How to Scale a Business Without Wasting Time, Money, or Talent | Mooly Beeri

Mooly Beeri: When we are able to identify the critical path, which are holding back the entire organization, we focus all of the improvement energy into those, and that's how we get very quick results for the customer,

Carmen: Hey everyone, this is Carmen Durvat and I will be hosting today's episode of Behind the Founders. Joining me here today is Mooley Berry. Mooley, welcome. Can you introduce yourself?

Mooly Beeri: Yes, hi Carmen. Thanks for hosting me on the show. Yeah, so I'm Muri Berry. I'm a longtime software guy, probably close to 30 years. Actually, if you count from the moment I started programming when I was 11, then probably more than 40 years of experience. I've been in many startup companies, in many corporates, in Microsoft for many years, and I've slowly drifted towards software excellence and I've done a couple of roles on that front. I've built the first software center of excellence for Philips. I've done a head of software transformation for large automotive companies, and all of that led to me starting a business called Betasoftware where we drive large scale software transformation in in in companies. And

Carmen: Okay.

Mooly Beeri: I'm originally from Israel, I live in Italy for the last few years and yeah I think that's more or less it for an introduction

Carmen: Great. Thank you so much for the lovely introduction. Definitely quite an international experience as well. I mean, I mention you mentioned that you've also had some connections with Brazil. you now live in Italy. what is it like to have all of these diverse cultural experiences?

Mooly Beeri: Yeah, I was I was fortunate to work in some large corporates where and this is this is pre-COVID, so when you had to work with some team, you had to travel to that team and actually

Carmen: Yeah.

Mooly Beeri: meet them in person. So I I've had teams from China, through India, Europe, North America, South America, and and I've had nice experiences working with with with most of them. And

Carmen: Mm-hmm.

Mooly Beeri: It's it's it's it's it's really an interesting experiences. I mean you you learn how to how to operate in different cultures, you learn how what matters to different cultures and what values drive their their day-to-day behavior. And so today, you know, I start my morning with meetings with China and then I

Carmen: Okay.

Mooly Beeri: travel with the sun, go to meetings with India and then Europe

Carmen: Yeah.

Mooly Beeri: and North America without moving from my chair. It's much more efficient, but maybe a little bit less fun. Yeah.

Carmen: Yeah, yeah. I can imagine it's kind of like moving throughout and traveling through through the globe while also being physically in the same space as well. That must be very interesting. yeah, I can imagine it's also really quite different as well, working with the different cultures. I can really relate because I I grew up in China. I'm half Chinese, so I know about the different communication styles and experiences as well.

Mooly Beeri: Yeah, I I had my my When when I when I built the software center of excellence for Philips, we were tasked by the CTO at the time to do an initial scan of the technical landscape across Philips. And at that

Carmen: Mm-hmm.

Mooly Beeri: time Philips was composed out of many many acquisitions across the globe, maybe close to a hundred companies. And each each of these companies thought Hey Philips acquired us, so we must be the best. We must be the best with our processes, we must be the best with our technical capabilities and What we did was to travel across all different companies through the span of about six months, visit each one of them physically and do an assessment of technical capabilities because the goal was to get at the company level to define the best of all practices. And so we had

Carmen: Mm-hmm.

Mooly Beeri: quite an interesting six months visiting probably almost every possible spot on I mean I can I can probably pinpoint some areas where Phillips didn't have a site,

Carmen: Yeah.

Mooly Beeri: but maybe apart from New Zealand and Antarctica, I think we visited almost everywhere.

Carmen: Everywhere as well. Yeah, that's definitely quite exciting as well to be able to soak up the different cultures. Well, what I love to learn a little bit more, Mooli, is you mentioned that pretty much at since the early age of 11, you have really been always the world of software has always followed you around. What I'd love to know is what inspired you to pursue the path of software development, software engineering.

Mooly Beeri: Well I was probably lucky to grow in the age where computers became very prevalent. So I went through the entire process of starting with my Texas instrument, going through my single spectrum Commodore sixty-four, the early XT eighty Pentium and and so on. And I started programming very, very early just for fun. And Yeah, it just it just naturally led to studying electronics and computers and doing my my degree in computer science, which led to position in Microsoft and yeah, and so it just it just evolved quite quite naturally for me. it was always very interesting to me to to see how technology evolves. We'll probably talk a little bit about AI later, but I've seen quite a few revolutions through the years and there's a new and exciting one now but I I've been through a couple of those and yeah it's it's just I'm one of these guys that I'm I'm lucky to work in something that I really like. And so

Carmen: Mm-hmm.

