Ritam Gandhi: Imagining your AI native competitor

Pauline Bertry 0:13

Hi Rita, welcome to the Dan, how are you?

Ritam 0:15

Very well. Hi, Pauline, all well, how are you?

Pauline Bertry 0:19

I'm well, I'm well. Thank you so much for making time.

Ritam 0:22

My pleasure, my pleasure. Thank you for having me.

Pauline Bertry 0:24

Yes, I appreciate it because I mean it's always easier, of course, when you start and to kind of invite people who you know. And I'm very grateful that you found time out of your busy schedule to talk to me, even though we've just met a couple of times before that. So thank you so much.

Ritam 0:42

My pleasure. I enjoy these conversations and it's a very interesting time in our world. So it's it's nice to share and brainstorm how things are changing.

Pauline Bertry 0:51

I I agree that that's I mean, exactly why I'm doing this. But then I mean, I think it would be fair to to start with with a story. I would love to ask you to to tell me one more time and everyone who is listening to us about like your past, your present, maybe some thoughts on the future, like what are you up to, what's inspiring you, what's taking your time now, etcetera.

Ritam 1:16

Yeah, for sure. So a little bit about my past. I moved to the UK when I was thirteen, went to boarding school, did a gap year. So I'm going quite far back here, not even just like career. I took a gap year, traveled around, worked in a pub, which was probably one of the best years of my life. It was great fun. Went to university in London and then joined Accenture because I didn't really want to work for a bank, which is what everyone was doing around the time I graduated. But the first project they put me on was for a bank. So I

Pauline Bertry 1:45

Ha okay.

Ritam 1:46

I ended up working for banks. But the nice thing was I essentially had the same boss for about ten years. I moved to another consultancy Capco with him and I went through this phase where I was really almost had this FOMO around the fact that there was a startup growth happening in in the late two thousand nine, two thousand ten, seeing seeing all these new startups pop up. And for many years I just kept watching that, you know, this was moving so quickly and then larger institutions were much slower.

Pauline Bertry 2:22

Hmm.

Ritam 2:23

So I really wanted to work in the startup community. and I thought I've come from a professional services background. So I understand that space. Why don't I set up professional services business for startups. and I I think, you know, obviously anyone who's got business experience will tell you don't build a services business for people who don't have money, which is basically already say startups. Anyhow we did. so we started pseudographine about twelve years ago and initially I contracted and and used those savings to essentially hire the first couple of hires in pseudographine and balance a book. So it was it's it was very It was bootstrapped, it was very sort of just practical in terms of how we got off the ground. and we I think the positive that I take away from trying to service startups is that they're, you know, in services we were told that when you're in agency land, you have to ask your client, do you what's your priority between speed, quality and cost? you know, and

Pauline Bertry 3:23

Mm.

Ritam 3:23

then we'll we'll we'll manage things. So you're always asked to kind of balance that triangle and and be upfront with the client that all three can't be achieved. So which one of the three is most important to you? And when I'd ask a startup founder which one of these three is most important to you, they just look at me blank. They're like, obviously all three. I want it good, I want it fast, and I want it cheap. Like and and

Pauline Bertry 3:43

And yesterday.

Ritam 3:44

yesterday, exactly. The fast bit is definitely important. But I think what was that my kind of glass half full view is that it taught us to be good and be fast. And be cheap at the same time. and it made us really optimize because we were forced into it. So, you know, when we started, we were all in London, and effectively a fully onshore capability, we realized that wasn't commercially viable. Again, because of the startup community. We learned that. we then tried to subcontract to other countries, but the quality really suffered. We built our own offshore capability in India. We realized that there were some things that were better in Europe. And we built a team in Portugal and and we just kept optimizing. And what we realized was different geographies, different people are good at different things. And you bring all of those people together, it's like magic, you know? And that's really what we what we what we kind of realized because we it wasn't that we could just go to a low cost location. We also needed the best people. and and we just optimized to be a high performance team. And fast forward a little bit, we are twelve years old, we have about a hundred people all in house, full time employees across those three geographies. We have one person Geneva, it's red herring because it's the first employee that joined the London office, but then he moved to Geneva. So so that's our Geneva office.

Pauline Bertry 5:07

Very cool.

Ritam 5:09

but we we we moved to work more towards the scale up end of the startup market and with enterprise clients. We have a lot of professional services partnerships. So we work with sort of top tier consultancies and partner with them to deliver projects. So a whole range of work. and we've also recently incubated our own products. So we are kind of blurring the lines of agency and product. and I think with AI it's it getting even more exciting because you can build. So it's a fun place to be right now. Lots of change happening.

Pauline Bertry 5:40

I mean I'm super curious on on like so ma so many fronts, but I I think what I I want to go into two topics as I told you. I want to talk a little bit about like the distributed teams and the people side of it and the like the incubation and the product and how you use AI, etc. But I'm I'm super curious to learn like how your perspective on Like what kind of people you are bringing in, how you are growing them, etc., changed over the last five years. Because what I'm observing is that there are companies who have been around for like very, very long time, and often they are not necessarily after like all of the AI changes. There are companies who appeared when EI was born. So obviously they are kind of AI native. And I think in in your case you you have a very unique profile of a company which was around before AI came into the picture. And you also seem to have gone through the transition of like adopting it internally from from what I understand. So how was this journey for you and for your team?

