Episode Transcript
[00:00:00] Speaker A: Foreign.
[00:00:09] Speaker B: Super happy to have you on Vinay. Can you kind of give me the story of Process Street? Where you guys started, where you're at now, how things are going?
[00:00:15] Speaker A: Yeah, sure, geez. I mean, we've been at it for eight years, so it's kind of a long story. But we originally started Process street just building a tool to solve our own problem. We're running a. An agency, a services business, and we were using a lot of remote employees in different countries and different time zones.
And we just wanted a tool that was almost like a virtual manager that was basically assigning work and not just telling people what to do, but also how to do it and then making sure that the work was actually done to a certain quality level. So it was more about adding a kind of layer of process and quality control on top of the normal task management tools that were kind of out there. That was the original problem. It's like a lot of the task management tools will tell you that you need to do something, but they don't really help you control the quality of that work in some way. And so we found we're using freelancers or we'd hire someone upwork, and they're like, we did the thing you asked me for and I'm like, oh, but the quality is not what I need it to be. So let me send it back for rework and let me send it back for rework again. Let me send it back for rework again. And that kind of like cycle of not getting the quality and the deliverables that we wanted was one of the reasons that we felt that the traditional task management kind of spreadsheet products weren't, you know, giving us the solving the problem that we had. So we built Process street and we started it as a generic kind of generic but horizontal process management products. So think process management like Asana for processes or Slack for processes or something. Right. So just any team, lots of small businesses kind of using it initially for any type of process.
Our original kind of thesis was very, I'd say ignorant, which was, oh yeah, all businesses need processes, right? Obviously. So our market is like all businesses. And that's kind of how we went into it. We're like, we're a business. Other businesses probably need this. And that's how it started. And we started building it, we launched it out in the market. We at least knew that we would be the first customer. So if nobody else wanted it, at least we'd have one customer.
And when I think about it, it wasn't a clear Long term strategy. But it also wasn't a bad place to start from and kind of give you that initial starting place to figure things out. So we built it, we put it out into the market and then we just started iterating, just launched it, started getting feedback from our friends and our initial user base and then just started building out functionality but still not really changing the strategy of this is a horizontal process tool for all businesses.
And we had some other strategic components in there. Like it was, we were a small business, we were very used to self serving into products, not going through a sales process.
Back then it was called self serve, now it's called plg. But we, you know, so we also kind of built the mechanics of the go to market motion in a similar way that we wanted to buy, right. So a lot of it was really just kind of taking everything that we knew as we got it out into the market. We got a few thousand customers, a few million dollars in revenue. We started to get a lot more comprehensive understanding of the market and the problem we were solving and the customers we're serving and the value that we're adding to those different customers. We could see different cohorts of customers and segments and who were really kind of engaging the most, retaining the best, expanding the most. And we really were trying to figure out, okay, where is this product adding the most value? Right. And that was a kind of long journey of like how do we kind of go from this broader like horizontal product to something that is more focused. And one of the reasons that we had to do that was because we just found that growth started to slow with the horizontal product. And there's a couple of components here. Like I think this is just kind of if, if, if the other listeners are SaaS founders, it's just kind of like a fun little journey that we went through where most of the traffic that. Well, if you're doing a PLG low ACV product, right, you don't really have many options for go to market channels. You know, mostly it's organic a lot of the products. So organic can compo compose of organic through word of mouth, which is what everyone gets organic through. Like virality hooks in the products. So you know a zoom call, like you send a link to somebody else, they join the zoom call. Riverside is a good example of that too. Where you send the link to someone else, they join the call. Now naturally you're looping in an external person is that product. So now this is distributing Riverside, right? Like if I came in, I'm like Oh, I want to start my own podcast. Maybe I would use Riverside. So that's like a reality hook inside the product or like marketing and content. A lot of people call this organic, but it's like founder led sales where the founder is building a brand on LinkedIn or like you're doing some type of content arbitrage or you're, you know, going and emailing a bunch of TikTok creators and getting them to like make videos about your product or something. It's like it kind of really doesn't cost much money if you can crack it. A lot of it is like founder time or maybe a few people on the marketing team's time. But a lot of people shove that under the rug of organic. Really. It's not really you should be calculating the price of the time of the founder or the marketing team, but whatever. But essentially that's kind of how a lot of the time those marketing or kind of organic channels can still work out to be really cheap channels. Right? Like if you get one marketing guy, any kind of cracks virality on TikTok or something, or cracks SEO in a way like that. You can get a really outsized amount of traffic and business for a very small investment. So you might just still call that organic or a growth hack or something.
