Welcome to another episode of Tech Unhinged. 
I’m Usher and today we’re exploring the future   of propt tech driven by AI and data science. 
Joining me is Costenza Balgoni, a technology   leader with over a decade of experience driving 
AI and data science across innovation across   different industries. Currently she’s leading 
the data and innovation team at generally real   estate where she launched the city forward 
platform. She’s also worked in the past at   Ericson Ferrero and KPMG and was also recognized 
among the Republicans uh 50 Italian women in AI.   Costanza thank you for being here.
Thank you for the invitation.  So Costanza, why don’t you kind 
of kick things off? I know you’ve   uh you’ve built a career around data science, 
machine learning, etc. So um I’d be I’d love   to know how did you get started and how was the 
journey you know kind of evolved over the years.  I actually came out from uh Italian humanities 
high school. So like me and math and data were   really like two words apart. But then I decided 
to study quantitative methods. Actually I studied   quantitative methods for um developing economies 
and I studied in UK and in that in that period   of time during my my university years I uh got 
in contact with uh pro programming at the time   was mainly MATLAB and so a lot of things have 
changed but uh I I definitely uh fell in love   with the with the topic with the with the approach 
of of that type of work and I very very very soon   switched to uh very quantitative and very data 
and very programming based kind of um research   topic and studies and after I graduated I decided 
to continue on this journey and I started a master   of philosophy in Pisa um in big data and social 
mining and I would say um I really fell in love   with the topic and the very early years of my 
20s has been pretty much an obsession for me. So   technology it’s um eventually my whole life and 
it’s possibly my hobby, my work and and and and   what I love and what I like. And I’ve started on 
off as a quantitative economist uh in my career.   I’ve been working for public agencies in for Italy 
but also for international organizations always   like an econ an on a quantitative economy or um in 
statistical departments and then I switched um to   corporate starting with tech and then continuing 
with with consultancy and and then consumer goods   and now financial services. I would say that uh my 
career has evolved a lot in these years cuz like   um of course uh I started uh working in technology 
but I also saw a lot of innovation uh first uh   then I moved to into more operative and practical 
uh positionings with KPMG where I was leading   entire products and projects and then also again 
um in my experience in consumer market I had a   very rounded role where I could see many things 
um along all the production type of a product but   also of a project and so I would say that I saw 
the topic from many point of view which is not   all the only one of the uh data scientist but also 
of the product manager of the project manager of   the leader of the responsible the P&L responsible 
and so on and right now I am leading this team uh   this team is actually um very new cuz like um when 
I joined July in 2019 generally was eventually the   second big real estate asset manager in Europe to 
have this kind of team and or to decide to invest   on this kind of team meaning that uh I’ve been 
leading this team from day one. I built it and   uh it was a very interesting experience as um 
first of all it was my first time in financial   services which is was of course not super 
intuitive for me in the beginning especially in a   in a very particular field as the real estate but 
secondly it was very interesting because you know   uh real estate is really hor the horror and the 
dream of every uh every person uh who does my job   because eventually it’s a green field so there 
is there is nothing there is no reference and   uh you need a lot of creativity but also So 
a lot of understanding of the future needs   um and you really and you really have this great 
privilege of defining what data driven means in   an industry like this that does that has long 
relying on intuition and business expertise.   So it’s a very fun job. There is a lot of things 
to build and uh so far I think we have we’ve been   doing a decent job and um we are loving it along 
in the way. Awesome. Awesome. So can we kind of   safely say that you’re living your childhood 
dream? But well it’s not my childhood dream   because probably when I was a child I wanted 
to be I don’t know a doctor or something else   but I would say that my university dream yes
university dream yes because know I I often talk   to people and then I because I wonder that myself 
as well that hey if I if I if you had a chance   what you be what would you truly be doing right 
now? So let’s say if you if you rethink everything   is this exactly what you had envisioned or not and 
a lot of times you learn you learn what you really   want later down the life as well right so
yeah that’s right  it’s a great privilege it’s a great privilege 
to be I’m always saying that is that is a great   privilege to be um in something that you would 
do it anyway so you get paid to do it and it’s   awesome but but eventually I would do it anyway 
absolutely so as you mentioned that real estate   um is a green Absolutely. There’s not a lot 