Ritam 6:58

Yeah. I think what's fascinating is that as part of this journey, it's also the f you know, I've been in this industry for just over twenty years. It's also the fastest pace of change that I've ever seen. So you can adopt a strashy and in three months that shrashi is no longer right.

Pauline Bertry 7:16

Viable.

Ritam 7:18

and I think what's worked for us well and we've learned is culture principles, values, those are kind of overarching facets, right? so for example, one of the things we've pushed quite hard is proactive failure. Like go and try things and fail and get them wrong and learn because otherwise we won't try. And I think the the question I ask ourselves, I ask our clients, I ask everyone it's my favorite question. Can you describe what your AI native competitor would look like? So for Cereographine, what would an AI native competitor look like for a client? What and everyone, I think we just have to go back to a blank canvas because right

Pauline Bertry 8:01

Mm-hmm.

Ritam 8:02

now everyone's trying to use AI to automate things. and in my opinion, that's not really the real use case. The real use case is to there's just new use cases that we aren't even aware of. So a lot of our thing is going blank

Pauline Bertry 8:16

Yeah.

Ritam 8:17

canvas. So as an example, you could use AI to write your code. So you're writing some code and you say AI autocomplete this for me.

Pauline Bertry 8:25

Mm-hmm.

Ritam 8:26

or you can say, I'm gonna follow AI driven development, which means I am going to structure my development process around AI writing code

Pauline Bertry 8:36

Ara.

Ritam 8:37

rather than AI helping me. So we I'm encouraging the team to build our own AI native competitor, whether it's designed

Pauline Bertry 8:42

I love it.

Ritam 8:43

product managers, whoever it is. And the other bit that we are really experimenting with and learning is understanding what different archetypes do well in what scenarios. So as an

Pauline Bertry 8:57

Hmm.

Ritam 8:57

example, one of the things we are learning is that new university graduates, which and we it generally our people strategy is we don't only listen to what everyone is saying in the public domain. Because every business is unique. So we are experimenting on our own and learning. So as an example, everyone's been talking about the fact that AI will wipe out the graduate hire market. But we

Pauline Bertry 9:22

Hmm.

Ritam 9:22

say instead of just believing that, let's try it out ourselves. so we have hired graduate engineers and they are absolutely smashing it because they are AI native coming out of university. They have like this is all they've known. So they're not having to unlearn the old way to do things. So we are also learning things that are different to what we read on the news. So we are seeing huge success with juniors and graduates. and and and so I think our approach generally is be curious, try and learn ourselves, make mistakes rather than just say, you know, I read this on a blog, so we're gonna follow this approach.

Pauline Bertry 9:58

I love it. I mean I've been really waiting to talk to someone who would tell me like we had a huge success on hiring juniors because I think like in majority of cases when I talk to people right now is like we cannot hire juniors because it's not economically viable. They it takes just too long to to upskill them and to get them to the point where they they can be productive. So how like I mean obviously when when you hire people out of your universities you say yes they are AI native, but how you are kind of helping them learn this judgment layer they would have built in two to three years generally if they were to join the same company five years before. So what what's your approach there?

Ritam 10:53

Yeah, look, there's no secret. Like I think they don't have the judgment layer, right? But what they do have is the ability to work in an AI native fashion. So

Pauline Bertry 11:04

Mm-hmm.

Ritam 11:05

what they do have is the ability to so so if you think about traditional I think what's happened over the last fifteen, twenty years is we've all established specialisms in the agency world, right? So

Pauline Bertry 11:15

Yeah.

Ritam 11:16

Everything has gone into if you're you know, I don't know marketing well, but let's say the marketing agency world there are SEO specialists, AdWord specialists, et cetera, et cetera. in our in a design world there's user experience, user research, UI, et cetera. In product

Pauline Bertry 11:28

UI.

Ritam 11:30

management there's like all sorts of things, project management, facets, product owner, scrum master, etcetera. In engineering there's front end, back end, DevOps, cloud, there is all the different languages and frameworks and the syntax.

Pauline Bertry 11:41

Hm

Ritam 11:42

I think what AI has changed. Is you don't need the specialist skill per se. So if you're an engineer and you're a front-end engineer and you know JavaScript, AI does quite quickly allow you to conquer Python and be a back-end Python

Pauline Bertry 11:57

Mm-hmm.

Ritam 11:57

engineer. and it's blurred the boundaries of specialisms because those skills, the AI is your skill deployed. so I think what's happened is let's say an engineer has been worked in streographing for seven, eight years, they would have been a specialist mobile developer in a specific

Pauline Bertry 12:16

Yeah.

Ritam 12:16

language and framework. Today they can be across any language and framework. They're having to adopt that mindset and that mentality, which a graduate already comes with because they have not known anything different. And it reminds

Pauline Bertry 12:27

Mm.

Ritam 12:28

me a lot, I always say what I'm watching right now is taking me back twenty years when I started my career. When I started my career, I was just asked to do everything. You know, I was an analyst. Like I the the joke was you're the P B A, D B A, QA,

Pauline Bertry 12:42

Yeah.

Ritam 12:43

all the and I think we can go back to that. So the positive is that they're not having to unlearn this era of specialisms. They are

Pauline Bertry 12:51

Mm-hmm.

