How to Boost Your Processes with AI

Vinay Patankar
Jan 26, 2024
Listen Now On
Apple PodcastspandoraSpotifyAmazon Music

Episode Description:

Today we sit down with Vinay Patankar, CEO of process.st, to explore how AI is transforming the landscape of business processes. We discuss the practical applications of AI in various industries, from fintech to healthcare, and how process.st is making these advanced technologies accessible to small and medium-sized businesses. Vinay shares his insights on the ethical considerations and compliance in integrating AI, ensuring it aligns with industry regulations and client needs. Looking to the future, we examine the potential of AI to augment human expertise and transform everyday business workflows. Join us for a thought-provoking conversation on the role of AI in elevating business efficiency and human potential.

Resources Mentioned:

Good to Great: Why Some Companies Make the Leap and Others Don't

Full Transcript:

Andrew Miller:
[0:00] Welcome to another episode of AI Unboxed. Today, we are joined by a former mentor and friend of mine, Vinay Patankar.
Vinay is the CEO of Process Street, which is the AI-powered platform for project, process, and compliance management.
It has helped thousands of leading fintech, asset management, real estate, and healthcare companies reduce risk and streamline operations.
Sounds like a pretty cool tool, but I'm a bit biased, you know, about this.
The first question I always ask, I know I just did an intro, but is there anything else, you know, you'd like to briefly go over, introduce yourself, your area of expertise?

Vinay Patankar:
[0:40] Yeah, I mean, not really, but just great to be talking to you again.
Obviously, we've got some history. You used to run our marketing team back in the day and we had some good times. So it's awesome to be here chatting with you again.

Andrew Miller:
[0:52] Same, same. Same. What would you say attracted you to the intersection of AI and like business process management and everything that you're working on now?

Vinay Patankar:
[1:04] Yeah, so we've been at this for a while working on our process management platform for a number of years.
And we just felt that a lot of the AI tools were just out of reach in terms of complexity and the kind of ROI of what what it would take for us to build on them to what the value we'd get, or we'd be able to deliver to our customers and the, you know, how much we'd be able to charge for it.
And, you know, obviously the, the kind of release of the new LLM APIs, you know, driven by open API kind of changed the ROI metrics a lot and like, encouraged us to explore a lot of the different, you know, paths that we could go down and functionality that we could deliver and really really changed the the ROI and the calculation of how much it would cost how long it would take to build and the value that we could deliver so it wasn't that it was something that you know I kind of have this long history with it we definitely did look at AI for a number of years for different use cases but again it just felt out of reach and so it was really that just the kind of change in the, the infrastructure available that enabled us to start offering that type of functionality to our customer base.

Andrew Miller:
[2:24] Yeah, absolutely. That makes a lot of sense. I know, I think it was years and years ago we were talking, and even at that time, we were talking about GPT, the first models that were coming out.
We were watching some of the videos and stuff like that, but it wasn't easily accessible to a developer and teams.
But with the launch of it, I mean, even just recently, right they um chat gpt just opened up its marketplace and now anybody can go in there and create like an app and like build on top of that but you have your own team there's developers designers there that can incorporate it and build it into your flow and it's just more accessible than it used to be in the past yeah.

Vinay Patankar:
[3:01] Exactly and for us you know it was it was really you know we don't feel threatened by ai it was really a um enabler it was a new type of functionality that that we could deliver and incorporate into our product.
And it's offered a lot of value. Our customers really, really love the stuff that we've built, but it just wasn't accessible for us before.
And so, yeah, it's been a really fun year kind of diving into that and building out a bunch of cool features and seeing how it's being used and the feedback from our customers.

Andrew Miller:
[3:30] Nice, nice. How would you say that AI might be revolutionizing your space?

Vinay Patankar:
[3:37] I would say that in process management, process automation process mining it's been used for quite some time and so we more service kind of the mid-market whereas in the real enterprise spaces let's say like insurance processes they've been using ai for a long time so they do a lot of analysis around predicting insurance events they do a lot of automation around forms and form automation and ocr and that kind of stuff so.

[4:10] It's been revolutionary. Healthcare is a good example as well, right? They've been deploying AI into healthcare.
A lot of like form transcription, data automation kind of work has been done for a long time, well before the launch of GPT and things like that.

[4:26] So it has been deployed into the kind of process automation space for a long time.
I think the big unlock for us as a business is that we're now able to deliver it to the SMB B market in a way that was completely financially unattainable before, right?
The cost of deploying a big healthcare automation solution makes sense if you're a big hospital and you can spend millions of dollars on it and it's being delivered by IBM and that's been around for five, 10 years or whatever.
They've been working on that stuff, but this has really made it accessible to a 300 person company or a 50 person company and allowed them to basically integrate similar types of functionality.

[5:08] So doing a lot of like document automation, forms processing, OCR stuff, helping with just simplifying build time and enabling them to come up with processes that they may not have thought of before or parts of processes that they haven't thought of before.
But probably the most value is actually just incorporating it into workflows itself.
So you're able now to use it to do kind of all these things that help automate day-to-day operations inside these businesses that before just would have been, you know, so expensive that it just wouldn't have made sense for them to deploy it in the past.

Andrew Miller:
[5:50] Yeah, yeah, absolutely. And speaking of that, could you share maybe like a real world example of one of these like traditional processes that are improved by incorporating, you know, AI into it?

Vinay Patankar:
[6:01] Yeah, totally. Totally. So one of our most common processes is employee onboarding.
And through the employee onboarding process, there's a lot of messaging that goes out.
There's a lot of forms that need to get filled out.
And so AI can help with processing all the forms that come in, taking all that data, using it to generate welcome messages, emails, onboarding packs, documents, documents and basically just all that type of work that needed to be done manually in the past you know AI can kind of use it for generating all that stuff.
I think that's a good example another example is just kind of processing information that's coming in from customers so you might have various customer requests, customer tickets, customer kind of change management projects that have to be done and AI can help with categorizing, synthesizing, doing sentiment analysis.

