The Inbound Harness: You Create the Demand. Agents Turn it into Pipeline

Julia Nimchinski:

And transitioning to our demo session, welcome back, Arjun Pillai, co-founder and CEO of Docket. What a treat! How are you?

Arjun Pillai:

I am well, you know. I am excited about the panel still. I’m still going on that energy. In between, I ran and got some lunch, but I’m ready to do this.

Julia Nimchinski:

Amazing, it was very well received, and yeah, the panel of the day.

Arjun Pillai:

Sorry, repeat?

Julia Nimchinski:

Yeah, it was very well received, the panel of the day. Let’s get into it.

Arjun Pillai:

Yeah, that’s awesome. Awesome. Thanks a lot, Julia. Great, to have this opportunity. Everybody on the conference, you know, you folks have been listening about harnesses and different kinds of harnesses. We… I believe that at Docket.io, we are building an agent with a very unique harness, and the uniqueness comes from the fact that the use case that we have is that we request you to put our agent to directly interface with your buyers. you know, sounds scary, right? Because, fundamentally, what most of the agents are inward-facing, and I’m… here, I am asking you to put it outward facing, right?

So, Julia, what we have built is essentially an inbound harness. Let me share my screen, I’m gonna show, like, 3 very high-level slides and jump directly into the product.

Julia Nimchinski:

Amazing.

Arjun Pillai:

Yeah, so like I said, we have built an inbound harness that allows marketers and revenue leaders to convert more of their demand into the pipeline. What does that mean? Imagine level 4 PM on your website. A serious buyer is on your website. They did their research on ChatGPT, your company showed up as a potential vendor, they reached your website.

At that point, what happens in today’s website that hasn’t changed for the last 25 years is that they get a static website, which is… reads like a brochure. and a And once they fill out a form, they know that they are potentially going to wait for one day, one and a half days, or two days for an SDR to get in touch with them, and the SDR is not going to answer their questions, the SDR is going to basically book a meeting for them, and then the AE comes. That world is over, right, where People tolerating the forms is not gonna happen.

What they want is, they want to jump right in and ask some very specific questions and get the answer to know whether this company, this vendor, is a good fit for me or not. And that is what our agents enable you to do. So, fundamentally, we believe that 1% of your website visitors are filling out a form, 1.3, 1.2, whatever that number is. But the other 99% of your website visitors, there are probably 5-6% of them who are serious demand buyers. Please don’t let them go dark. It is not a traffic problem, it’s a capture problem, and we help you fix it.

And we help you fix it by providing you with a demand agent that turns your buyers into agent-qualified leads. And let me show how that… that looks like.

Arjun Pillai:

Let me, go out. By the way, all of you, I’m… screen sharing is paused. Let me reshare, because I think it’s, did not. Yeah, I’m supposed to do the whole… There you go. Okay. So, by the way, all of you folks can go onto our website and then play around with it, right? So this is not a demo where I’m showing something that I cooked up using AI, and then, you know, it shows well, but it’s an actual thing. So, imagine on our website, somebody comes to our website, we have all these pages that we have written out.

What we are continuously seeing is people don’t want to sit down and read through all these pages, right? Instead, they simply want to ask a question, just like how they would ask on ChatGPT and get an answer. So, we have integrated into our website, our agent. We call it Docket Frameless. We believe that this is the next evolution of the website, where it is an agent-first experience, where agent is not sitting inside a widget, it is actually part of the native experience that is happening on your website. So, as you are hovering and going through the website, what you are seeing is The contextual callouts are changing. as I’m reading different things on the website, depending on the source from which I’m coming in, as I’m going through, these contextual callouts will change.

And you’re also seeing that the agent, at this point, is more like a co-browser with me. It is not just a widget, it is available anywhere I go on the website, and I can start creating or having that conversation with the agent. I can browse through. I feel good about everything. How fast can I deploy? What security controls are included? Okay, those are all relevant questions, so it makes sense. I want to interact and start having that conversation with the agent.

Docket AI Agent:

Arjun, nice to see you again. We talked about answer accuracy and getting started, so what’s the one outcome you want to improve first? Conversion rate or qualification quality?

Arjun Pillai:

So, right out of the gate, you are seeing that there is memory. We have short-term memory and long-term memory. This long-term memory that you are seeing, it remembered who I am, what was the context of the conversations, which pages I visited, what was the campaign that I clicked, what was the webinar that I attended. The agent knows all of these things. It also knows about 60% of the cases which company I am visiting from. And it has additional information for contextualizing For a healthcare company versus a software company versus a manufacturing company. Hi, Ora, nice to see you.

I know we have interacted in the past, but I have… I forgot pretty much everything. Can you tell me what Docket’s Marketing Agent is all about?

