilization? It has everything to do with it. Because in the end, you have a demand and you have a supply. You need to match your demand to the supply. It's a demand supply optimization problem. That's what this is. So if you can have those kind of conversations with your customers and say that, hey, you know what, I can really help you make sure you make the most out of the infrastructure that you have put in place by targeting those customers, then you're having a much more meaningful conversation, much more core, fundamental conversation with your customers. So those are the kind of conversations that I think we all long for. So I think the three years from now, the sales organizations are going to be a lot more strategic in how they approach, how they plan, they go to market strategies and how agents are standing side by side. Really, you know, part of the workflow is done by agents. When I need to come in, when I want to come in, I come in and I do my work and I get out of the way that the agents continue to do their work because this daily grind of me doing everything is going to end. That's not going to happen anymore. So it's only the things where humans will add value, where we love to do things that we want to do is what we will be doing.

Zig Ziglar said, logic makes people think and emotions makes people buy. There is a conversation I just had with a CEO this morning where she said she's using an AI agent to help her with meditations, to help her stay fit, to help her eat the right things and do more exercise. So she gets daily reminders from an AI agent. So, and I thought, why not have a, an AI agent explain the customer to you in terms of what type of person? Like, you know, if I was your agent, I would say, Abhijit, you're a fun person to work with and you're very intelligent, very, very educated in your thought leader. And you are somebody that develops high trust. And I think that if we could build something like that for the salesperson, they would feel more comfortable talking with people that are different from them.

Yep, absolutely.

Well, I want to thank you for being so generous with your time and so insightful with your answers. And it is my mission to help you spread the word.

Thank you for doing that for us, Gerard.