[6:58] Prioritizing territory assigning it can do all sorts of different interesting like kind of, Things that might have taken, maybe you would have to build a huge kind of conditional logic tree to maybe figure it out beforehand.
But even then, sometimes it might not have even been possible.
And it can kind of turn that into a very simple prompt or command that makes it way more accessible to somebody who doesn't have that kind of technical skill and understanding of these complex logic trees to be able to do this type of pretty advanced stuff.

Andrew Miller:
[7:33] Wow. Wow. Well, it basically sounds like an entire ops team that has come in there.
And of course, I referenced the marketing side, right? So you go inside a HubSpot and you say that this person's coming in, right? This prospect or this lead.
And you have to go in there and manually set the different conditions saying, okay, it's going to go to this sequence. If it's there, this is their geo, this is their IP.
This is what we're going to be sending out. Oh, make sure you do this.
And then there's still always like that person or a couple people that slip through that is like, Like, how did they get in there?
But that isn't something that people can just like easily go and set up, especially like an SMB, you know, they don't want to, they don't have the capacity or the resources to hire a full ops team to come in there and set that all up.
Having an AI built into it that has like this basically like pre-logic built it. And then it's like thinking and learning from it.
That that's a huge time saver.

Vinay Patankar:
[8:26] Yeah. So there's a few things like one, let's just say you have someone's city.
And then you're like, you have three different regions, you know, you have one third of America, Central America, and then the other, you know, you kind of say splice up America into three regions.
It's like, you'd kind of have to have this list, this mapping table of like all the cities in America.
And then you'd have to kind of know where all those cities, like what state all those cities are in, and then assign a whole group of states to like a rep or a team.

[8:59] But with AI, you can just ask like, you know, what third of America is this city in, right?
And it will automatically kind of do that routing that otherwise would have been like a very complex table and logic tree you'd have to manage.
So that's kind of interesting, but probably more interesting is like if there's free form texts, right?
So if they've submitted a contact us form and it's like, tell us what you're contacting us about, right?
In the contact form, like tell us your, tell us the legal problem that you're dealing with as like a law firm or something like that right tell us the tell us the website project you want to work on right and it can it can take that freeform text and start to classify it and prioritize it and route it in a way that you would kind of have to have hundreds and hundreds of these weird trigger words right like to kind of catch any type of freeform text that was submitted into that form and try to figure out like what type of problem is this person trying to solve and how How much of a priority is it for our sales team or how big of an opportunity is it for us?
Where should we kind of route this or prioritize this kind of request?

[10:06] And so when the things are coming in in more natural form, it's an email, it's a support ticket, it's a contact us form.
Then it gets really interesting, right? Because it's still annoying to route the cities, but you still kind of have a finite list of cities.
When it's a free form text it's just like can be anything right um even it can be like other languages and stuff and also understand it right someone could submit it in some you know in japanese or something and it will still like know what's going on yeah.

Andrew Miller:
[10:33] That that's always been like the bane of many people's existence they have their contact form in there and there's that one option that's like hey just share your concept inside there like share your whatever and it just goes in a bunch of different directions so you have to sit there and like read it and try to like actually understand it, but having something there that can do that for you, that just reduces the labor, you know, labor costs and workforce there and just allows you to focus on higher strategic, you know, efforts.

Vinay Patankar:
[11:01] Yeah. And then you can take that and you can also use it to generate responses or kind of help you further on in the flow with other things that are going on as well.
So yeah, it can really just help, you know, streamline a lot of these, these workflows flows that your customer support teams or sales teams or professional services teams, et cetera.

Andrew Miller:
[11:20] Are doing. Would you say it's able to like, I mean, it's categorizing these additional like open fields.
You think like making like a database of common themes and trends from that is something that it's also able to do is something that's interesting because then it could, you know, build out these like scripted responses, you know, down the road.

Vinay Patankar:
[11:40] Yeah, once you have that categorized, you can store the category and then you can do analysis on that category. Yeah.

Andrew Miller:
[11:46] Awesome. Could you maybe describe a fascinating AI application that maybe you've recently worked on?
I know like Process Street is all encompassing. So an example from, you know, Process Street would be great.

Vinay Patankar:
[11:59] So, I mean, a fun one we did recently was we did this kind of as a bit of a marketing gimmick to just kind of work with our customers.
Is we built this new year's resolution workflow and it was something that we you know we just did and sent out over new years which was pretty fun where you could kind of punch in a bunch of things you were trying to do like we kind of had two paths you could either say what are your what's your resolution around like you know systemizing your business and building out more more processes and systems for your business or what's like a personal resolution solution that you might have.
And we built this workflow where it's like, okay, I want to systemize my businesses.
It's like, okay, I want to systemize my sales processes.
And it would kind of come up with a whole bunch of interesting suggestions and ideas and an action plan for how you could go about starting to, you know, create a plan for 2024 to, you know, systemize that area of your business that you kind of suggested as the focus.
Or if you did like a personal one, it'd be like, oh, I want to learn Spanish or something.
It would kind of give you an action plan and a bunch of tips and resources on how you could you know start learning spanish in the new year so that was a that was a pretty fun one um internally we have a we have a bunch of interesting ones like let's just say uh.

[13:22] So we do these in our meeting workflows.
So we have workflows for different meetings that we're running and you might have a, so at the beginning you might have, you rotate who's doing note taking, who's the kind of, who's leading the meeting in that for that particular example.
So AI is kind of like picking and routing and saying, okay, you're the note taker today.
You're the meeting taker. It's almost like a dice roller, right?
Where it's like kind of picking who's doing things.
Then it will then it will look at like well the um the different topics that have been submitted it will create an agenda and then like the note taker will take notes about what's going on then it will create action items from the notes it will create a summary from the the notes in the meeting and so it kind of works to really streamline the meeting process and make things a bit more fun and um fair yeah.