Mooly Beeri: it's relatively easy for me to get up every morning and and get to work because it doesn't feel like a burden.

Carmen: Yeah. Doesn't feel like a job, it's something that's really aligned to your passion as well.

Mooly Beeri: Exactly, yes.

Carmen: Yeah. Well, Mooli, I'm just taking a look at your LinkedIn profile and I see that before you decided to find better software dev, you were at aptive before. And I have heard a lot about Active. I believe they were a subsidiary of BMW building software for I believe it was autonomous driving or something like that. What exactly were you what was your main role back then when you were at aptive?

Mooly Beeri: Yeah, so I I I joined Aptive as the head of software transformation. And

Carmen: Mm-hmm.

Mooly Beeri: basically the the goal was to implement the same methodology which I implemented across Philips, but this time in an automotive setting. And Active was still is a big power Powerhouse in terms of software engineering in the automotive world. And they basically worked with me to implement across the corporate the same methodology, the same system that I'm deploying now with better software across multiple businesses. And that's

Carmen: Mm-hmm.

Mooly Beeri: that's meant to help. the business in three ways in achieving faster time to market in achieving higher quality and in being able to deliver both of these in with lower costs

Carmen: Okay.

Mooly Beeri: as as the automotive industry is facing more and more cost pressures so

Carmen: Mm-hmm.

Mooly Beeri: Yeah, so I was working with all teams there across the globe for almost four years. and before moving on to other clients, I've also built a small team there that would keep the same framework going within the company. It was a nice,

Carmen: Mm-hmm.

Mooly Beeri: nice experience and a nice way to see that the framework that I've developed with the help of two colleagues in Philips for

Carmen: Mm-hmm.

Mooly Beeri: healthcare is applicable also in the automotive world.

Carmen: Yeah, absolutely for sure. There are definitely many ways how the development of software can automate and also improve different elements, whether it's in healthcare, whether it's in the automotive space as well. how big was aptive actually? Like how many employees are there?

Mooly Beeri: it was quite big, I think. there were about five thousand software engineers across the company and

Carmen: Mm-hmm. Okay.

Mooly Beeri: I wasn't working directly with five thousand engineers, probably with all the

Carmen: Yeah, of course.

Mooly Beeri: managers at some point in time. And it's similar scale to the scale that we had in Philips where I I think we worked with three thousand five hundred engineers worldwide. this this this this framework and approach to software excellence it scales quite nicely. I mean right now

Carmen: Mm.

Mooly Beeri: I have clients where We have 10, 15 software engineers and it's working perfectly well for them at a small scale. But it also works well, I'm working now with Mercedes, for instance,

Carmen: Okay.

Mooly Beeri: and it's working also well on on much larger scale.

Carmen: Okay. Yeah, that's very interesting. Yeah, I was really curious because before I worked at Cherry Assistant, I used to be part of a very similar company as well. I work for it was a sussip subsidiary of Volkswagen. So we do quite similar things, very similar company size where we really work on developing full stack to basically improve the sustainability. to also a scale the autonomous driving and of course the general it was I forgot it has been a while back, but generally improving the overall user-friendly experience by developing car sharing apps within the Volkswagen group as well. So yeah, it was really interesting to

Mooly Beeri: Mm.

Carmen: be able to connect a lot on that because I'm like, okay, I've had really similar background before moving into cherry assistant. Okay.

Mooly Beeri: Yeah, and I think the this experience with Active is is still keeping the business going quite nicely. Because Active was quite powerful in terms of the capabilities in software, a lot of

Carmen: Yeah.

Mooly Beeri: the leaders in Active moved on to become engineering leaders in other companies. And while

Carmen: Yeah.

Mooly Beeri: they moved on, they are very often calling calling on me to

Carmen: Yeah.

Mooly Beeri: replicate the same thing in their new company. So

Carmen: Course.