There's few channels where you can get customers for that cheap and usually they don't last, they get competitive, et cetera, et cetera. So we kind of did that same playbook and to be fair, you know, we, we, we made millions of dollars kind of with those organic channels.
And you know, from, from most people's perspectives it's very successful. But at some point that kind of those channels dried up and they weren't sustaining growth for us in the long term. And I think for the real PLG products to work, they both need to crack like an organic channel, like marketing, SEO, growth hack, something, something. And they also need some type of really strong virality component like Slack is inviting other users or Calendly is like sending the thing out or whatever it is. And the organic traffic kind of starts the flywheel for the PLG engine. And then the virality of the PLG engine kind of starts to grow off the top of that initial spike that's brought in through whether it's SEO or the founder or TikTok or some type of like cheap channel. Right. The problem that our product had was that we found that processes don't really have like a real viral component. One of the problems with processes is there's this There's a build phase.
So instead of being like Zoom or Slack, where you can kind of like within 10 minutes you could have 10 people on using the product, like chatting with each other or all join a call together. And there's no real setup or build or implementation kind of phase for the product. You can just kind of almost immediately get value within the first hour. And a lot of the time for multiple people can get value within the first hour. That's like a strong component of a PLG Virality product. The other type of PLG product that works really well is like a single, single player use case. So where you can get value as an individual user, use the product on your own. You don't need to convince your boss, you don't need to convince anyone else. You don't need to put in a credit card, you can come into the product. So Airtable is a good example of this, where maybe you can like come into Airtable, you can build your own little relational database. You can spend a week or two on it on the free version, and then you can kind of like showcase that product and say, hey, look at this thing that I built.
I built it on my own, I'm getting value out of it on my own. And then maybe other people in the team can also get value out of this database, but you can get value out of it on your own. As a single player. You don't need to convince other people in the company to use it. As soon as you need to convince other people in the company to use it, there's this additional layer of friction, right? So when you can go in and use the product on your own, test it, and you don't have to show it to anybody else until you can prove that you've gotten value out of it. That kind of creates this almost protective shelter for the employee inside the company to actually go in and invest in the product and test it in a way without necessarily embarrassing themselves or exposing wasting other people's time in some way on a solution that doesn't actually end up getting picked up and doesn't offer value. If you say, hey, we need 20 people to test this product before we can decide if we're going to use it or not.
And you take 20 people's time and you say three hours of each of those 20 people, suddenly you're using a lot of resources to kind of like assess this product before it can actually be used. And so anyway, for PLG products to work, they need one of those two components. They need like a viral Component or they need like a single pair use case.
Processes don't have either of those.
By nature, there's this build phase, so it takes a long time to build the process before you can deploy it because you need to make sure the process is right and accurate and integrated and tested and blah blah, blah. And then secondly, processes by nature are not a single player product. The whole point of a process, if you think about it from first principles, is like you're designing a process so that one can manage many, right? Like that's the purpose of a process. Like you put something in place so that you can deploy that out to multiple other people to follow the process. So as a concept, it's not a single player concept, it's a multiplayer concept, right? So yeah, we basically found both of those components were stopping the natural PLG engine from continuing to drive growth. And so we started to explore other channels, like how can we keep growing and open up other channels. And as we started to look at a lot of the more traditional, non free slash, really cheap organic channels, whatever paid traffic or events or outbound or, you know, any of the normal things, we found that they were really expensive or competing with Oracle and IBM and these companies with unlimited budgets.
And they also have really high acv. Sometimes they're selling their products for tens of millions of dollars so they can burn millions of dollars to acquire a customer, basically.
And none of the unit economics would work with a low ACV trying to use those higher cost channels. And so we went through this journey of figuring out how can we actually drive up our ACV so that we can start to explore and use some of these other channels.