Andrew Miller:
[14:13] Nice nice um maybe for the the first example of actually like launching this new year's resolution you know workflow that you sent out there what do you think well what were some of the challenges of actually bringing the application to life or that process to life.

Vinay Patankar:
[14:27] There's challenges around formatting i think that was an interesting one that we had to deal with so uh when we're kind of building out the the responses and if you said oh i want to learn spanish let's say versus you just said learn spanish as your input you would when we kind of of like presented you with the plan, it would say, oh, your plan is to, I want to learn Spanish, right?
Instead of your plan is to learn Spanish.
So we'd have to kind of use AI to reformat all the responses in a way that kind of made everything look really nice when it was spat out into kind of like a final document.
And so I think that's like a really interesting thing because people can input in these fields in all sorts of different ways. They can, you know, not use capitalization correctly.
They can add extra fields here or there. They can add filler words.
They can kind of type things in it's a lot of like kind of reformatting that freeform text in a way that looks really professional and is formatted cleanly and like outputs into a document in like a really structured clean way um i think that was one of the challenges uh just yeah just just messing around with the prompts so just trying to get the thing to have more actionable outputs and and clean outputs versus kind of flowery, long paragraph-y kind of sentences.[15:50] Yeah. But not too much. It's pretty, it's pretty easy these days to, to use the, to use the tools. Yeah.

Andrew Miller:
[15:57] Your, your first example about formatting, I feel like I've, I've heard that so many times in like different use cases.
Uh, so an example, like using one of these image generators, right.
It spits out text that even if you tell it to put text, right.
It adds like an extra letter here, or it's like mixed all around there.
And people that are not graphic designers, they're looking at that and like, oh, okay, maybe I'm going to send another prompt and say, hey, remove this, hey, remove that.
It doesn't, it just gets worse and worse as you keep going through there.
So you definitely need some of that human aspect brought on, layer on top of like these models to make sure that things do come out correctly.

Vinay Patankar:
[16:35] Yeah, totally. And that's kind of one of the things a lot of people like about using AI with Process Street combined, because we have these these um kind of approval flows built in so as you go about and the ai is doing stuff it's writing emails it's writing messages it's creating documents for you we can we put in these like approval flows so that you always have these humans well you have the option to basically have humans need to double check something and approve it and they can kind of regenerate it before anything goes out to a customer or an employee whereas when you're just kind of like you know have ai write this email and automatically send it then uh it's a bit of it's a bit of a The diastrol, yeah.

Andrew Miller:
[17:13] Uh-oh, uh-oh. It's like, I don't remember saying that. Oops.
I know that you work with, you know, compliance management and you're getting into like areas like healthcare.
And one of the big questions that always comes up around AI is, you know, ethical considerations.
You know, as you're working on, you know, rolling out, I mean, Process Street's rolled out, but incorporating more AI functionality in there.
Are there specific, you know, ethical considerations that you're considering whenever you're building, you know, developing this work.

Vinay Patankar:
[17:45] Um, I mean, we just, we just follow whatever the compliance standards are, right?
So there's like a number of things like say HIPAA for healthcare that we, that we follow.
And that basically builds in a lot of that stuff.
So there's certain things that you can't do. You can't share, you know, important patient information or sensitive information about like people's medical conditions and stuff like that.
And so we have controls around things like that. We also have have controls around like if people actually just want to turn off ai and actually not not use it at all um in that or not allow anybody to kind of use it at all in in their in their um organization, um but you know it's a pretty like our products are pretty like flexible products right where you can do a lot of things and we just give you those building blocks and we can't really manage manage everything right um so yeah we do we just we just try to make sure that especially with things like health care or financial financial teams that we're giving them basically the controls that they need to basically manage how they want um the the system to perform right so whatever their kind of ethical standards are we give them the controls to basically turn those those things on and off as they choose as a customer.
We don't kind of make those decisions for the customer. Yeah.

Andrew Miller:
[19:09] That's super smart. That makes me wonder, you know, a lot of companies launch and then they're building AI natively into like their entire functionality.
So there's not that option to switch things off because then their product won't function at all.
How did you go about doing that? Like from the get-go, you were like, hey, we're going to like keep this separate.
They can turn a toggle, it turns on or just, can you walk me through that a little bit?

Vinay Patankar:
[19:34] Yeah. I mean, firstly, I think that they'll probably struggle to sell into certain industries if that's the case, right?
So there are lots of industries that are not heavily regulated and they're probably going to have a lot more flexibility.
Tech startups are probably a good example of that. Maybe not FinTech or health tech, but just kind of like normal tech.
And that's going to be fairly limiting to the type of market that they can go after, but that's fine.
There's plenty of opportunities in those industries as well.
Our platform definitely, we find, offers a lot more value to customers that have a regulation and have processes that they have to adhere to to kind of meet those regulations and stay in compliance.
And so that's a lot more important for us. But also, because we weren't AI from the beginning, we'd already built out a very robust solution before we implemented AI.
AI, when we added it in, it was just a kind of module on top of what we already had. It wasn't part of the core infrastructure.
So that made it easy to kind of silo it out and create kind of controls and permissions around it.
So, yeah, because we'd already built a lot before, it just kind of was pretty natural for us. Nice.

Andrew Miller:
[20:50] That's smart. I know you're working, again, in these highly regulated areas.
Is there maybe a specific ethical dilemma that you could share that you faced while maybe moving into that space or incorporating AI into it?