Ritam 12:52

able to be generalists. So there they have an advantage over experienced individual contributors. What yes, they don't have is experience and judgment, and that you can't substitute. So what we're finding is almost like a barbell effect. Where people are willing to rise up and use their judgment as their main superpower and they're senior, they're doing really well. The the the juniors who are willing to learn fast and and be flexible in a more fluid role are doing really well. I think those that are clinging on to specialism individual contribution are not doing well.

Pauline Bertry 13:26

Yeah. I have a very similar observation. I do think like we for for a very very long time we were talking about kind of T-shaped individuals in terms of profess like professional skill set and I think what AI is bringing to the industry is that it is making the expectation out of the Both bars higher, which means that yes, you need to have like an extremely wide horizontal bar to be able to kind of cover end-to-end spectrum. But at the same time, if you want to specialize, that's also fine, but then you need to be so deep into like a technology or expertise or knowledge that you are. Like you need to be better than all the human kind of averaged knowledge combined, which which is basically the the model that that you can access. And and I think it's almost like what I'm observing is almost this kind of polarization of of expertise. It's either you are going even broader or you're going like extremely

Ritam 14:43

Very deep.

Pauline Bertry 14:44

deep, extremely deep.

Ritam 14:45

And the question is what you go extremely deep in, because there are many things that now AI does so well. So, you know, you could be a researcher, use sorry, AI researcher, yes. So also what you need to go deep in is changing, right?

Pauline Bertry 14:57

I agree.

Ritam 14:59

that's the big evolution. So we are finding it's very exciting because everyone's having to relearn, retrain how they do things. And I think everyone's that talking, you know, the vast majority of people are saying hiring juniors is bad.

Pauline Bertry 15:12

Hmm.

Ritam 15:13

I completely disagree because of my own experience, you know, and and I think it's also made me learn you don't every company is different. You have to see and look, I think for us what one thing I have to say is, I'm grateful for the wax fact that spirphine generally has very low attrition. So, you know, just naturally when we were a small company, we hired a lot of juniors then because that it was a affordability factor.

Pauline Bertry 15:35

Yeah.

Ritam 15:36

But luckily most of those people have stayed so actually, you know, over 12 years now we have a very senior team. so

Pauline Bertry 15:41

Yeah.

Ritam 15:42

there is also a factor that when we bring in junior, you know, to give you an idea, out of a hundred people, we'll maybe bring in five junior hires in a year. So it's it's only it's a very small proportion. And so we are actually already quite top heavy and quite senior as a business. So again, there are different nuances. Every business is different. Yes, if you're a business which already has so we have more of a I don't know, diamond shaped pyramid or more, upward shaped pyramid. Yes, it's true. If a business is like this and you already have more juniors, you're going on adding more juniors, maybe they will have another challenge. but

Pauline Bertry 16:10

Yeah.

Ritam 16:11

for us, you know, five juniors are coming in and they have fifty seniors to mentor them. So that's also

Pauline Bertry 16:16

Yeah.

Ritam 16:16

the difference.

Pauline Bertry 16:19

But what's what's the secret there? Like you said that most of the people stayed over time. Generally, like professional services is the industry where the turnover is one of the highest in comparing to comparing to other industries. if I'm not mistaken, I think for a person between twenty four and thirty five years, the average tenure in professional services is like two and a half years or or something like this. in your case like looks like the tenure is much higher. So what's what's the secret, what's the story there?

Ritam 16:55

Yeah, I think the average tenure of I look I don't know the exact style, but I think the average tenure of the current serving pseudography employees, you know, about four and a half years. and considering, you know, most of our growth came

Pauline Bertry 17:05

Almost double.

Ritam 17:06

recently, and what most of our growth came recently, so you have to factor that in. We also, you know, you're factoring in new joiners. I don't think there's a secret. I think our approach has been so we have we have certain values which represent who we are and one of them is to be compassionate and I think one of our things is that th I worked in a job for ten years before starting studiography and there are just a whole list of things that I remember then were not done for me, that I did not think cost more money. It cost more effort, yes. But it didn't cost more money and it would have made my experience of being in that job far greater. So, you know,

Pauline Bertry 17:46

Mm.

Ritam 17:47

a simple example, we are going through our annual review cycle in the UK at the moment, and we you know when when I was at Accenture, my memory is, you know, I just get a letter once a year, you know, like this is your change and pay, etc. But it was very impersonal. At least Five Road reports to me and someone who has a reportee that they don't want to speak to. nearly everyone the in the in the UK office, I will sit down for an hour or hour and a half. I will prepare for that meeting for an hour or two and have a very honest conversation about their performance, how they're doing, how the business is doing, and and and speak openly about numbers. And I think the main thing that we do try and embody is that we want to treat everyone the way we want to be treated. And you know

Pauline Bertry 18:30

Hmm.

Ritam 18:31

we so there is that. I think we We have certain principles, you know, we are a B Corp. We are we we also want to do work that's interesting and good. and

Pauline Bertry 18:44

I love it.

Ritam 18:46

yeah, so I think I think it's been less of a prescribed set of policies that have led to this. It's more a way of being and we just wanna be a nice place to work and we want people to come in and do their best work, enjoy it. we wanna support them. and look, we get lots of things wrong as well. So then we acknowledge it. we do an annual sort of internal survey and we say the one thing we wanna change is that every year when we ask you have we acted on the feedback from the previous year, that score should be the highest score. That's our North Star. We wanna see that

Pauline Bertry 19:19

Hmm.