And that was kind of the exercise of, okay, if we want to charge more for the product, we need to understand who we're delivering the most value to. Because if one company is only getting $10,000 worth of value and another company is getting $200,000 worth of value, clearly we can charge a lot more to the company who's getting $200,000 worth of value than the company who's just getting 10,000. Right? And so very basic kind of principle. And then it's like, okay, well how do you do that? And then we started digging into all the data around retention, expansion, use cases, segmentation, et cetera. The thing that we found was that the companies or the teams that get the most value out of processes are the companies and teams that have a strong compliance component to the work that they're doing. So they're not just creating processes because they're Trying to drive a little bit of efficiency and trying to get things organized in a nice way or maybe trying to save, you know, a middle manager having to do some work on Saturday afternoon or something, that's not usually a priority for the business that usually gets pushed under, get, you know, drive leads and sell business and whatever, talk to customers. But for the teams or industries where they're heavily regulated or there's a strong compliance, compliance component to their work. So, for example, HR or finance or legal IT security, where there's a lot more stringent structure, there's external regulators, there's standards like Soc 2 or ISO that they're complying to.
Those teams have to have processes and it's not an option. It switches from a vitamin to a painkiller. Right. And so that was kind of the thread we started pulling on and we're like, okay, so where we notice that the companies, where there's a thread of compliance running through it, either they're a, yeah, they're like a finance team or a legal team or they're say a fintech or a health tech or a prop tech company where they're operating in a more regulated space, where they're interfacing with external regulators and whatever. Those companies show a lot more engagement, better retention, better expansion and you know, are much healthier customers. And then we can, that's kind of what we see from the data. And then as we go in and do qualitative exercises with those customers and really try to understand the impact and the value that we're driving, we can see that from a pure revenue and business impact perspective, we're driving a lot more value as well. So that's where we've started to hone in. And today you'd call Process street, like an AI platform that bundles, you know, work management, so work for automation and project management with compliance. And so we have a compliance agent that is, imagine you just have like a, a compliance person in your company and what they're doing is they're constantly monitoring changes in external regulations. Oh, Trump's like rolling out all these tariffs. Right. So now we need to consider how we're going to roll those into our organization. And that's usually like a compliance person that's watching like the external regulations or they're monitoring the SOC2 standards and the ISO standards. And then they're also monitoring all the work that's going on inside the organization and they're making sure that all the work that's happening inside the organization is actually following those external compliance standards. And so that's what we do. You know, I listed off a lot of the kind of areas that we, that we focus on, but really it's around helping people build, manage and automate processes in areas where they really have to have those processes.
They're critical business functions for either the team or the industry.
[00:13:58] Speaker B: And what things have you generally seen that are maybe ripe for processes or automation in a lot of businesses? And I'll kind of give you some context here. So I'm always surprised that my friends who were talking about work and they'll be like, oh yeah, I'm doing like this and this and this and this thing. And in my mind, often I'll say it, I'll be like, you can totally automate this. Whether it's in a tool or with an assistant or a, like you're just wasting a lot of your brain power on sort of low level tasks. So I'm curious as to what your experience has been and what businesses are maybe often wasting more time on than they should and not freeing up the kind of high level thinking.
[00:14:30] Speaker A: Yeah, I mean, what businesses? I think it's not like a.
Answer the last question first. It's not like there's a particular industry or vertical where it's like these types of businesses are better or worse. I mean, yeah, you know, very heavily regulated, slow moving, where there's a lot of risk kind of. Businesses are definitely adopting things slower, right. Like for example, healthcare. It's like if the risk of an automation breaking is someone losing their life, then they're going to be a lot more careful about automating that and put in a lot more balances and checks and systems and you know, redundancy into those and be a lot more cautious around how they actually implement it versus the downside is like, you know, you misspell an outbound email for your marketing agency or something. Right. Like then you're probably a lot more open to taking a lot more risk. So yeah, you know, smaller, smaller companies that have less risk and less to lose and are dealing with less sensitive information and types of products and services are adopting faster. Right. Bigger, slower, more regulated, more risk kind of adverse industries and businesses are moving slower. That's, that's probably a general, like, that's. But that's the same with pretty much all technology. Technology adoption, right. Otherwise I'd say, let's just say, okay, within the SMBs that can actually adopt quickly where.