Vinay Patankar:
[21:07] Hasn't really come up a lot besides that these companies have these certain regulations that they have to adhere to and we just make sure that we give them the capability to adhere to those regulations and like i said a lot i think a lot of that's built into it but um again it's like the product can be used in so many different ways that we're not really making those types types of decisions in the product where like, here's the kind of toolkit, um, here's the options that you could use.
If you need to follow these certain regulations, then these are the options that you need to, you know, turn on or turn off.
Um, but besides that, we're not really making those decisions.
You know, it's just like, it's like a word doc or something, right?
It's kind of like, oh, are you going to control like what people write in that word doc type of thing? Um, and it doesn't really, yeah, it doesn't really make sense for us to police like what people kind of put in our, in, in, you know, in our product that way. Yeah, absolutely.

Andrew Miller:
[22:09] And I think you touched on it earlier. I mean, you thought about how you were going to market with this, you know, you built in, you understand your, your customers and who you're working with.
And so you, you knew that there's already regulations in those industries.
You went and researched those, you understood there's like HIPAA compliance or GDPR, there's SOC 2, there's all these other things. so whenever you were ready to do it, you're like, okay, we'll build this in.
It's great. It's powerful, but...
They might have pushback in these spaces about AI, especially since it's so new.
Let's give them the ability to not have that be a major objection or blocker.
And if it does come up during like a conversation, it's like, hey, you can just toggle it off. And they're like, oh, okay.
So you can, you can move forward. So I think having that foresight before, you know, building it out and thinking through it made all the difference.
And that's probably, that's why you're not, you know, faced by all these ethical dilemmas. It's like, oh no, we need to roll this back. Oh no, we need to do that.

Vinay Patankar:
[23:03] Right. Right. Yeah, I think that, yeah, exactly. And then also the product already has a lot of these controls in place.
Like I talked about, we have these approval flows and things like that, where you, you know, if you are concerned about AI doing this or that, we already have all these kinds of controls in there where you can say, oh, we'll add an approval step, add three approval steps, right?
Like add your, make sure your legal team approves this. Like we can kind of control all of that.
We already have that kind of functionality built in.
And, you know, a lot of the other things like, like training data and stuff, right?
Like if you're going through the API, it's not using any of your data for training purposes and all of that kind of is private.
And so a lot of that's also thought of by, you know, the providers as well.

Andrew Miller:
[23:47] Yeah, yeah. And I know I've talked to other guests on this in the past and the big thing is it didn't, the like ChatGPT wasn't always like sharing that, hey, you shouldn't put private information in there or these LLM models.
It's like, hey, we're actually using this info that you're giving us and building it into our training models.
And so there were a lot of issues there. Now there's those notifications saying, hey, don't put Alec Baldwin's number in here and contact him.
He just popped it in my head because I read an article about him.
But don't don't put all his info in there because, hey, this isn't private.
And we're putting it even though we try to neutralize it, it's still going into a larger data set.

Vinay Patankar:
[24:25] Well, yeah, I mean, from my understanding, everything that's going in through the APIs.
So if you're basically paying, then none of that's being used for training, right?
It's only the free users that are basically contributing to the training models.
And that's kind of one of the benefits of paying is that if you're paying either for GPT Pro, or you're paying through the API for each of your calls, then you know, you're paying for privacy, right?

Andrew Miller:
[24:48] That makes sense. Now, taking a step back and looking at a larger, maybe not full 10,000 foot view, but still a higher view, I know you're in B2B, SaaS, in tech, you're very, very plugged in.
What would you say is maybe a groundbreaking yet underutilized AI tech, technology, app, whatever that is being used in the space?

Vinay Patankar:
[25:16] Yeah i think there's a lot of really so i mean there's gonna be a lot right it's gonna be a huge huge amount but i think that where you're gonna see it offer a ton of value and like really large companies get built is um the very the very like.

[25:36] Kind of core business processes so hr accounting sales like those are going to be the ones that there's going to be these huge huge things built i recently um invested in a company that is basically rebuilding quickbooks ai first and so it's this kind of like it's like a finance product right for accounting and finance so it's like quickbooks and payroll but it's just everything is AI first so you know all your expenses are automatically run through AI and categorized and posted all of your forms are like automatically sucked in through AI all your receipts are sucked in through AI like just every single thing just like all the little bits and pieces that like a bookkeeper would need to do or an accountant would need to do all your kind of tax documents and tax filings are automatically organized and like set up for you for the year It's just like every little bit of the process of managing like the finances for a business, like every single step is kind of just getting pushed in through like an AI kind of module, essentially.
And you're seeing just, you know, what used to take a team of people, you know, a whole year to do.
It's like someone part-time can just kind of like do all the finance work for a business because it's just so many of those little individual steps just get fully automated or at least like 80% the way automated.

[27:02] I think you're just going to see that kind of happen through these main functions of a business where you used to have, you know, in a finance team, you might have like a payroll, accounts receivable, accounts payable, bookkeeper, an accountant.
Content that can all just get like squished into like one person just managing an AI platform, um I think you're going to see lots of value in legal like I think soon you're going to just, like a lot of things that you would tap your legal team for you're just going to be able to talk to an AI bot and get a lot of that value um a lot of like document automation and contract automation and that kind of stuff um, But it's going to be it's going to be everywhere. Yeah.

Andrew Miller:
[27:43] That makes a lot of sense. I think I might trust like an AI legal bot a little bit more than sometimes an attorney, just because it has the full data set behind it. And it can give you that specialized knowledge.
Of course, there'll be those prompts in there that says, hey, I don't actually know that, which is what the attorney says anyways.
But, you know, it'll have a larger, you know, database that they can like reference and say, hey, this is what it says right now.
And it's like 100% accurate. it so that's very very very cool.