Mooly Beeri: I after aptive right now I think we have three at least three more clients where the

Carmen: Mm-hmm.

Mooly Beeri: person I'm working with is the same person whom I worked with in aptive but they

Carmen: Okay.

Mooly Beeri: just moved on to a different leadership role in a different automotive company. So

Carmen: Mm-hmm. Yeah, that makes sense as well. Well, let's talk a little bit more about better software.dev. So, as you we've discussed a bit earlier, you've had all of these experiences in terms of growth transformations, helping to scale at across different industries, automotive, healthcare, and tech. What inspired you to found, take that step and found your own business rather than working for you know, for and supporting other different organizations.

Mooly Beeri: Yeah, so I I I worked for a few years with Philips on this and then I worked for a few years in Active and it was basically driven by by by the need because I started getting requests from people who left Philips or left Aptive whether I can help them replicate the same thing.

Carmen: Mm-hmm.

Mooly Beeri: in their company. And at the time it was a the it wasn't a thing, it wasn't a framework. It was myself doing some work. And then I had requests to do the same work in in three or four different places. So

Carmen: Mm-hmm.

Mooly Beeri: that that pressure kind of led into the formalization of the approach, making it a framework, making it something that can be deployed not only by myself but by also by people that I train and starting off as a separate business that can serve and is serving now multiple customers

Carmen: Yeah.

Mooly Beeri: with the same approach. So it

Carmen: Okay.

Mooly Beeri: wasn't it wasn't so much as I was I woke up one morning and said I I feel like it's a good day to start a business. It was more like

Carmen: Yeah, of course.

Mooly Beeri: driven driven by the demand which

Carmen: Yeah.

Mooly Beeri: which basically forced me into I mean I could I could have chosen just one of those businesses and go work there, but I wanted to

Carmen: My friend.

Mooly Beeri: do more than one and then

Carmen: Course.

Mooly Beeri: this this constraint led to the formation of the business.

Carmen: Yeah, of course. It's something that doesn't happen overnight because it's an accumulation of a lot of businesses, clients reaching out to you to help you to to help them to replicate the framework as well. So if let's say the listeners tuning in to this episode and they were to hear your business better software.dev for the first time, if you could summarize down to one Particular mission, what it is that you're trying to achieve. what is it that you're trying to support for the different businesses?

Mooly Beeri: Yeah, so w we're helping engineering leaders become world-class engineering leaders. And the the main thing is that we're we're helping leaders focus on improving their business with the same rigor they focus on delivery. So the improvement of the business, making the business grow and excel, needs the same attention as delivering a product or delivering a milestone or delivering to customers. And so we basically make the improvement an integral part of day-to-day work and By doing that, we basically deliver about thirty to forty percent fewer customer issues every year, about fifteen percent faster time to market, and almost a third less in terms of cost of non-quality.

Carmen: Okay.

Mooly Beeri: And that for for any scale of business. So we're we're basically our our mission is to make more and more leaders world class software leaders.

Carmen: Okay. And what kind of cuss clients have you worked with? Which backgrounds are they from?

Mooly Beeri: well we have clients from healthcare, from automotive, from retail. and I think also one from the financial world. and that's probably what we have done so far, yeah.

Carmen: Mm-hmm. Okay. And from just partnering up with these different businesses, from the experiences that you had, so I see that you found it in just around five and a half years ago, what would you say from the five years of working and having exposure with the different clients, what would you say has been the biggest challenge that many of them face while trying to optimize and become more efficient? engineering solutions and so on.

Mooly Beeri: Yeah, it's it's the the biggest challenge is that most of the most of them are under pressure to improve of course

Carmen: Mm-hmm.

Mooly Beeri: and the transformations you the the the typical transformations they can buy require a quite a massive upfront investment they're quite disruptive

Carmen: Mm-hmm.

Mooly Beeri: and it takes quite a lot of time until you actually see the the results and

Carmen: Mm-hmm.

Mooly Beeri: And they are usually under pressure to deliver more with less, and they need transformations that will show an ROI immediately. And that's

Carmen: Mm-hmm.