Or let's say within tech companies that can adopt quickly. I think a lot of it comes down to More the culture of the business.
I think that, you know, there's two main ways to, to automate.
You can have centralized teams or functions or people that are kind of like builders that are controlling a lot of the automations. And so it's going to be your ops teams or maybe your IT teams or whatever. Those companies, unless they really invest in those resources a lot of the time, do move slower than companies that have more of a decentralized, democratized approach to automation, where they're really enabling every employee, every person in their company to get their hands on these tools. Giving everybody accounts to the workflow products and training everybody on how to use the different tooling and really kind of allowing people to go out and build their own automations that, you know, can create more chaos but can also create a lot more innovation and efficiency.
When you get these teams that are the bottlenecks, they can really slow down and control the automation flows. And once things are built, they can become very delicate and people resistant to changing or tinkering with them because it risks breaking them. And so, yeah, having a more kind of decentralized approach to automation, I think those are the companies that are know, adopting faster in terms of what's right for automation.
Everything can be automated, but it, it, it's gonna. Different things are gonna have different ROIs basically on your business. And so for example, if you're a law firm, that's a really, you know, like you're a corporate lawyer, you probably don't want to automate like your client communication because that's kind of like a really important component of your business. But if you're selling a, you know, a $50 dropshipping product, then you probably do want to automate your kind of customer service.
But we did a CFO roundtable the other day in Palo Alto and we actually asked this question. We were like, what is the most burning? Like if you could, if you could automate anything in your business with AI, what is the most burning thing that you would want to automate? And CFOs are good, good people to ask this too, because they have visibility across all the business units. So they're not just kind of thinking about their own function of finance. They're thinking about, if my job is to drive as much profitability as possible for the business, where could I get those gains anywhere in the business?
And it kind of came back with exactly what you think it would be. It was like sales, marketing people and operations, right?
It's kind of just like the main stacks, sales and marketing I was like, yeah, if AI could get me more growth, right? If they can get me more revenue, that's always the first place to start. Because if you can get an extra a hundred grand in revenue, then you can use that 100 grand to solve like your other problems, right? Whereas if you're, if you're just kind of speeding up some type of operational process and it's not, you know, unless maybe that allows you to fire someone and get that revenue back, a lot of the time you're really just driving efficiency, but you're not really seeing too much of a immediate impact on the P and L. The sales and marketing is where everybody wants to start and then kind of usually it moves from sales and marketing to more of your back office operations. So operations, finance, hr.
The other thing that kind of came out that I think is interesting is that where we're seeing a lot of automation happen first is what you'd call like the level one jobs or the junior jobs. And so you're seeing like your SDR layers or your junior content writers or your junior analysts or your junior graphics designers. Like a lot of those jobs are getting automated away and you're finding that a senior person with like a bunch of AI tooling can kind of do just as good a job, if not better a job because they're better at curating and they have a better eye and editing whatnot than a junior doing that job. So that's kind of where we're seeing it be used the most initially. And I think those are probably the low hanging fruits. But it's very much business dependent depending on where the opportunity lies for your business.
[00:20:13] Speaker B: And you kind of touched on it a bit there, AI automation. So I've been a little bit disappointed in AI so far in the three years since it's been out, two and a half, whatever. I have not found a ton of use for it in my process. Some things that most of them pass off to AI, but either I can do it faster myself or faster handing it off to a human or whatever. So I'm curious as to your thoughts on human versus AI automation. Maybe could you give us a breakdown of sort of the strengths and weaknesses of each and where is AI really excelling as far as automation goes?