of there’s there’s no bar set. I hope you’re   setting one, but you know generally where and 
what things can be done. So but but I it’s it’s   a very traditional industry. Um you know you have 
a very traditional way of thinking but we and then   and applying principles and then it’s not like 
uh stock markets or forex you know exchange or   currencies where you’re still dealing with 
data you still have numbers you still have   everything but it’s kind of in that sense. But 
in the last few years I I feel there’s there’s   been some exciting changes you know the way they 
have kind of digital transform yourselves they’ve   um started looking at data correlations between 
you know an asset itself and how it’s providing   value and and all of those things and hence it 
kind of came into that stage where this this   entire transition was referring being being 
now referred as propt tech generally right so   um do you think enough has happened to justify 
giving it a term of its own. There are many   things to say to answer this question, but I 
would focus on a couple of them which in my   opinion are the most important. The real estate 
it’s a rather simple business if you think about   uh it’s like uh you know you buy something 
that hopefully will all the certain values   um and will be leased or will be sold speculative 
at a at a better figure you bought it. But it’s a   quite straightforward business. It’s it’s not 
made of a lot of data. It’s not made of a lot   of physical processes along the way. Of course, 
it’s not a simple business, but it’s not a yuber   complex business. I would say we are pretty 
straightforward uh how the business model works   and how the operative model works. It doesn’t 
make it easy, but from a from a data perspective,   we are very far from other industries. Sure. 
And also the real estate holds a very specific   um characteristic that if you think about 
is the only asset into the financial market   space it has a physical presence. It exists. 
It’s it’s something you can touch. You can see  like even in private equity in the alternatives 
you have a stake of a company. The company exists   but what you have is not that physical. If 
you think about this two characteristics   eventually were both uh I think uh keeping the 
business itself a little bit behind in term of   digitalization but are the sames that are pushing 
now forward the the the digitalization throughout   the different initiatives that go along the way of 
this big perimeter that is at the propt tech and   uh I think that uh overall the quantity and 
typology of data that have been taken into   consideration within the decision process 
of the real estate were quite few and   uh we I think that we did it maybe first in in in 
Europe but now a lot of companies are stepping up   on this it’s physicality of the real estate that 
needs to be taken into account also crucial in the   world we’re living now where things are changing 
very frequently and there is for example all the   great topic about the alternative data usage 
for investment support so thinking that you   need to understand that an an asset a building is 
within a it’s like in a living organism influenced   by what is around but is also influencing what is 
around. So I think that this is a very fascinating   perspective right now especially for all the 
predictive models and all the way we support that   their decision to build more resilience portfolio. 
So in general that that part of the data science   and also of the AI somehow are really interesting. 
I would say they are rather unexplored. Uh the   alternative data we actually did a lot with 
city forward but I but there are a lot of   niches and spaces and piece of processes that can 
be yet explored more and and developed more and   understood more. I think also that the AI in this 
regards will allow us to really exploit a lot of   other datas, a lot of other u somehow correlation 
and modeling exercising around the physical world   which makes it very exciting and very cool. On 
the other hand, the uh meant as the LLM models,   the foundational models and so on of course will 
are going are and are going to impact their real   estate as well as other uh asset class as well 
on all the autoation side which will be I would   say universal multi-industries. or optimizing 
processes, automating pro automatizing processes   as much as um as possible to really get into the 
differentiated value that is performed by the   human being. So I’m thinking about all the work 
of reporting but also the data management but also   some part of our um financial processes and so on 
that can be very very very much exploited with AI.   uh leaving the people with what now the industry 
likes to call high value added activities which   in my opinion is not really about low value added 
activities and uh high value added activities it’s   more what is the eye is very good at at the moment 
and it’s very good also in doing what we thought   were high value added activities for example 
creativity we always saw that creativity was   a high value added activity while AI is very 
good at doing it for example and uh and so I   would say more I would I’m always speaking more 
about ambiguous and non-ambiguous processes. So   the processes that can be explained easily with 
a flowchart without many exception. These are the   processes that in 2025 a machine should do it.