Mooly Beeri: Basically the the program that we have in in better software is designed exactly to cater to those needs. So we are we've designed a program that doesn't require any upfront investment. So if you partner with let's say McKinsey and you want to do a transformation for your business, they would come in, charge you a hundred and fifty thousand dollars, leave you with a hundred pages of documentation and say good luck.

Carmen: Mm-hmm.

Mooly Beeri: in implementing those changes. And the way we work with clients is that we don't charge anything upfront. It's like

Carmen: Mm-hmm.

Mooly Beeri: a monthly subscription that you can cancel every month. And we

Carmen: Okay. Okay.

Mooly Beeri: measure the improvements and the financial savings on a month by month basis. And so

Carmen: Okay. Mm.

Mooly Beeri: the customer can see every month how much money they've saved actually by working with us and we're going

Carmen: Okay.

Mooly Beeri: to close this year with an average ROI of 8.5.

Carmen: Okay

Mooly Beeri: and so it's it's it's really catering to those those problems and most of the

Carmen: Mm-hmm.

Mooly Beeri: engineering leaders are always under extreme pressure to deliver more and deliver faster and deliver at lower cost so

Carmen: Mm, mm-hmm.

Mooly Beeri: These these are the typical challenges.

Carmen: Yeah, yeah. No, that makes a lot of sense as well, especially because when it comes to software, it does involve a lot of different components that make up for that as well. So in terms of the process, you mentioned that once they partner up with you, there's a monthly subscription. prior to that, how does the general process work? Is there, do you do they go into a discovery call and then that way you tailor up a plan? But how does the process generally work?

Mooly Beeri: Yeah, for the first four weeks we're doing a very detailed discovery. Of their engineering practices. We basically, maybe it's a bit too detailed, but we basically go through the software development lifecycle and we derive from the ISO standard 35 different competencies,

Carmen: Mm.

Mooly Beeri: which span the entire software development lifecycle from analysis to design, implementation, testing, maintenance, and release. And we assess each one of these 35 competencies. on an effectiveness scale from A to G, which

Carmen: Mm-hmm.

Mooly Beeri: gives us at the end some kind of a heat map of the organization where we can see how effective the organization is across the entire software development lifecycle.

Carmen: Yeah.

Mooly Beeri: and this gives us a very good starting point to understand where this specific organization needs to improve, what are the bottlenecks,

Carmen: Mm.

Mooly Beeri: what are the top three areas which actually require improvement. Because

Carmen: Mm-hmm.

Mooly Beeri: The standout approach to improvement today is is not very accurate and not very scientific. If you look at engineering leaders, they will allocate improvement budget to each of the department more or less equally and hope to get some benefit. we call this the spray improvement. you just

Carmen: Okay. Okay.

Mooly Beeri: try to spray a little bit and hope you get something. But

Carmen: Yeah.

Mooly Beeri: most of the areas are either already good enough or they're not the critical path, they're not the bottleneck for the organization. And so

Carmen: Hmm.

Mooly Beeri: When we are able to identify the critical path, the one or three areas which are holding back the entire organization, we focus all of the improvement energy into those, and that's how we get very quick results for the customer, because we only focus the improvement, all the energy on the one to three items that matter the most. And we do this iteratively. So every time.

Carmen: Hmm.

Mooly Beeri: We finish an iteration, we measure the performance again, we identify the next three items that require improvement, and we just solve them in the next iteration. And and so that that becomes a very effective way of structuring improvements. We're actually saving a lot of effort because it happens quite often that

Carmen: Okay. Mm-hmm. Okay.

Mooly Beeri: I Within these discovery sessions, we find out that the organization is in the middle of purchasing a new tool to manage requirements. They're going to invest

Carmen: Yeah.

Mooly Beeri: $200,000 a year in a license to some new requirement management tool. But then we analyze how many defects are associated with unstable requirements.

Carmen: Okay.

Mooly Beeri: How many problems result from immature requirements? And we see that requirements are not the main problem for that organization. They are accurate

Carmen: Mm-hmm. Okay.

Mooly Beeri: at 95% of the time. And

Carmen: Mm-hmm.

Mooly Beeri: therefore, of course, you can spend 200k on replacing your requirement tool, but the benefit to the organization at most would be 5%, because requirements

Carmen: Mm-hmm.