Ritam 19:20

we're constantly acting on feedback. we I think if you give people autonomy alongside of it creates a lot of i i ownership and excitement in the work they do and that also helps to retain but look it's yeah it's I feel blessed. I always ask. I I'm not I'm not sure what the secret is. I don't know that I can't I can't always pinpoint it. And but yeah, i you know you know there's professional services there's just one one asset that matters it's your people and we're very lucky on that front. I think one of the challenges that I do think about is, you know, AI is making it harder to create headcount growth and headcount growth just naturally creates a bit of a pyramid and personal growth. So I think that's one thing we want to watch carefully.

Pauline Bertry 20:11

Hmm.

Ritam 20:11

you know, if we are static on headcount growth but still growing in other ways and revenue, et cetera, how does that affect people? because traditionally it's been like, you know, you'll get promoted, et cetera. So

Pauline Bertry 20:21

Yeah.

Ritam 20:22

that's something we have to think about.

Pauline Bertry 20:23

Interesting. I think that that's a question that has come in a lot recently on the kind of how do you enable growth in the teams where the head count growth is not is not planned or is not is not expected. I and I think one of the best Angles that I've heard and that also resonates with what you are sharing is like it's called lateral growth or like horizontal growth, basically enabling people to take ownership and kind of build almost their own pieces of the business that would basically put them into the position where they would be building a new business. service or a new service line, I mean we can call it however we want, from scratch, because then they would need to recreate the same structure with the all of the like best practices, et cetera, that that you had in parallel to to the core business that you are building. And I think there is a lot of conversations in in this area right now, especially with like AI, et cetera, as we said.

Ritam 21:35

A really good point. so we it's it's a really good point. So, you know, one of the challenges in professional services is that you go from having a bench to being fully utilized in in in quick succession and often so in twenty twenty three we had a big bench. everyone's payroll, everyone's employed and and yeah, one of the things I you know, one of my worst fears is to ask someone to go for business performance reasons. And It's something that I, you know, I wanna always try and make the last resort in whatever way possible. And, you know, we're there are different strategies, we try and make sure we have s sufficient cash reserves, et cetera, to facilitate it. So in twenty twenty three we had a big bench and one of the things I do within Cedro Graphene is just naturally, you know, we have engineers, designers, product managers, you know, finance ops, all the ancillary functions, HR. But I personally manage the product managers and The product management team had some capacity. And I said, you know, what is the next step if you're not becoming a product director and you know, you're not growing in in in sort of CPO, you're not growing that track.

Pauline Bertry 22:36

Yeah.

Ritam 22:37

What is the lateral growth track? And I thought, well, actually a product manager is quite close to being an MD of a business. You know, that's the you know, it's a generous. And I said, Why don't you start a business from within Cedar Graphene? And we did that and we incubated a business. And now it's doing a s a significant ERR. But then we also said, look, you can incubate a SaaS business or you can incubate a business within our business that serves our business. So as an example, we've created a product called Pulse, which basically measures productivity and quality. It's basically agency tool to measure productivity and quality of your

Pauline Bertry 23:09

Yeah.

Ritam 23:09

engineering function. so it's almost a product that's serving ourselves as a client, you know. But this it's being run in a way as if it's an independent business that is accountable for certain metrics. You know, so it's you know, odd do you have enough daily active users, et cetera? but yeah, that has been a really and I think with AI it's gonna get even more interesting. You know, you can build stuff faster, cheaper. The zero to one market has become really short.

Pauline Bertry 23:35

Yeah. I have a question on on that. because I mean I think that's like the element that that you're talking about, which is like the incubation of your own product, etc. It's something I hear a lot when I talk to people in professional services at all sizes, like from two to three people to like McKinsey, BCG, Accenture and and bigger guys. And I think what is always kind of the question is how do you balance the like need of the service business and the need of the product business? So let's say like if tomorrow your service business needs an additional person for a client project or like

Ritam 24:20

Do we do we take them back?

Pauline Bertry 24:21

And an yes, do they do you take them I mean, that's not a direct question, but how do you like

Ritam 24:26

No, it's a great question.

Pauline Bertry 24:28

how do you how do you balance that?

Ritam 24:30

So the it's a great question. So we didn't initially. So before twenty twenty three, we used to constantly, you know, we'd have a bit of capacity, we'd build something and we'd throw it away because we didn't focus on it. I think what we've changed now is if we've tested something and we believe we want to back it, we treat that product as applient. So even though, you know the the the the billiability in intercompany billing is at cost, we treat it as a highly paying c client in terms of prioritization of resource. and it's a really hard discipline because you know if a big client comes in as a big opportunity, there's a bit of tussling. But we've tried to abide by that principle to whatever extent possible. We say we treat it like a client. So as an example, if a massive opportunity comes in and we have to take a resource off, but that could happen with the real client and we'd go to them and we'd say, look, We've worked with you for ages. We've had this massive opportunity that's 10x the size of revenue and we need this particular resource. We will make it up to you. We will give you two other resources instead of something. However, you would treat a normal client. we do internally. so we we we've picked that principle. We're like, otherwise we'd rather not do it. We'd rather just have a bench. and the other bit is mindset, right? So a services business, and I think this is where we're seeing everything conflate across product and services. But the mindset piece in my mind is that what's happening is people are questioning the time and materials model. And and you know the the the different mindset between product and services is services your vested interest is to inflate the amount of work involved because it inflates the time and you get paid by time. In products it's got nothing to do with the time you spend. And

Pauline Bertry 26:23

Mm-hmm.