[00:20:41] Speaker A: Yeah, I think today AI is excelling as a copilot, right? So in that example where, let's just say you had an editor and a junior content writer, right, in a team before and the junior content writer is writing content and then giving it to the editor and the editor's editing that content, proving it, and publish and sending it out. Like, in that flow, it's pretty easy for the editor to actually replace the junior writer with AI. It's not. It's because that structure is already set up in a way where there's a more senior person doing QA on the work that's coming through it already. It's not like the junior writer is just writing stuff and then pushing it straight out into the wild without any other like, like approval or quality oversight. So things where there's this. There's already a natural structure of like, you've got a junior and then you've got a senior and replacing the junior but still allows because the juniors are almost supporting the seniors in those roles. So the AI is not doing all the work. They're almost becoming like a supporting function in that team. But there's still a human in the loop.
Another example, I think where a lot of people are getting huge gains is work that you may send to a consulting firm. So the core call centers are kind of like the big ones right now, where it's like, maybe you would have hired a thousand people in a call center before and you're paying them 500 grand a year.
Now you can have an AI agent that maybe you only need like the level two and level three people in that call center now when there's an escalation. But like the level one call center can be replaced by AI and then only if someone gets confused, they get escalated up to like the level two or level three, like real humans. So that's happening a lot. Again, it's just like this level one component, because that doesn't require full autonomy. It requires, like, usually there are these kind of structures above the level ones that are ensuring that the work gets done and. And so they're more operating like a co pilot or a supporting function than like a completely independent agent.
I think there's probably some independent agents that are working pretty well in call centers at this point.
And some of the chatbot stuff for customer support, pretty basic things.
I think the knowledge bots are getting a lot of value. So if you hear their stories around, like, JP Morgan and McKinsey have basically sucked in their last 200, 100 years worth of information and internal information in those companies. Right? So every McKinsey case study and interview and project that's ever been done has been dumped into an LLM. And now all the, you know, whatever, 50,000 employees at McKinsey can like go in and kind of ask this LLM information and it's now like digging through 100 years of history. Like yeah, we did these projects with these customers in like these other countries and they connected to your project in some way and maybe you could use them and maybe you should talk to this other person in the company because they were involved in it. Like there's a huge amount of value being extracted out of just like organizing these giant data sets. Especially for big enterprises where, you know, if they're paying a consultant 300 grand a year and they can save that 10% of that consultant's time just on like digging around through SharePoint folders or whatever, that's meaningful. And they can do that across 10,000 employees. That becomes meaningful gains for SMBs. It becomes a little bit trickier. Definitely. Content creation is I think where we're getting a lot of value for it. Oh, and coding, I mean in the engineering teams there's a huge amount of value immediately for sure as well. So I think that's where we're seeing it most right now is in marketing and engineering. In marketing, it's a lot of content creation and sales enablement. So for example, like we do a lot of stuff with call transcripts. Like after the AES go on calls, those calls get like analyzed in different ways. They get packaged into documents. We create like deal docs out of the calls they get sent to customers. You know, they get used in deck generation and proposal generation and things like that.
But it's all kind of working as a copilot mostly where there's still a human checking the work before it goes out. We're getting closer and closer. I think that Google's new kind of set of agentic features is going to be really impactful for SMBs. And I do think that the enterprises are already getting a lot of value in certain use cases. I just heard that Geico was going to lay off 50 to 70% of its employee base and replace them with AI. So it's like there's not. It's like Geico is like an online insurance agency. And so there are these kind of businesses that are perfectly positioned to really get like tons of tons of value out of it. So yeah, I think it's going to be slowly chipping away at the market as the models progress and as the applications become more robust. The applications are still pretty, still pretty iffy in a lot of, a lot of circumstances, but they're making progress and they're moving from kind of SMB to mid market now and I think they'll get there eventually, yeah.
[00:25:35] Speaker B: A couple of the use cases you mentioned, I have found quite a bit of use in so things like engineering or like querying a knowledge base or being able to like chat with a code base and knowledge base, something like that I found like that'd be super helpful. Also kind of sales follow ups or kind of call summaries, that kind of thing. I have found some use in that.