Docket AI Agent:

Sure. I’ll give you a clear overview of what Docket’s marketing agent does, and why it matters. Docket turns your inbound traffic into real buyer conversations, then qualifies intent and routes the right accounts to sales. It answers with approved knowledge, qualifies using your chosen rubric, and books meetings, so reps get context instead of a form fill. Want to see how qualification and handoff look in practice?

Arjun Pillai:

Yeah, sure, what does qualification look like?

Docket AI Agent:

Let’s walk through how qualification works in a live conversation, and what sales receives. Qualification happens inside the chat. Docket reads what buyers ask, how they respond, and the context of the account, then uses your chosen rubric to decide if they’re ready. When they are, it routes them to the right owner and can book a meeting. Sales gets a concise dossier with questions, signals, and context instead of loose notes. Want to map?

Arjun Pillai:

So… so you’re seeing that the agent pulled up a slide, much like how a human would pull up a slide and is walking through the slide. It is asking me follow-up questions and qualification questions, not like an interview. It’s answering me and then weaving in the qualification questions alongside, and it’s a near-human experience with an avatar, with a face, with voice, with text, with slides, with videos, with demoing. All of that is included in the agent. No, that’s not what I want. I know you do routing, but here is my question. How do I ensure that your answers are accurate?

I am really worried to put an agent to directly speak with my buyers.

Docket AI Agent:

Let me explain how Docket keeps answers reliable and safe for buyer-facing conversations. Docket answers from approved, governed knowledge instead of improvising, and it differs when something is out of scope or not available. That reduces hallucinations and keeps buyer responses consistent and traceable. The calendar is open, so pick a time that works best.

Arjun Pillai:

So it is basically saying that, let’s go deeper into it, let’s book a meeting. And fundamentally, what is also happening is, behind the whole thing of the calendar, there is a full routing that is happening, where it is picking the right person to route it to, based on enterprise routing, segmentation, territory-based routing, all the rules that our customers have. All I can do… all I have to do is simply confirm the booking, and the meeting goes to the right, salesperson to their calendar natively integrated. So there’s no need for, like, a, you know, more complicated, routing or different tools or anything like that.

It’s all in one stack. Now, if I go ahead and show you the backend of the system, this is where a lot of marketers really like. Part of the value of the agent is the downstream data that it creates. So if you go to the agent dashboard, which is what I’m showing, all of the conversations, you get the entire transcript, as well as the audio recording of the conversation right here. And… What our agent can also do is the agent can… this person has interacted with our website 61 minutes with our agent, believe it or not.

61 minutes. This is something that we see consistently, by the way. A lot of the buyers will come interact with our agent once, 4 days later, they’ll come back again, 7 days later, they’ll come back again, 15 days later, they’ll come back and then give us the email and book a meeting. So, our agent did the conversation, created a summary, pain points, discovery questions, next steps, all of this is Created or, like, gleaned from the transcription that has already happened. And our agent also has the ability to write it all back to the CRMs, or marketing automation, and all of this.

This is all part of that harness, right? You are supposed to engage, you are supposed to qualify, you’re supposed to route, you’re also supposed to alert on Slack, you’re also supposed to write it back to your CRM, or marketing automation. This is what we call as the inbound Harness, right? It has the right prompts, it has the right memory, it has the right tools to make sure that the agent is delivering the experience that a buyer needs and a seller needs. So then, you have all this amazing data. This data was not available to marketers before, because marketers had no idea what exactly were the questions that people had in their mind, now they can actually do it.

Now, we go one step further. We take all of this data and give access to our customers through an MCP. So, if I go here… Let me close it. let’s say I want to go and say… Let’s do this. So this is my Claude. And I asked a question like, how did my marketing agent deployment in Docket.io do last week? How it is doing on a conversation, a dashboard would be great. Right? So Docket’s MCP is connected here to a point where our customers can simply go in and ask all sorts of questions, and the agent is gonna say how many widget loads happened, how many conversations happened, conversation stage by stage happened, what were the most questions that actually were asked.

I can say, what were the… let’s dig into the conversation. Let me know what are the repeated pain points my visitors shared to the agent last month. And then it pulled the 3,368 conversations between August 9th and September 9th, and then it Looked at all of this. Meetings that aren’t real, 83 conversations had that. Traffic that won’t convert? 75 conversations. Leakage after the click? 48 times. So these are the main pain points that our customers are trying to solve with Docket. So now what I should do is, as a content marketer, I can write content against all of these questions, and that is creating a windfall for me.

So the chat GPTs and perplexities of the world are going to kind of take this, index it, and that drives more traffic to our website. So. the earlier chatbots were these individual chatbots that were, like, just chatbots or glorified forms. Now we have gotten to a point where an agent, an inbound agent, is sitting on your website. It is tracking everything happening upstream, whether it is Marketo or HubSpot, Outreach, all these systems. taking all that context, having the right conversation, and pushing it into downstream systems, right? Salesforce, or Outreach, and all these downstream systems. And also giving you an MCP to do your work better.