Ritam 26:24

I think building our own product and incubating our own products has made us a better services business because we understand what it takes to bring it to market, et cetera. because our internal ethos, I always say is look, ooh, the best you can for the client. Fastest, cheapest, best. Because even if you build less because you spend less time, that client will come back because they'll see value. But I think building our own product business, we've started to see where there is value.

Pauline Bertry 26:49

Yeah, it's also fair because you are working for the type of businesses you're also building internally, right? I think you said you are you

Ritam 26:55

Correct. Right.

Pauline Bertry 26:56

moved to a to to a different part of the spectrum, kind of closer to scale ups versus

Ritam 27:04

Yes.

Pauline Bertry 27:05

like zero to one startups, but I think it's it's the same universe. we met, you shared that I mean I know now as well that you are extremely excited by Like what AI is bringing to the workforce? You've played with a lot of internal use cases. So I mean I would love to maybe open this topic with a bit of broad question, like how do you see the value of AI in in professional services, and like specifically agency world, digital boutiques world, etc. So how how is it coming to into

Ritam 27:42

Yeah.

Pauline Bertry 27:43

the picture?

Ritam 27:43

So again, like with most things I feel like I end up having a bit of a contrary view. But I you know, everyone says AI is gonna kill the agency business, it's gonna kill services, etc. It's gonna kill SAS, it's gonna kill this, that, the other, you know. AI's gonna kill everything. Yeah, we're all dead. Everything's

Pauline Bertry 27:55

It's gonna kill yes, all of the big consulting as well.

Ritam 28:00

over. Game over. so my my side so I went for this talk by this chap called Greg Crabtree, who's written a book called Simple Numbers, and he said, you know Right now you go for any talk, you go for anything people talk AI, right? So everyone has a view on AI. And he said something really interesting. And he said, In the end, people pay in the end, the markets evolve to pay for labor. If there is no labor, there is no arbitrage. Because you just and and so the concept is he's like, Well, if AI does everything, we would just won't make any money. People pay to get other people to do something. That's how the economy works. And it's always worked that

Pauline Bertry 28:32

Mm.

Ritam 28:32

way. So, you know. When the tractor when the industrial revolution happened and tractors came, it's not like people started growing their own vegetables and not paying people to make vegetables for them, right?

Pauline Bertry 28:41

Yeah.

Ritam 28:42

Because in in effect, you could argue growing vegetables is free because it requires air, soil, you know and and and and water.

Pauline Bertry 28:50

Time.

Ritam 28:51

Yeah. And and but the point is people pay for it because there's transport, etc. But all those elements involve labor. And I think I came away from that talk and I thought about the fact that All we have to do as consultancies, agencies, services, businesses is figure out where the next need for labor is, and

Pauline Bertry 29:13

Hmm.

Ritam 29:14

where the next need for human value addition is. And of course, if we're going to be inefficient and not leverage AI, someone will outcompete us. So the farmer who uses a tractor will outcompete the farmer who physically goes and, you know, digs up. the mud and puts the seeds in and physically takes a water can and waters it.

Pauline Bertry 29:37

Yeah.

Ritam 29:37

And I think that is what that is why I always ask people, who is your AI native competitor? You know? You are building an AI native competitor for the HR world, right?

Pauline Bertry 29:50

Yeah. I need to talk to you about the evolution we went through because I think I didn't tell you yet, but that's outside of the podcast.

Ritam 29:58

Fine, fine. but yeah, I think everyone should be thinking about what is their AI native competitor in building.

Pauline Bertry 30:08

I love it. And from what you like from all of the use cases that you've played so far, what is and I mean of course what you feel comfortable sharing. what is like an AI native competitor of product studio business? like how does it look like for you and what are the elements of it you've already played with?

Ritam 30:36

Yeah. I think we are like the like we're trying to distinguish between AI enabled, AI powered, and AI native, right? AI enabled, AI

Pauline Bertry 30:44

Uh-huh.

Ritam 30:45

you know, AI enabled is like, I'm writing some requirements, let me ask Claude to check them and write

Pauline Bertry 30:50

Mm-hmm.

Ritam 30:51

them up and rewrite them. AI native is let's reimagine our entire workflow. Right?

Pauline Bertry 30:56

Uh-huh.

Ritam 30:57

So as an example, we have processes amongst the product management team. Those pro processes are we maintain a risk register on a project to make sure we identify potential risks. We do a v variety

Pauline Bertry 31:10

Mm.

Ritam 31:10

of different things. What we're doing now is we're saying, well, in the world of AI, how should we do it? You know, does it does it mean that now AI is just constantly scanning for risks across the project plan and code base and Atlassian Jira and Confluence have we use and and so on and i like a constantly dynamic risk register rather than once a week? Look at it. you know, so we're questioning everything given that we have this new capability, and we

Pauline Bertry 31:38

Mm-hmm.