Mooly Beeri: are already 95% of the time correct. And so

Carmen: Okay. Mm-hmm.

Mooly Beeri: you're better off spending these 200k. on some other area, maybe on test automation, maybe on continuous integration, maybe on coding agent, but that decision can only happen after you have the exact diagnosis of the business performance across the software development lifecycle.

Carmen: Okay. Yeah. One of the things that stood out to me was that you mentioned a while back one of the metrics that you use to measure, to show one of the reasons why the clients should continue to partner up with you is the fact that they are saving a lot of financial expenses as well to show that they've been really strategic in terms of investing and also knowing how to invest and optimize. What would you say are the other few key indicators that you use to measure the s success on how well they're doing in terms of scaling in terms of the transformative solutions.

Mooly Beeri: Yeah, so we we measure with them a couple of key performance indicators. first and foremost is the amount of time the team is spending on non-productive work, on technical debt, on items that are not value added. So every time the team has to spend time on fixing a bug or analyzing a customer ticket or reworking some requirement that wasn't properly defined, this is wasted effort. This is work on technical debt. And so we measure the percentage of time the team spends on non-productive, non-value-added work. Then we measure customer satisfaction, we measure the cost of a defect, the cost of bugs in the system, because we see The cost of defects is exponential and tied to the moment it was injected versus the moment it was detected. And so we understand what's the injection point and what's the detection point for most of the bugs, and we calculate the cost of a defect, and we and then we work together with them to reduce all of this. So reduce the number the absolute number of defects that are generated in the system, then reduce the gap between injection and detection, and and by that reducing the the overall cost of defects and allowing the team to spend more of their day on value added deliveries and It's it's very typical that we show 40% reduction in defect count. I even have one extreme example, it's not a typical one, but it's it's a real one with a team that I worked with last year in Mercedes. They had a special special case. I won't go into the details, but in three months by analyzing their specific constraints. We were able to reduce their incoming defect rates by 90%. And imagine imagine

Carmen: wow. That's a huge difference. Yeah.

Mooly Beeri: the team now ha having 90% more time to. And so there's more capacity, and then management can decide if they want to deliver more with the same

Carmen: Mm, mm.

Mooly Beeri: team, or they want to shift people to other priorities and deliver the same with a smaller team. So we we we basically free up capacity for. the the company.

Carmen: Yeah, I think capacity is definitely obviously one of the most common challenges that I also hear whenever I communicate or talk to different clients of ours as well. It's always about capacity, how to prioritize, optimize so that we're able to prioritize or focus more on the business growth and the business output as well. Speaking of capacity, let's talk a little bit more about your team setup, team structure. currently, how many of you are in the team, and what would you say are the main roles and functions that you have allocated and divided?

Mooly Beeri: Yeah, so we have we have a group of coaches that are working alongside myself. I'm also working as a coach with teams. And I don't have any full time employees, so these coaches and everyone else is working as a contractor and they basically can work more or less depending on the load. Right now apart from me, there are two other coaches, and we also have a project manager and a data analyst and the chief of staff. and they're all coming in and out depending on load. and so the model is quite scalable. I have a few people quite senior that are trained on. the framework on the model and they're very happy to take on teams as we as we expand. they have their own day job. Maybe at some point where we have some long-term contracts with large scale companies maybe there will be a few people that are full-time working with with us but for now this is the setup yeah

Carmen: Okay. So most of them are contractors or they're here for short term project based project, depending on the needs and the priorities.

Mooly Beeri: Most of them are working with me for a long time, a really long time, but they

Carmen: Okay.

Mooly Beeri: they the the their their load changes throughout the year depending

Carmen: Okay.

Mooly Beeri: on the number of concurrent clients that we have. But it's not it's

Carmen: Okay.

Mooly Beeri: it's usually the same people. I mean I have a a group that is is going with me probably three, four years already. Yeah.

Carmen: Okay. And where did you how did you know them? Did you is it through word of mouth or did you recruit them? How d how do you bring them on?

Mooly Beeri: I've worked with two of them in the previous company. actually with three of them I worked in the previous company to

Carmen: Okay. Yeah.

Mooly Beeri: a certain extent. One was just we we opened a position and we did some hiring and we we found him and another lady yeah w we had we had we've done some project together and then w We both enjoyed working together and she just stayed on as a project manager. So yeah.