So this is what an inbound Harness… I just wanted to show this to people so that everything that we are talking about as, you know, a harness is theoretical, so I wanted to show what a real inbound harness looks like, which is built for a very specific use case, right? This agent is not supposed to give you relationship advice, right? If you go and ask our agent, hey, give me relationship advice, it’s not gonna tell you. Right, you can try. Go to Docket.io, try to break the agent. If you break it, send me a LinkedIn.

That is learning for us. But, fundamentally, it’s not going to answer those questions. It’ll slowly bring you down, I’m here to answer Docket questions. Or on our customer website, it’ll say, I’m here to answer these questions. So, this is how you harness it. As a vendor, we give this inbound harness to you, and we will give you all the additional tools alongside, whether it is calendar integrations, routing, things like that. Now, you use it to improve your inbound. Automate your inbound. Get 15% more leads into your pipe from the same traffic that you already have. you know, we have about 60, 62 customers at this point in time that have deployed the agents on their website live.

We also have a page on our website that shows a bunch of the agents that we have deployed on our customer website, if you want to check them out. But that’s what Docket does, and that’s what an inbound harness would look like.

Julia Nimchinski:

Love it. But it’ll be more timely, and I mean, beautiful, just, you know, continuation to the panel, super practical. Thank you, Arjun. Curious, how does the pricing work? Can you talk a little bit about that?

Arjun Pillai:

Yeah, our pricing starts at about $36,000. It’s a very transparent pricing. You know, we go to $36,000, $48,000, and then the enterprise package. Some of our competitors… there are some of legacy competitors that got acquired by other bigger companies that are way more expensive, that are actually not AI-native, and they have, you know. the way optimization and all is different. We actually are priced in a very fair way, transparent way, go and ask the agent, the agent will give you the exact same pricing that I gave you. And the pricing depends on how many agents that you are deploying, and what are the features that you need.

If you need advanced features, if you need Docket Frameless, which is a native experience on your website, obviously you will pay a little bit more. If you need, like, 7 agents, then you pay a little bit more. But typically it starts at $36.48 and then goes up.

Julia Nimchinski:

In terms of your vision for this category and the roadmap, what can you share?

Arjun Pillai:

Yeah, good question. See, we believe that this chat or the interaction layer is actually one place where we can activate the intelligence that we are capturing. Look at the entire RevTech stack of a marketing ops person or a RevOps person, right? There are very few systems that are tracking a person at a cookie level, at a pseudo-anonymous level. There is, like, Marketer and HubSpot. But the data that Marketer and HubSpot is housing today is practically unusable. RevOps people cannot do much with that data. So, what we are, like, Salesforce is not tracking that data, Outreach, Salesloft, none of these systems are tracking that data.

Docket is tracking everything that a marketer and HubSpot is tracking, and more. And then we are putting an agent on top of it to infer the context of every single buyer. Where exactly is this account? Where exactly is this buyer in the buyer curve? So our view is that website agent is one place where you can deploy us. You can deploy us inside your app. Those are the conversational experiences. In a month’s time, you are going to see us deploy an email agent that automatically nurtures your existing 5,000, 10,000 contacts that you have lying around, basically, in a very, very personalized fashion.

And over a period of time, you are going to see us, deploy more and more agents on the field. Fundamentally, Julia, the way we think about it is, 20 years or so earlier, Marketo, HubSpot, MailChimp, Iloqua, Pardot all came in. That was, like, the big change that happened for marketers. Before, it was billboards and newspapers. Marketing automation kind of changed that. But in the last 20 years, there hasn’t been a big change to that, right? The marketers are still using the same old systems. Marketers who I speak to, they know that instead of 10 people team, they’ll have to do the work with 3 people or 4 people and do more.

But they don’t know exactly how to Be it, like, 3 or 4 people team, and still accomplish whatever they are accomplishing and more. They can do it by going into what we call as the Agentic marketing paradigm. So the mission of the company is to help marketers achieve Agentic marketing, and the inbound is, like, the starting point for us, and we are going to deploy more agents to help them get there.

Julia Nimchinski:

There’s a lot of, I wouldn’t call it signal, but there is a lot of noise happening on LinkedIn per se, about the role of CMO dyeing. What are your thoughts, and how do you envision the future of it?

Arjun Pillai:

That’s wrong. That’s where I would start. See, a CMO, the amount of decisions that a CMO is taking on a daily basis, right? AI is nowhere close to taking those decisions, right? So the role of CMO is not dying, it is changing. Instead of driving an army of content writers and agencies and all of that, what the CMO will do is they’ll have a limited set of really smart high agency people who are working with agents to achieve the goals. When I think about a CMO job, there is all the demand gen things, like the paid ads and things like that, that becomes a role.