Ritam 31:38

are questioning how we should do it. So we're not saying, we maintain a risk register which we update once a month, so we will now get AI to write it, or we'll use voice to dictate what the risks are to AI and we'll transcribe and write it. We're saying, Well, what should we be doing in the world of AI? Should we just be running a dynamic risk register, just constantly flagging risks? same thing with the way we create presentations. We're like, well our head of design created a tool that allows us to pump out slightly tweaked versions of presentations for clients where we're pitching to them based on it's still using the pseudographene brand, design guidelines, etcetera, but tweaking all the templates for that client. So using slightly tweaking the colors so it matches their tone. putting the footer and header so that it matches their company name, et cetera. But it's just it's constantly thinking about, well, in the world of AI, how do we do things differently? So now for example, when I send a generic credentials presentation to a client, I don't send the generic one. Because in the world of AI, it takes me a second to make a slightly custom version with their name and logo, so it's a bit personalized. So now the way I even represent ourselves has changed. And I think we're just constantly asking ourselves, Let's go back to a blank kind of say we're building a product management team from today, right? what is what is the future? And I think it's really hard to unlearn the way you do things. So in the beginning we're like, why

Pauline Bertry 33:08

Hmm.

Ritam 33:08

do we have to do this? But we are yeah, we're just going back to asking ourselves. So the other example from a product management perspective, we always write requirements for a zero to one. So you come up with an app idea, we write the requirements, we do

Pauline Bertry 33:20

Yeah.

Ritam 33:21

detail discovery. We do detailed design, we create wireframes, we do all sorts of things. Now we are experimenting a path where we don't do any of that. We just build it using lovable or whatever you want to do, you know, which is

Pauline Bertry 33:35

Mm-hmm.

Ritam 33:36

that's our that's our that's our discovery. Like it's just version one. because it's very low risk, there are no real users, etc. And then get feedback. You know, so we've we've gone rather than taking three months to build an MVP, but why we there is a five minute version. Because I think what people misunderstand about what AI is changing is everyone's like, is it adding efficiency of 20% or 30%? I'm like, no, that's not what's happening. AI either 100x is your task or has no impact. You and I are having this

Pauline Bertry 34:04

Yeah.

Ritam 34:05

conversation, it has AI has no impact on, you know, if if we record this podcast over an hour, AI is not going to make us record this podcast in 10 minutes. It's not going to make us record this podcast. It has no impact on this podcast and the I guess efficiency with which we record it. But then if I'm writing a unit test on some code, it might ten exit or a hundred exit. So I think

Pauline Bertry 34:26

Yeah.

Ritam 34:27

people are looking for like, well, is it twenty percent? I'm like, no, it's it it varies depending on the task you're doing. So I think we're trying to rethink what that allows us to do and how it allows us to do those things.

Pauline Bertry 34:37

That that's that's so true. And I like like specifically I I would love to kind of jump jump in here because I I have a few thoughts to share. What I like the best AI tools I play with are the ones that let people be people and take away the part that is boring. Like take Riverside, the this is the tool that we are using

Ritam 35:04

Yeah.

Pauline Bertry 35:04

to record. I mean, of course, they are in the hyper-growth mode, they still have some challenges they are overcoming, like the bugs, etc. But what I love about this tool is that it gives me a great experience of talking to you. at the same time, it makes my experience of editing 300 times faster. Because, like when I need to cut something that I said wrong. I just get the transcript and I can very easily cut it instead of, you know, like in five,

Ritam 35:37

Yes.

Pauline Bertry 35:38

seven years ago going and trying to find like the perfect

Ritam 35:42

Yes.

Pauline Bertry 35:42

second. If I need to make a thumbnail for the for the video, I just give it input and it creates one. So it like it lets me enjoy the experience of being the podcast host, talk to people, and then of course like there is a bit of work as well to do to To make it like to make the final product, but I almost don't spend that much time on it because the tool enables me to do so.

Ritam 36:14

Yes.

Pauline Bertry 36:16

And and I think on the other side, also an interesting observation I have with the five-minute version. And I think that's also something we will eventually need to unlearn with AI. My frustration, like my acceptance for the level of frustration versus certain tools, is becoming extremely low because of AI. Because like before, what would happen is that okay, you have a tool, it has certain constraints, you will just live with them because that's the only option you have. Now,

Ritam 36:50

Yes.

Pauline Bertry 36:50

like the minor frustration I get, I'm like, I'll just go build it. So then I go build it. I spent a lot of tokens and in like in a lot of cases, the five minute version of it is not necessarily the best alternative. And I think that's a a reflex as well to kind of learn now that not everything needs to be built.

Ritam 37:15

So I couldn't agree more. I feel like people like to swing from pendulums, right? Like they're like I bet in six months there'll be, you know, an AI bubble crash or something and people like, AI was just a fad. I'm like, no, like two things can be true at the same time. You it is transformational for some things and it's not transformational for some things. It's still in its infancy. It's like the dial up era. We don't know how much is gonna change. But you're absolutely right. This the this thing is that like, someone's paying, you a hundred pounds a month license for something, they're like, I don't want to pay it, I'll just build it myself. I'm like, no, that's not how it works. Because then you have to test it, you have to check it works, you have to constantly update it. I just think, like with everything, things will evolve. We will evolve as

Pauline Bertry 37:53

Mm-hmm.

Ritam 37:53

services businesses, SaaS will evolve. I think the one bit that I am finding really interesting is there's evolution and there's conflation. And I think what's up

Pauline Bertry 38:05

Hmm.

Ritam 38:06

A lot is getting conflated. I'm seeing product business provide services. I'm seeing services businesses, so agencies build products, products building the agency arm. So, you you're you're seeing this with OpenAI and Anthropic, etc., they're building their own consultancy

Pauline Bertry 38:20

Yeah.