Carmen: Mm, okay. Yeah, that's always it always helps when you have some familiarity of in terms of previous collaborations as well. Definitely makes a huge difference. So are they all yeah, sorry.

Mooly Beeri: Yeah, you know the so at at this point in in my in my in my career I have probably about ten thousand connections on LinkedIn. And if I need if I need some some capability, I I I must be already connected to someone who has that capability. The this illusion that I can just find some random person that that can that that can serve this need and I don't have anyone in my network who can do that. It's always much

Carmen: Mm-hmm.

Mooly Beeri: much easier to work, especially if you work remotely, work with someone

Carmen: Mm-hmm.

Mooly Beeri: that you have already worked with physically and you know the person and you've you've got some kind of relationship on going already and and so it makes it much much easier to start collaborating.

Carmen: Course, and also about building the trust as well. yeah, because most of the time, something that I've also learned as well as a recruitment partner is that you can find people, they have exactly all the qualifications and the skills that you're looking for, but it's also about the the collaboration. If you can build this rapport or this trust to be able to collaborate with that person, that makes a really huge difference as well.

Mooly Beeri: Yeah, at the end of the day we are connecting through Zoom or Google Meet. It's about the quality of communication and whether you can you can both understand each other and spend relatively little time in explaining what you need and get the right output. So the best people I work with, they need very little guidance, they already understand what I need. and they just go off to the work, I get the the result and it's 90% in the direction of what what was needed and so you get a lot of value with relatively little spend in terms of direction and instruction and so on.

Carmen: Yeah, and of course that will prevent any potential delays in the process as well to be able to deliver the output that you're looking for to. Yeah,

Mooly Beeri: Yeah, exactly. Mm-hmm.

Carmen: okay. Well, looking at your team setup right now, you know, you mentioned that perhaps in the future you're thinking of perhaps getting more long-term support. Looking right now, if there were any aspects of what you're working on at better software dev, where you think, okay, I'm spending a lot of time working in this particular function. If I'm able to delegate it immediately, if I had the right support or the right person to trust, what would it be?

Mooly Beeri: I think our setup is quite is quite efficient as as it is now. I can't say that I have a specific need right now. I'm I'm experimenting now with delegating some work to my AI agents, so that seems to be working quite nicely. and yeah, but I don't think there's any specific struggle right now that requires some some immediate attention. I think our setup is for the most part working right now.

Carmen: Mm-hmm. Okay, that's really great to hear. You mentioned something about you you're delegating some of the operational aspects to agents. Are they also remote? Where are they all based actually, your team and your support?

Mooly Beeri: my team is is quite global. I mean I have a couple of guys in the UK, one in on the West Coast in the US, one actually in Hawaii, another

Carmen: Okay.

Mooly Beeri: one she based Serbia

Carmen: Okay.

Mooly Beeri: so it's it's really global global theme

Carmen: Yeah. Okay.

Mooly Beeri: and one is actually right now for the last month he's been traveling in India but yeah so

Carmen: Okay. Yeah.

Mooly Beeri: it they're they're they're all over the place

Carmen: Mm-hmm. And in terms of like the different time zones, do you have some core working hours or is it all flexible?

Mooly Beeri: Yeah, it's it's we we each of them has their own assignments and we have with each a weekly catch up and we adjust. I mean if they travel then we can we can find a different time. I'm I'm luckily I'm more or less in the middle so I can adjust my day. actually this morning at nine

Carmen: Yeah. Yeah, course. Hmm.

Mooly Beeri: nine AM I My first meeting with the team in China, I think it was there four PM.

Carmen: Mm-hmm.

Mooly Beeri: And then what is it like two hours ago I had my meeting with the team in North America on on the west coast. So I think

Carmen: Mm-hmm.

Mooly Beeri: it was also there nine AM. I'm not sure. But anyway. And yes.

Carmen: Yeah, yeah. Because you're right in the middle as well. China's like seven hours ahead and the East Coast is like seven hours behind. Yeah. Mm-hmm.

Mooly Beeri: Yeah, exactly. So so we're we're pretty flexible. And anyway, most of the time the work is not done in in in the meetings. I mean the meetings

Carmen: Mm-hmm.