Vinay Patankar:
[28:11] Yeah i mean doctors too right like you know even before chachi bt like watson was you know ibm's watson was was more powerful and could outperform a lot of doctors and it's just like you know a doctor that's like every every day there's hundreds of new you know research papers and studies like published right and just an ai that has literally literally read everything ever and every day it reads every single new thing that gets published, right?
Like, you know, probably hundreds and hundreds of hours of like reading that it's just doing every single day based on new data that's coming out. Um, You go and you see a doctor that's whatever in their 50s or 60s, went to med school 30, 40 years ago, has just been cruising for the last 20 years with his patients.
Surely that person is not as up to date on all the new cutting edge science that's coming out.
You know maybe the very top like doctors and the top hospitals that are the highest paid and that the you know the billionaires go to or whatever like maybe they'll be able to compete but your average doctor in an average city that's kind of like just you know being cruising in their local practice for the last 30 years or whatever it's unlikely they're going to be able to outperform an ai right um so for sure.

Andrew Miller:
[29:33] For sure probably probably like that the median ai uh you know knowledge base and everything will outperform the vast majority of people out there just because of like the wealth of information you can hit.
Like you said, there are always those outliers for sure, those people that are really like the top.
And then you have to also believe that like hallucinations and other factors that don't lead them down weird biases have been rectified or kind of like built out a bit so that you don't get some weird answers and responses.
But as we progress, I mean, it's rapid, rapidly changing and getting better and better every day.
So that's, that's soon. That's, it's very, very, you know, soon to come.

Vinay Patankar:
[30:13] Yeah. Like you were saying with the, the image, with the text, right.
You know, now you get some, some text responses that are a little bit weird.
They've got an extra letter here or there, but like six months ago, you couldn't do text at all.
And so it's not going to be long before that text is perfect.
Right. Like in five years or something, right. There's gonna be zero errors in text. Yeah. Oh.

Andrew Miller:
[30:31] For sure. For sure. Sure. Well, looking at this same, you know, train of thought, where do you think maybe AI is falling short in like the tech field or healthcare or whatever you want to go down?

Vinay Patankar:
[30:44] Yeah, I think that one of its biggest problems is it still just doesn't really understand things well, right?
Like it's still very far away from a human who can fully like understand a concept.
It's you know and and tackle it in like an innovative new way um that it's not, where there isn't kind of like training data somewhere for it to reference um, and i also think that it has a big challenge in.

[31:20] Um, synthesizing ideas in a way that really draws out the most important, um, kind of most impactful parts of that idea.
If you ask it to like summarize something, for example, it might give you like a general summary, but it might actually not include the most important points.
This, this kind of surfaces a lot in marketing, right?
If you're like, Like, summarize this kind of web page or this white paper and put it in a sales email for me.
It doesn't really know how to pick what are the ideas that are going to sell the best, right?
What are the ideas that are going to actually convince somebody the most?
It's going to give you more of this kind of generic summary or it's going to summarize each paragraph into like a shorter paragraph.
But it's not going to really like pull out the pieces that are really important.
Important um to say write an email or for like copy on a website right like ai is very far from being able to outperform the best copywriters right because it gives you it gives you average copy it doesn't give you the best copy um and so yeah being able to really kind of like understand oh what are the you know value triggers or emotional triggers that you're trying to like hit when you're say writing a sales email or crafting the copy for a landing page.

[32:42] Or even, you know, creating a presentation or whatever you're trying to do, doesn't really know how to kind of pick those ideas that are going to have the biggest impact and convert the best or drive your message or kind of anchor emotionally with somebody.

[32:58] So it's kind of just doing this generic kind of reformatting of text without really understanding the point of what you're trying to do.
And so I think there's a lot of room for improvement there, yeah.

Andrew Miller:
[33:13] Absolutely. And I guess that's where the human-led has to come in, right?
You have to look at AI as an assistant to help you there.

Vinay Patankar:
[33:23] Co-pilot.

Andrew Miller:
[33:23] Yeah. It's a co-pilot, right? Right. And then if you really like drill down on these like LLM models, they're really just looking for the best next word that makes sense whenever they're pushing it out there.
So they're like, oh, that makes sense. That makes sense. And it's it's pretty good.
You know, it gets pretty good, but it doesn't, like you said, get to the heart of a lot of these pieces and you lose a lot of details.
I've run so many tests where I'm like, keep all the core details of this.
Do not remove it from this.
And it still gives me like this weird generic thing that I'm like, okay, you kept a part of it, but that's still, that doesn't work for me.
You know, I need to go through it again.
So I think that's a really good, you know, shout out that. Yeah.
People are saying that copywriting and content marketing and all that is like going to get taken over by AI, but you really need that human side of it because a lot of this stuff is just, you know, it gets lost and it doesn't hit the mark.

Vinay Patankar:
[34:18] Yeah. Yeah. No, like look, average copywriting will get taken over, right? You need to write like an average email.
You need to write like an internal document that's getting circulated.
That doesn't really matter how well that performs.
Like, sure. But you, you know, your job, you know, you want, you want it to craft like a Superbowl ad script or something, right?
Where like, you're, you know, you're, you're, you know, the, the cost is so high, you're going to get a better result by like finding the best agencies with the best copywriters and the most creative thinkers and a lot of the time you know you're doing something where.

[34:51] No one's ever thought of it before. If you remember that Coinbase Super Bowl ad where it was just a QR code that was bouncing around, it's like, hey, I wouldn't be able to. Now, it probably could because it's been done before.
But those kind of innovative ideas, that's where it really falls short.
And what's the other, like the Einstein quote, right?
Where genius is being able to take a complex idea and explain it in a sentence or explain it to a five-year-old, something like that, right?
It's not very good at that. It doesn't have that kind of genius.
What it can do is it can take someone who's done that and take that one sentence and turn it into a three-page essay, right?
Like, it could do that well, but that's kind of like the opposite of genius, right? That's kind of like extrapolation, right?
So, yeah, it doesn't have that kind of ability to really, like, hone in on these high-leverage ideas and, you know, copy and sentences and value points.
And so I think there's still a lot of opportunity there. Yeah, but it can help humans brainstorm that stuff.
It can help humans eliminate things.

[35:55] So it's definitely, you know, it's definitely got its place, but still a long way from being actually a human and having actual intelligence. Yeah.