Ritam 38:21

networks and partnerships. So and I think it's going back to the fact that it's proving that the human element is where you create value in the end. You know, if everything is fully automated, there's zero value because it's just it's a race to the bottom.

Pauline Bertry 38:35

Yeah. I I agree. And then I think also there is another angle to it, is like the the vertical element of the expertise. And I don't remember exactly where I found this research, but what they were talking about is that if let's say my expertise is in accounting and your expertise is in product design or product management. If you use AI to ask it accounting questions, you will be overwhelmed by the quality of the answer, you will be like, Wow, that's amazing. But me as an accountant, I would be like, not really. And vice versa. Like if you like if I as an accountant use AI to create a website design or like a product service, I will be like, my god, that's amazing. But

Ritam 39:30

Yeah.

Pauline Bertry 39:30

at the same time, you when you will play with it, you will be like, yeah, okay, that that's average, but maybe not that distinctive. And I think for me, what what this means is that basically the bar is getting higher for like what is expected when when we are talking about expertise.

Ritam 39:49

Yes.

Pauline Bertry 39:51

and then you still need like this one to ten. knowledge peace, which doesn't come from knowledge, it comes from experience, from like having done this 300 times. And I think

Ritam 40:06

Yes. So

Pauline Bertry 40:07

that's what what the human angle will be about.

Ritam 40:10

Hundred percent. So look, people still say that AI's done something and it's corrected it or fixed it. So I always say like, you know, the human expertise in a field, like it's not like, you know, I'm gonna ask Chat GPT for a medical prognosis, I'm gonna go to a doctor. and inevitably and you know, I speak I spoke to a lawyer the other day and he was telling me the People are just using AI now to like sue people and send them legal letters. It's really frustrating because all the all of them have like some flaw or some mistake that they haven't thought through. So

Pauline Bertry 40:46

Curse.

Ritam 40:47

the specialism, I don't and and and so we we did this sort of study internally where we talk to a lot of people about how they feel about the extent to which they use AI in their work from an agency perspective. And and how they determine to what extent they'll use AI in their work. And they said the way they determine it is by looking at the risk profile of the work they're doing. Because the higher the risk, the more they need to check it from a human perspective

Pauline Bertry 41:14

Yeah.

Ritam 41:14

and contrast AI. So if you're building on a banking application that has millions of users, you will move slowly. You will check everything as a human. You'll have loads of governance.

Pauline Bertry 41:23

Yeah.

Ritam 41:25

But if you're building something like you come up to me and you're like, hey, I have an idea. I want to find a way to build a tool where someone can collate feedback on an individual based on their Slack exchanges over the last year. and I'm like, okay, let's see if it does something cool. I mean we can do that really quickly because we are not risking that someone's gonna use it and it's going to you know, transfer money and and and have any kind of financial commercial impact and there aren't gonna be millions of users. So I think what we're realizing is where the risk is low on a task, AI is the perfect solution because you can take the

Pauline Bertry 42:03

Yeah.

Ritam 42:03

risk. You have a probabilistic view that ninety five percent of the time it'll be good enough. and that's how we are working on that that's the approach we're taking.

Pauline Bertry 42:12

Yeah. And I think on on this one there is another element which is a bit not not exactly the same what what you are mentioning, but it's very close. It's called generative UI. So it's bas I mean most likely you've heard about it, but it's the trend where like the web interface is becoming the product of a one-time conversation where you don't

Ritam 42:36

Right.

Pauline Bertry 42:37

Like like same same way you do it in in the chorrect

Ritam 42:39

Like a dynamic interface. Like

Pauline Bertry 42:41

correct and I think what what you are saying is like exactly going into this direction because now you can almost build certain tools VDI because it's so fast that will help you achieve things that you only need to do once that you wouldn't be able To do five years ago because I wasn't there. Like an example, I built a super simple tool that is helping me to process my LinkedIn connection data into different buckets. Like before that, I would need to do it in Excel. It would be an extremely challenging exercise. Now I built like my internal own, like Tinder-like interface where I can look at the profile and decide in which bucket They go, and I see people doing this more and more.

Ritam 43:33

Yeah, it's fascinating. I think and this is why I think the real I don't know, the real discovery still is all these new ways of doing things that we didn't think of. That's AI native in my mind. So the generative UI is AI native. Like we didn't think of

Pauline Bertry 43:50

Yeah.

Ritam 43:50

it and we couldn't do it pre this world. So and I and I think that path to discovery is still very fresh, you know. Before the App Store existed, no one would have thought of an idea. would have never never come up with the idea of Uber. You know, they would have so there are all these things that are happening as a result of the capability. And

Pauline Bertry 44:09

Hmm.

Ritam 44:09

I think we're still very early in that curve. the reason I find it very exciting is I think disruption creates opportunity. and I started the business in twenty fourteen and you know the first year we were figuring things out. I was still sort of contracting part time. So really twenty fifteen. We haven't really seen any disruption. You know, like two thousand I think seven or so, I can't remember, but that is when the iPhone came out, you know, that was the real, you know, last disruption was the mobile wave.

Pauline Bertry 44:37

Big thing. Hmm.

Ritam 44:39

And prior to that it was the internet, the late nineties. So, you know, most of my career, especially most of my well, all my time running the business, we haven't really seen a disruptive technological change. Yeah, there was a sm short wave around how blockchain impacted things, etcetera, but this is very different. This touches on everything. So I do see this as the big disruptive wave.