Mooly Beeri: are just to ensure we're all walking in the same direction and

Carmen: Mm-hmm.

Mooly Beeri: maybe provide some some guidance, but most of the work is is offline. I mean I

Carmen: Mm-hmm.

Mooly Beeri: For me the the right balance is I I speak to someone for thirty minutes and they're off to do a couple of days worth of work,

Carmen: Yeah.

Mooly Beeri: and we can look at the the the result.

Carmen: Yeah, yeah, that makes sense as well. It's all about efficiency and being intentional, purposeful in those meetings, too. So, what I love to learn and I love to jump on to a different topic is the topic of AI, right? So we are now in this day and age where AI is surrounding us, AI is everywhere, we're also being asked to stay informed, utilize it from our day-to-day business and so on. How are you thinking about AI? Within your business.

Mooly Beeri: Yeah, it it's it's it's evolving. about a year or two ago it was your profession is going to be eliminated, no one's gonna need software engineers anymore. And I had friends who stopped recommending to their children to go study computer science because

Carmen: Okay.

Mooly Beeri: it was thought that the profession is dead.

Carmen: Mm-hmm.

Mooly Beeri: But I think I think as we evolve we we we we understand the capabilities, we also understand the limitations. I have many clients using AI in the different phases of the software development life cycle. Some are using it to elaborate requirements, some are using it to write code, some are using it to test or to code review or deploy, etc. And and and so I think I have two two perspectives to share on on AI. First is the tool itself or the technology itself is not precise. Okay. It's there's there is a saying that I like, which is when I use AI on something that I don't really understand, I'm really impressed. When I use AI on something I'm good at, I'm always very critical or very

Carmen: Mm-hmm.

Mooly Beeri: disappointed. And so the the capabilities are incredible f in getting you from zero knowledge in some area or zero capability up to 60-70%. So if you want to get quickly a prototype, if you want to do a quick research, if you want to know something you don't know now, AI will get you to 60 or 70% in

Carmen: Mm-hmm.

Mooly Beeri: in probably the the fastest time we've ever seen something in in in in history But that also creates the illusion that you can use it also to get from 70% to 100% accuracy. And this is

Carmen: Mm-hmm.

Mooly Beeri: where we spend most of our time because it's just not designed to give you a hundred percent accuracy. And then

Carmen: Mm-hmm.

Mooly Beeri: people start fighting with trying to get better prompts and trying to get the feedback loops and trying to adjust the the the agent network and put more agents in place and different Different types of LLMs and and so on. So first thing to really understand is that what it can do for you and

Carmen: Mm-hmm.

Mooly Beeri: how to use it in a way that makes the most out of that specific capability. So

Carmen: Thank you.

Mooly Beeri: I'll give you two examples from my world where it makes perfect sense to use AI and where it's difficult. So every company has a list, which is probably endless, of bugs and customer reports, and incidents, and all the findings throughout the testing cycles. This can be easily tens of thousands, hundreds of thousands of issues that are accumulated. And using AI to make sense of that data and find patterns in it and find reasons for. common reasons of problems that the business is suffering from. This is a great use of AI. It's really the perfect tool to find patterns within this pile of data where no human can read 10,000 bugs and come up with a pattern. Okay, it's it's really difficult. So this is this is great and I always encourage my clients to take this pool of data, give it to the AI and try to distill three or four themes of issues that are haunting their business. But when you try to create code, ask AI to write code, then you get inaccuracies in the code. And you either spend more time reviewing it by humans, or you fall into the trap of saying, I'll let AI review its own code. Which

Carmen: Mm-hmm. To save time, yeah.

Mooly Beeri: yeah, which which basically means you're postponing everything into testing. And then the test team might be saying, Let's generate the tests with AI, and they're also testing with AI. And so you find cases where the first time a human being actually interacts with the system is the customer. Okay, because everyone else generated their artifacts with AI. And then, of course, the customer, which is still human, is not very happy with the the result, and and and this thing usually collapses. So The second perspective I have here is that when you use AI, use it on an area in your specific organization where you need help. So it comes back to the understanding of where are your weak points. If your requirements are currently close to 100% stable, getting AI To write requirements is a mistake. It will lower the level of accuracy. Maybe it will give you a little bit more speed, but that doesn't matter if you end up implementing the wrong requirements. If your developers are producing really high-quality code, you can try an AI agent side by side with them and measure the accuracy of what they deliver or the defect density in what they deliver versus what the AI delivers. And so the the the advice that I give to most of my clients is first understand which area is the one that needs improvement, and then pilot the AI performance side by side with the human performance.