Andrew Miller:
[36:05] I think a lot of our listeners might have instantly got the image of themselves telling ChatGPT, shorten that, shorten that, please shorten that.
You know, because that three page essay that comes out there from this little thing, it's like, no, no, no.
And so that's just like, I guess, first world problems, right?
We're going in there just please make it shorter.

Vinay Patankar:
[36:27] Yeah. Yeah. There's a bit of a risk in that too, right?
Because then it might like not include the most important things in that shortened version. Yeah.
So a lot of the time you need to know, like, these are the three value points we need to hit because we know these are the three things that resonate most with our audience or our customer. or whatever, and then you're like, make sure you include these, but make it short.
But if you don't give it that instruction of like, I'm telling you, these are the most important things because we've tested this aggressively.
We've, we've spoken to a hundred of our customers, like whatever, right?
Like I've just, I've read all of David Ogilvie's books and I understand like the power, the power words that need to be used or whatever.
Like if you don't kind of like give it that type of direction, then it doesn't, it doesn't hit the mark.

Andrew Miller:
[37:12] A hundred percent. I mean, and that's, that's why it's an assistant.
And that's why we have to think about that as like, it is a great assistant, but you can't let it like just run the whole thing for you because then it just starts going into like mediocrity and the stuff that's shipping is just, yeah, if you just want to be like everybody else and never stand above the crowd, sure.
Sure, you know, send it, but giving it that guidance and that direction is what makes it powerful and saves, you know, true like marketers and everybody else, you know, a lot of time because you can like guide it.
And then you take that, like you said, as, as guidance, uh, for moving forward. So.

Vinay Patankar:
[37:45] But there's, there's still lots of like places for it, right.
Where, you know, you're like, Oh, okay. You're a office assistant at like a dentist practice.
And you're going to to write an email with like here's like a summary of the the the you know the services you received or the types of work that you had done and this is like your next check-in date and like you know thanks thanks for being a customer right like that doesn't need to be this doesn't need to have like some expert copywriter like write it right it's just a very basic kind of transactional email and for those types of things it's great right like it's it's taking labor off like a human who who was just doing a kind of mundane kind of repetitive task and it's, it's, you know, it's, it's really freeing them up to do more of these kinds of creative, important tasks.

[38:33] And so, yeah, it definitely has its place. And there's lots of places, you know, I think with the market has had on, it's maybe there's a lot of, it's not as useful, but for a lot of these just like operational jobs, right. It's, it's super valuable. Yeah.

Andrew Miller:
[38:46] Absolutely. Totally, totally agree.

Vinay Patankar:
[38:47] That's actually where we see a lot of our customers use it, right.
It's more in these like operational processes where there is a lot of value you because like the the the interaction or the the job that's being done is not this super high value job it's just kind of this more repetitive mundane job that a human's doing and it can really like lighten that load a lot yeah repetitive.

Andrew Miller:
[39:07] Tasks right that's why we've created a lot of like services and tools out there it's to help reduce those repetitive tasks that a human could be not well not have to do they could focus on higher value items so cool man uh if we if we look a little bit towards the future, right?
What emerging AI trends would you say you're most excited about?

Vinay Patankar:
[39:34] Good question i think one of the things that we've been waiting for is for gpt to kind of, bring in a lot more image analysis and document analysis into its api so that's something that we're looking to implement into the product but you know add in any image add in any upload any scanned form or a photo of a form and help like will help you process that right so say you're doing like a real estate inspection and you're just going around and you're like oh there's like a leaky faucet here and you take a photo of it and you upload it like leaky faucet in the bedroom and like fill out all this data for you right like from from the photo or if you yeah you upload a pdf or a form that's been like written written out with a you know pen and paper process all that ocr it organize it categorize it like so i think that there's a lot of interesting stuff with kind of.

[40:28] Computer vision basically and integrating that into these workflows and there's a lot of people people working on that right now but um i don't know if you've seen like meta's meta just released like the an update to their their ray-ban glasses but you can kind of just like so you can just like what you know you're wearing the glasses and they've got camera and like a headphone kind of built in and they they kind of integrated an ai they've got it's called llama that's like their their their siri and their llr model um you can see it in like if you're in you know messenger chat or instagram chat or something you can kind of like query the the ai now um but you can just say like walk up to a plant and then look at it and say like tell me how to it's kind of like a half dead plant and it's like tell me how to fix this plant right and it will be like oh it looks like you're looking at this type of plant it looks like it's suffering from like not enough water or it needs like some extra calcium or something right and it was just like so you can just kind of walk around the world um you know accessibility is amazing right like you could just be if you're blind you can just walk around and it's like it can tell you what you're looking at everywhere um, But, uh, yeah, kind of just being able to, to interact more with the world through photos and videos, I think is something that I'm really excited about. Yeah.

Andrew Miller:
[41:43] Yeah. Yeah, absolutely. I'm not sure if I'm thinking about that one or maybe Google launched something similar where they had like the Google, like, uh, like sunglasses, like with lenses that were like tapping into what you were seeing and then it's like giving you information and then you can like start searching and stuff like that.
It might've been made meta. Meta. I can't remember, but I did see something similar to that.

Vinay Patankar:
[42:04] Yeah. I don't know if Google launched one recently. They had the old one, Google Lens.

Andrew Miller:
[42:08] Yeah. The old one. Yeah.

Vinay Patankar:
[42:09] Meta has just like, they're like, it's like a partnership with Ray-Bans.
If you go into any Ray-Bans store, you can buy them now and they just look like normal Ray-Bans.

Andrew Miller:
[42:16] Super interesting. Very, very cool. How do you see the relationship between AI and human expertise evolving, you know, over the next few years?