All the content-related things, that becomes a role. Then there is probably an investor relations, if you’re a big company or an analyst relations, or like, you know, like PR or event marketing. So. The entire work of marketing will come down to 3 or 4 people, and then there is a marketing factory, which is an Agentic system harnessed for that particular company that they can use to drive all the different tasks. And the CMO is essentially the person who injects the objective and the taste into that organization. So when I said our mission is actually to transform that CMO’s job into the Agentic marketing job.

If we do that, then we are a successful company. Literally, that’s what we are trying to do.

Julia Nimchinski:

And in terms of GTM as a function, how do you envision it? One year from now, five years from now.

Arjun Pillai:

Yeah, I believe that all the tasks that we are starting to do with humans will start to start with an agent. Inside Docket today, almost every single thing that we are doing, it starts from an agent today. Think of a product feature launch that we are doing, right? We literally tell our agent, which is already harnessed in a particular way, hey, we want to launch this particular feature. Figure it out. Because it already knows when we launch, here are the things that we typically do. here are the creatives that we typically need. Here is the typical promotion that we do.

Here is the amplification that we can do. It already knows that You know, Gary is our VP of Marketing, Lauren is our head of marketing, Kavya is our content. So, our factory already knows these things. It has the tools to ping them on Slack, get their inputs, come back and do the kind of things that they are supposed to do. So, I believe that humans… we… for a really long time, we have been talking about human in the loop. I think that’s gonna change. Humans will be out of the loop. The loop is being done by the agent with inputs from the human, but the loop is getting completed, just like that, right?

So it’s a factory that’s completing the loop, and human is outside that loop. It’s not… You are not triggering it all the time, you are not breaking it all the time, you’re not approving it all the time. Sometimes you do, but most of the times the factory is gonna run on its own.

Julia Nimchinski:

Love it. And last question, in terms of security, you showed the, essentially, the marketing dream. Yeah. So many use cases, how it could be immediately leveraged, obviously, but since we can review every chat, every conversation, can you just address it in terms of security, compliance, all that?

Arjun Pillai:

Yeah, yeah, yeah. So, see, when you go to any particular website today, there is a consent banner that pops up, thanks to GDPR. It’s just a pain, honestly. But, there is a consent banner that pops up. When you say accept, or when you don’t even say accept, what is happening is you are accepting that some of your data is actually being tracked for marketing and ad purposes. That’s what you are basically seeing accept 2. And, like I said, whatever we are tracking is already getting tracked by a cookie of a marketer or a HubSpot. in the same umbrella of marketing purposes or ad purposes.

It’s the same thing that we are also tracking, right? We as a company, don’t track people at a person level unless they have made it explicitly clear. When our agent says, hello, Arjun, it was because I, Arjun, has given that data to the agent. Otherwise, we are only tracking the people at an account level, which is, you know, compliant and all of that in all the necessary things that we are doing. We also adhere to all the consent management systems, you know, right to be forgotten, GDPR, all the compliance-related things. And from a security standpoint, we are a SOC 2 Type 2 company, ISO 27007, all of that, right?

It’s an ironclad system, both from a security standpoint, privacy standpoint, and an AI security standpoint, because we don’t learn on any of our customers’ data.

Julia Nimchinski:

Beautiful. Where should our people go?

Arjun Pillai:

I… I would love for you folks to go to Docket.io and experience the agent, right? Just play with the agent, ask it different kinds of questions, try to break the agent. That is probably the best place that you can start to experience what the future of a website or a B2B go-to-market motion looks like. then to be in touch with me, I would love for you to connect with me on LinkedIn. this is what I do, you know, I keep sharing whatever the learnings, all the experiments that we do, all the mistakes that we make, all the successes and the fails that we have, because we are fundamentally an AI lab for marketing.

So, we put a lot of experiments and failures and successes out there, so we would love for your interaction to be there. Tell us what worked, tell us what did not work. And yeah, if you think your inbound motion can get a boost with Docket, we would love to talk to you, obviously.

Julia Nimchinski:

Amazing. Thank you so much.

Arjun Pillai:

Thank you for the opportunity, Julia, for everybody watching. I appreciate your time, thank you, and looking forward to speaking with you.

Julia Nimchinski:

Thank you, Arjun. Take care. Bye-bye. Thanks. And that’s a wrap for our Day 2 of the Agentic Harness Summit. Thank you so much for watching. Thank you to all of our moderators, speakers, community, partners. Please join us tomorrow. We have an amazing day. Day 3. Mike Maples from Floodgate, Head of AI at ClickUp, Jaya Gupta from Foundation Capital, CMOs of Salesloft, 6sense, Insight Partners, and many, many, many more. So, do share your feedback, share your questions, comments, concerns. We are trying our best to make it as valuable for you as possible. And yeah, see you tomorrow.