Carmen: Mm-hmm.

Mooly Beeri: There is another last trend that that starts creeping in is that the the AI companies are slowly starting to charge the actual price of their processing to clients. And so I see more and more clients saying, I could use AI here, but then I'm going to run out of my tokens or run out of my my budget, and I can't be spending ten thousand dollars a month on code review with AI, because

Carmen: Mm-hmm.

Mooly Beeri: that's that's more expensive than my own work during that that month. And so that also starts creeping in. And I think somewhere we will find the balance.

Carmen: Yeah. And it's also really about weighing the pros and cons of using whether it's like an experienced human or experienced person or versus something automated through an AI as well. That's also something to consider and compare and see what's whether it's more efficient to find something automated, quicker, and faster. But of course. There isn't a guarantee that there is a hundred percent accuracy or of course an experienced human who can of course still feed into the human element as well.

Mooly Beeri: Yeah, it's it's it's as I as I said earlier, it's the probably the third or fourth wave of new and exciting technology that I'm seeing in my career. And it it's it's following a similar pattern that you everyone gets excited and then it's starting to be abused and then the real price pops up. And then you start asking yourself, is this really worth the the price? The the the the the latest analogy I can give is when cloud computing became very prevalent, everything went to the cloud, and then companies started paying for the price of hosting all of the services over the cloud, and the the the prices were not something businesses could bear. And then came this realization of what do we actually need and is it actually does it make sense to host so much on the cloud and I think I've seen stories where companies are spending all of their AI budget on just converting Word documents to PDFs or

Carmen: Mm-hmm.

Mooly Beeri: doing some other things which are quite trivial to do in other ways. And The more the AI companies will be forced to show real profit, not just promises of profit, the more the cost will go on to their consumers and the more it will drive the use of AI only in the places where it's actually providing value.

Carmen: Yeah, definitely for sure. It's all about like connecting it into the right value, also into the right output as well. So, Mooli, just one final question that I always ensure to ask every single one of the my guests that are attending behind the founders is if You could share one or a couple of insights or advice that would be important for aspiring business founders or someone considering to start the entrepreneurship journey. What would it be?

Mooly Beeri: Yeah, I w I I think the the one of the best advice I I can I can give is focus on the problem you're trying to solve. Today it's so easy to come up with an idea and then spend your next half year on building that idea out and you have all the tools and you can spend tons of money and tons of time on converting this idea into an app and a website and promotion and marketing, but the there's there's not enough thinking of what the right problem is. What is the actual buyer motivation, what is it that you're solving in the world.

Carmen: Mm-hmm.

Mooly Beeri: And the more you spend time on that, the more you stay with the problem, the better the solution comes. And so that's that's one piece of of advice. And for for founders who are trying to start their own business, I would say the best starting point, there is there is this belief that you can just go. marketing cold calls, cold emails, broadcasting your views and and and get your first ten hundred thousand clients but if you're like me if you're in consulting if you're selling some service the best source of at least the first clients is you already know these guys these are people that you've worked with in the past

Carmen: Mm-hmm. Mm-hmm.

Mooly Beeri: they are you've already built the trust with them, they will hire you because they trust you. and that will be your first proof point. Then you can expand, but starting with the people you already know, that's already I think it's it's a good it's a good a good starting point.

Carmen: Mm-hmm. Yeah, definitely very valuable to focus within the current network that you have to be able to help you expand and grow and also build up the client base as well. Okay. Well, Mooli, thank you so much for sharing really valuable, helpful insights. And I'm sure that the listeners tuning into this podcast would definitely value and appreciate what you have to share as well. Thank you so much again for coming on to speak with me. This is Behind the Founders. Be sure to stay tuned for the next

Mooly Beeri: Yeah.

Carmen: episode. Thanks again.