[00:25:52] Speaker A: Oh yeah, Email management is going to be a big one right now. Like you saw, I, when Zapier first came out, I built my own little like AI inbox management system that's you know, mostly just auto categorization. Reads the emails, categorizes them. You saw like Fixer launched, but then now Superhuman just released a bunch of these features built into Superhuman. So I've like shut off. I basically knew this was going to happen so I just built it temporarily. But I've been waiting for the actual email products to build this stuff in. So Superhuman just released auto labeling, auto archiving, auto draft response and auto auto reminders. I think like when you don't like, you know, reminding you to follow up with someone, that's huge. That's like already half of what most people use their Athena systems for, right. Is basically just email categorization and management. And so that's going to get built into everyone's inbox. That's going to be huge in terms of just like the, the raw amount of time it just saves the world. It's also going to be a huge negative hit to outbound, outbound email marketing, which is interesting because you're getting like a whole new layer of auto filtering being built into every product and it's much better than any of the other like older Technologies. Like it 100% finds every app and email and automatically archives it. Right. Like so that's going to be an interesting, I think challenge for outbound email. And then I think the big breakthrough is when Siri and Alexa and okay, Google or whatever Google's assistant's called, actually get something that feels like talking to ChatGPT kind of built in natively into all the devices, then you're going to see a giant explosion across the world.
[00:27:39] Speaker B: Well, switching us back to automation really quick.
So in my life I understood that automation is important, but oftentimes it feels like it's lower priority. And I know that's not true. Like my logical head says like, nope, that's not true. You're being totally crazy. But on the emotional side it's like, oh, like that can wait until the morning. Like I can hire a human to do something that I know I should be like in building into a software or something. So I think my question is like automation feels painful and it's sometimes annoying to do, but it's so important. So sometimes why does it feel unimportant? Why does it feel like a burden when really it should be like top priority? Do you have any thoughts there?
[00:28:17] Speaker A: I think because, well, it's probably not a single answer, but something that comes up frequently and thrown under the bus. But I have this conversation with my girlfriend a lot.
She runs her own business and is very much on the knows she should be automating more things than she does and is constantly kind of doing stuff manually. I think that there's, there's a, an unknown component, right? Like if you don't know the tools, you don't know if it's going to work. If you don't know how long it's going to take to build something, you. And maybe in the past you've, you've spent four hours do it trying to build something. It hasn't got to the point where it really is doing it in the way that you want it. And you're like, ah, screw it, I just have to get this thing out. You just ditch the project and go back to doing it manually.
I think that can cause people to hesitate because they're just not certain that if they invest three hours to try to automate this, that they're actually going to get the thing that they, the product that they want. Right. Maybe they'll take, maybe it will take them 10 hours, maybe it will take them 20 hours and maybe they still won't get something that's fully usable and they just don't know. And so now you're just not certain.
If I invest the time in automating this, I'm going to actually get the result that I want. But you are more certain that if I just do this manually, I am going to get the result that I want. So I think there's this kind of factor of the unknown uncertainty that causes a lot of hesitation.
I think a big part of that is education or experience. They just don't know enough about the tooling and the options to have confidence that they're actually going to be able to build something that can do the thing that they want. They're not aware of all the, the functionality and potential roadblocks that may, that may come up. So yeah, it's kind of a tricky chicken and the egg kind of problem.
This is where I mean, for us, for example, we work with a lot of partners and this is why we work with a lot of partners, because we found that when a customer just comes into the product and they're not experienced with doing a lot of automation, they don't have a ton of experience with different products and haven't built these types of systems and flows before.
They just, sometimes they just, they want to know every single feature in the product before they even start to build anything because they're just worried that they're going to get three months in and then suddenly they're going to hit something that's going to, that they didn't weren't able to like, forecast and see that's going to block them from being able to get, you know, the result they want. And so working with partners who really, deeply understand all the, you know, not just our product, but also lots of other products and how they all work together and making sure that they can actually deliver the result the client wants and give the client confidence has been huge for us, especially as we're selling for a much higher price point.
[00:31:09] Speaker B: Yeah, well, if people want to learn more about you, about automation, about processes, about what you're doing, where should they go?
[00:31:15] Speaker A: Head to our website, process St, sign up for a free trial and also follow me on Twitter @Vinay P10.
[00:31:22] Speaker B: Vinay, awesome chatting with you. Thank you so much for hopping on.
[00:31:25] Speaker A: Yeah, of course. Glad to be here, Brady, thanks.