Vinay Patankar:
[42:28] Yeah so i think basically ai will replace the bottom half of performers in most knowledge work so if you're you know below average designer below average programmer below average lawyer below average marketer right like then your job is going to get replaced by ai and i think it's going to supercharge the above average workers so they're going to be able to do the job of three or five people and they're going to be able to have higher quality outputs because they're utilizing these ai co-pilots to kind of handle a lot of their mundane work but they're adding on that layer of human creativity filtering the results guiding the ai in a way for it to help kind of them uh supercharge their work right um so i think that yeah you either yeah you need you You know, if you're kind of just like someone that's coasting, doing the minimum work, and, you know, hasn't really put in the effort to master your field, I think that you should be concerned.
I think if you're a top performer who's really passionate and is constantly learning and innovating in their field, then you should be excited because you're going to be able to accrue a lot more value because you're going to be able to offer a lot more value by using these AI co-pilots.

Andrew Miller:
[43:48] 100%. 100%. I think that's a really good shout out. I mean, especially if you just niche down into the tech field with everything that's been going on, you know, over the past, you know, couple of years with the mass layoffs with everything like that.
I know that there's been like reconfiguration and there's been some, I'm not saying there haven't been some amazing top 1%, 10% talent that were let go that happens with like the reorgs, but they, those top percentage have an opportunity now to like.
Do even more and like find a better, you know, fit for them and everything around there.
So I think evolving and building these into their own processes and internal workflows is just going to elevate them to new heights as opposed to those people that are just like, well, and it's not the government, right?
Well, I'm here at the government. I'm gonna be here for like a hundred years and then I'm gonna get a gold watch and it's over. It's like, this is, you know, capitalism.
This is the space that we're in that it's all now, especially on the VC side, we need to see the roi we need to see the money that's being returned to us and it just gets tougher and tougher so you have to be.

Vinay Patankar:
[44:47] A high performer.

Andrew Miller:
[44:48] And utilize these tools that are around you.

Vinay Patankar:
[44:50] Yeah not just the vc side any public company that has shareholders that they have to report to right um so you're you're seeing like google's just like slashing right now and it's just replacing people with ai like already right like it's just going it's just going to be more and more until these companies just get super, super tight. Yeah.

Andrew Miller:
[45:07] For sure.
Let's look at some like personal insights and maybe like lessons just in general.
What's a lesson that you've learned in your career that you wish you knew earlier?

Vinay Patankar:
[45:20] The thing that I wish I had done earlier and I'd learned earlier was just to start my entrepreneurial journey faster.
So I think that I didn't start building websites until I was kind of in my, you know, mid-20s, and I just feel like if I was on that from when I was 16 or something, then I would be way, way further ahead.

[45:46] That's always something that I, that I wish that I'd done. And I feel that kind of like once you're in that entrepreneurial mindset, you, you just unlock things a lot faster and you learn a lot faster.

[45:59] And I think there's also, and it's not even necessarily like starting your own business.
Like for example, the first thing that I think for me that was really like an entrepreneurial unlock was working in sales where I almost just had to manage my own desk.
Desk right and it was kind of like running a little business in and of its own i think it's, getting out of that and you know once you're once you're like higher up in an organization, you you are entrepreneurial because you have your budget and you have your you have your targets and it's kind of like a lot of the responsibility rolls up to you to kind of figure out how to make things work but if you're just like you know working at mcdonald's or working in like a as a you know, in a job where you're not kind of using any of those entrepreneurial muscles and kind of decision-making muscles, right.
Understanding how to place bets, how to allocate resources, understanding how to prioritize, understanding how to like drive a target forward.
Then you're just, you know, you're, you're way behind somebody who's kind of getting that, those learnings earlier.
And so I think I would try to get myself into situations earlier where I was kind of exercising those muscles, whether it was starting a side hustle or whether it was trying to get into something that was more like a sales job or something earlier on where it was more up to me to figure out how to make it work versus me just like taking orders and just kind of doing what I was told. Yeah.

Andrew Miller:
[47:24] No, I love that. I love that. And I think even if you're not going going to go out as an individual and like start your own business, understanding the other levers that pull or move an organization forward is crucial because then you can take a more holistic approach in your work.
You can understand the different components in there. And you only know that if you push yourself beyond your comfort zone, like you were saying, I think it's, what is it?
Noah Kagan has like a challenge where he makes everybody go out there and try to negotiate on something, no matter what it is.
You go there, you try to buy something, you have to negotiate it down and you do that multiple times just to get you beyond that comfort level.
And that gives you a different perspective. And it's like, oh, that can actually happen. So you have to push that muscle and grow.
But I think that's a really, really good insight that anybody can apply.

Vinay Patankar:
[48:17] Yeah. That actually reminds me of another lesson that I could share, which is, getting closer to like getting closer to the people who are doing the things that i really want to do like another lesson was like um or another example is when i was growing up in australia i like would only read about say tech entrepreneurs you know in, the financial times or in the economist or something you know in these kind of big magazines, and the only people that kind of make it into those magazines are the public company ceos and whatever right and it just like the gap felt so big between where i was and like the types of people who i thought were like kind of the entrepreneurs running businesses but once i.

[49:08] Got once i traveled to america and i started going to some conferences and i started like meeting meeting people in person that were more closer to my age, had much smaller businesses.

[49:20] And I just was able to kind of like sit down, like, so, you know, and ask them questions and just like physically kind of experience, you know, in arm's length, like people who were doing things that I was aspiring to do, it made it a lot more attainable. Right.
And so kind of staying in your, this is kind of part of that maybe getting out of your comfort zone, but, or what you mentioned is like, once you, you once you start to kind of see it right once you start to get kind of closer to it and see that it's physically possible it kind of changes from this idea that's like in a movie or something, right to more reality and i think this is why people who may have like entrepreneurial parents or whatever kind of tend to get a big bigger head start because like they're so they're so close to it their whole life right but for most people you're just told like oh go to go to school get get a job like you're you know or you're just like sitting on your computer all day and you're not like going out and kind of like interacting with with these these types of people in the real world that can you know even if the gap is not very big maybe all you need to do is do something make a small change because it's you're not like physically close to it it makes it just feel a lot further away and so trying to actually physically get close um whether that's going to events or going to meetups or finding a mentor or like anything like that, I think can really make a huge difference.