Pauline Bertry 45:04

No, th this is a this is interesting what what you are saying because my career started in two thousand sixteen, like the the proper part of it. And I mean I think a lot of people in like my generation are now kind of anxious and worried and like, my god, what's going on? Like will AI kill all of the like digital jobs, etc. And I think you are you are right, it's mostly because Like when we haven't seen the laptops arrive in the workforce. We were kind of born ish with them. same we kind of in the university we had the phones coming in. We didn't feel the impact of the two thousand seven, two thousand eight crisis because we were way too young. And that's the first like real disruption we are living through. But I think it's

Ritam 45:57

Yeah.

Pauline Bertry 45:57

happening every like ten to twenty years. So maybe that's just something to, you know, accept and live with.

Ritam 46:05

Yes, yes. That's why we're seeing like the the people who are finding this most comfortable are the very junior recent graduates and the very experienced people. So people with twenty, thirty years experience, because they've seen other disruptions previously. They've seen change. So they're not scared of change.

Pauline Bertry 46:21

I love it. That's that's a very good that's a very good directional and aspiration inspirational advice. But I have a last question, just to be kind of being mindful of time. you already mentioned a few things that that you are playing with. What are the like several use cases you hope to experiment in the future? for for you and for graphene or maybe for the product that you are building, etc.

Ritam 46:52

Yeah, so the use cases we are trying to experiment with is I so I'll I'll tell you kind of my focus. I'm loving the the simplicity of the MCP capability.

Pauline Bertry 47:05

Mm. Yeah.

Ritam 47:07

And so I you know, we're building several MCPs. They're quick to build, they're cheap to build, they're fast. and it is It i I can't understand why you'd, you know, now build a visualization tool, you know, it was it just immediately allows you to query your data so well. So that is the one area we're experimenting. We built three hundred products for clients over the last twelve years. That is a pool of data. How can we learn from it and leverage it to serve clients better? So MCP just is become this kind of experimentation route for us. we are really looking to s think about what is the future of design? Like you talked about the generative UI, like We feel that has not been disrupted. So you have the iPhone interface, which has been the same for now nearly 20 years. you have the landscape view of a computer or laptop, which has been nearly the same. and so we are thinking of the future of design and we're doing r loads of experiments, predominantly

Pauline Bertry 48:02

Mm-hmm.

Ritam 48:03

around how can you create a more dynamic interface that changes as you talk through.

Pauline Bertry 48:08

Mm-hmm. Very good.

Ritam 48:11

we are rethinking the so you know we come from this tradition of zero to one, like getting things off the ground. We're rethinking actually how you condense and compress zero to one so much that it's nearly free to understand if an idea is workable and get feedback. And to focus really on how do you use AI to get traction. Because we believe people will build things and experiment but how do you use it to gain traction? And how do you create that product feedback loop to grow that traction? and and I think the last thing I'll say is what we see as a future and we're experimenting on is AI is probabilistic, traditional software is deterministic. Almost everything we're working on experimenting with is trying to figure out how those two work together. Because we

Pauline Bertry 48:58

Hmm.

Ritam 48:59

believe all software will have some probabilistic AI element within it and it will have traditional determin, you know, your registration or login is still gonna be your deterministic piece. And

Pauline Bertry 49:07

Yeah, yeah.

Ritam 49:09

so we're trying to experiment and play with How how those two work together.

Pauline Bertry 49:14

Yeah. I love the I love the angle on zero to one and attraction testing because we kind of ended up playing with with that as well. So now we have basically everything that we are launching we are basing it on the hypothesis based on the research that is informed by like my conversations with people and also some like external sources. It is building the hypothesis, we kind of input feedback and then all of the kind of elements of the product go to market strategy, if if we can call it this way, it is built b based on that. We are about to launch the two first campaigns with with this approach. So let's see if it's working. I don't know.

Ritam 50:07

Yeah. No, and I think this is why we have to think differently.

Pauline Bertry 50:11

Yeah. I agree because I think it is giving us access to not only speed, and I think that that's what I'm observing mostly is that AI is giving access to speed, but I think it also giving access to kind of a lot of analysis capa like analytical capabilities that we are often not not leveraging enough. So I was testing it. a lot on like analyzing my own pitches and how I talk about what we are building. And then at some point, like if you instruct it well to give you feedback in the right way, you're like, I maybe need to work a little bit on

Ritam 50:47

Yes.

Pauline Bertry 50:47

my presentation skills. So yeah, I I agree with you. Any kind of closing guiding thoughts on AI and future of design for everyone who is listening?

Ritam 51:02

I think there's a there's been a lot of conversation around people being trained and you know, being coached and and and so on on this topic. I think my guiding thoughts are experiment, play. and I say this to everyone, that actually the best value you'll get is just from playing, trying different things. and and yeah, so my my sort of my my closing thoughts are Experiment, you'll be surprised. Just try things, whatever tool you use, George. you know, I've seen amazing things be created just on chat, not even cowork, not even code, clot code, just just using chat. and and so yeah, just play around for for the use case based on what will make the most amount of impact for you personally in a custom way.

Pauline Bertry 51:51

I love it. Thank you so much. Thank you for taking time to share your thoughts. I

Ritam 51:56

Tá junto.

Pauline Bertry 51:57

love the exchange. Bye bye. Have a great day.

Ritam 51:59

Guys, the right money. Thank you.