[50:42] And I wish I did that earlier as well. Yeah, absolutely.

Andrew Miller:
[50:46] I mean, I guess it's that believability effect, right? You...
When you're actually there and you see somebody else can do it, it's like, oh, I could do it as well.
And you see that this is something that's possible as opposed to just in like the financial times and these huge, you know, people that we build up in our minds that are just like almost, you know, like a whole nother level of person.
I wanted to use an example. I think it was like Malcolm Gladwell talked about like that nine minute mile, that person that like forever, I don't remember who it was, but like forever couldn't break a mile.

Vinay Patankar:
[51:17] Yeah.

Andrew Miller:
[51:17] Yeah. The four minute mile. There we go. Couldn't break it. And then all of a sudden the person broke it.
And then like the next year, like five people broke it. And then the next year, like a bunch of people broke it. It was just like, Hey, it's possible.
We can actually do it. But like for all these like years, nobody could actually do it.
And it was just that believability factor that, Hey, I can do it.
It's a mindset shift. So I think that's huge, man.

Vinay Patankar:
[51:36] Yeah.

Andrew Miller:
[51:38] Um, are there any, maybe like major reading recommendations or resources that you would recommend people like check out maybe AI related, or just like generally that you're like, these These are my go-to and you should probably check them out as well.

Vinay Patankar:
[51:52] Yeah. One of the books we've been on a big strategy blitz this year.
So my favorite one from like 2023 was Good to Great.

Andrew Miller:
[52:03] Okay. Yeah, absolutely.

Vinay Patankar:
[52:05] And that was a really great book around trying to kind of dissect what makes a business great and how to kind of achieve that through very considerate focus and prioritization.

Andrew Miller:
[52:21] For sure. That's a classic. Everybody should definitely read that.
I mean, Jim Collins knows what he's talking about.
So let's see, wrapping things up a little bit here.
What would you say your moonshot project is? And this could be AI, you know, into process street, but your moonshot project for the future? Sure.

Vinay Patankar:
[52:40] I wouldn't, I don't know exactly what the project would be, but one of the areas that I'm really passionate about is kind of unlocking human potential, like, like, like human kind of brain potential, basically.
So I think that humans are the most interesting, you know, valuable things that we can observe in the universe, basically.
And i think we have this and the reason that humans are most interesting and valuable is like we have this brain that basically can create explanations create art create knowledge create technology and we're the only you know thing that we know of that can do that right.

[53:28] And we have this limited capacity of kind of brain cycles right we have whatever it is seven and a half billion people they're awake for 16 hours a day they can probably pay attention for six to eight hours a day that's kind of like your finite cap of resources of like this kind of resource that's arguably the most important resource right like everything that around us kind of comes from this resource right like we'd all basically be dead we wouldn't have food we wouldn't have shelter we wouldn't have medicine we wouldn't have anything if it wasn't for this brain cycle going on right um yet so many humans are trapped in ways and this can kind of tie back back to AI, but so many humans are trapped in situations where they're not able to use like their mind properly.

[54:09] Um, they're in a war zone. They don't have hunger. They don't have food.
They don't have education.
They're trapped in a, in a, in a labor job that where they're not using their brain, they're using their body.
And, um, you know, I think AI is a great example of how we can kind of free up more of those minds to solve the most important problems of humanity and enable these kinds of brain cycles to, you know, if we had, if you think about like.

[54:33] Let's say um kind of einstein's era which whatever 100 years ago 70 80 years ago or something right when he was born um you know how many people had college degrees how many people had um education, up to like advanced levels in math it was probably i'm just guessing i have no idea but you know three percent of the world five percent of the world or something right so you if you could find one einstein when five percent of the world is kind of educated in that way you could think that if If you could educate the entire world in that way, maybe there'd be 20 Einsteins running around at any given time.
And then imagine the kind of like impact that you could have by having 20 of those people kind of solving problems at once versus just one, right?
And so trying to kind of unlock more human potential that way is something that I'm very passionate about.
I don't know. It's a very big, you know, there's a lot of kind of different like directions you could go with that and different types. You know, there's lots and lots of problems that are basically blocking that, right?
So I don't know exactly what the project would be, but I think kind of pushing that forward would be valuable.

Andrew Miller:
[55:39] No, I think that's beautiful. And that falls right in line with the moonshot, right?
I mean, this is a big idea that can make a massive impact on the world.
So I love that. I mean, as you develop that, always hit me up.
I'd love to. I mean, Masters of Behavioral Economics.
That's so cool. I love where you're going with that.
How can our listeners follow your work? And you?

Vinay Patankar:
[56:00] Yeah, so the best place to check out, you know, our company is Process Street, process.st, or you can follow me on x, Twitter, at VinayP10.

Andrew Miller:
[56:10] Perfect, perfect. Well, I know that I learned a lot from this conversation, and, you know, it's always a pleasure talking with you, Vinay.
I'd love to leave, like, any last words of wisdom that you have for our listeners that you'd like to relay.

Vinay Patankar:
[56:26] Yeah, don't be afraid of AI. You've got to embrace it. You want to be one of those above average performers that's, uh, that's crushing it.
And, uh, I think if you don't, you know, you're going to be left behind.
So I wouldn't be afraid. I would be diving in headfirst.

Andrew Miller:
[56:39] Absolutely. Absolutely. Well, thank you so much for your time.
Appreciate it. Um, and we'll, we'll catch you later. Take.

Vinay Patankar:
[56:45] Care. Thanks. Great catching up. Good to see you.

Andrew Miller:
[56:47] You too.