I 100% agree. In fact that’s that’s the kind of   conversations we have with modus the people that 
hey look at your problems and then see which ones   are the ones that you can very clearly articulate 
and define and then yes those cases best cases for   AI agents. So before we kind of go into that 
further, I’ve seen in conversations a lot of   times people mix AI data science and they kind of 
talk about it together and there’s this confusion,   right? So how they complement each other and 
everything. Can you kind of help explain and   uh you know listeners can also understand that 
what’s the core difference between the two?  Let’s say that AI eventually u was already part 
of the data science uh long before the term AI   became pop. Um yes so machine learning itself 
and because AI eventually artificial intelligence   meaning transferring to machine capabilities of 
humans and uh we have we have been already using   uh since the 50s uh models of machine learning 
that were actually teaching machines to do human   performances uh long before CH GPT and GPT models 
we had NLP um models that were already working   into the tech space and the generation space. 
Um I think that from a technical perspective   what changed a lot is that the paradigm on how we 
were working in data science really changed. Um I   always say that we we actually started our career 
with a sort of linear way of doing data science   that was actually following a certain pattern of 
consequential algorithms that were then used and   developed and then deployed um the LLM process the 
attention process so the transformers themselves   and really made this things parallel which is very 
different and it’s uh very interesting. However,   um I feel that the number of use cases into 
the corporate world. So I’m speaking about   the corporate world with the exception of very 
few industries for example the tech, some part   of pharma, some part of energy where technology 
is really part of the business uh bringing you   topline. they are possibly somehow putting too 
much effort into or too much effort or too much   um I would say hope into a transformational 
effect of the of the AI itself without having   done I would say the foundational with machine 
learning and and eventually somehow even process   re-engineering. So this links with what we 
were saying before the AI is not magic wand   and also AI deployment AI as LLM foundational 
models and transformers are actually a lot   uh less um usable in corporate perspective a 
lot less than the current organization think.   uh right now uh we have uh many important task 
for enterprises for example capabilities of   extracting data from SQL code the best 
model on the market right now it’s giny   2.5 experiment I’m not if not if I’m not wrong 
and on complex queries the performance are 56%.   Think about the reporting. You what do you need 
for reporting? You need an agent that is able   to extract this data. So natural language to SQL 
then pivoting this data and then produce a reboard   or be able to produce chart and visualization. 
If already on the first step of this process,   the LLM will fail 50% of the time.
Yep.  Is possibly not very transformative. So somehow 
u I think we have to have a very curious and a   very proactive approach to this technology. I’m 
speaking again about enterprise environment,   corporate environment but also very realistic 
approach and always remembering that the level   of sophistication of the data science right 
now the old data science so the old models   can be very yeah the old data science can be 
already very beneficial for a lot of businesses   um that look both at automation somehow but also 
a decision support a decision science uh tools  awesome so I’m guessing So, Gen AI generally 
has now really helped or stormed to an extent   the problem of NLP, right? So, previously you had 
to spend a lot of energy, a lot of time building   an NLP model. I think that’s the biggest thing 
that they’ve actually been able to accomplish   is that make that task fairly menial, fairly 
easy to do, fairly there will be associations,   there there will be errors, but at least 
it works 60 70% of the time, right? So,   I guess that’s the biggest impact that 
that you can see that Genai has brought in.  Yeah. the productivity for of course 
it’s very important and then again on   my sector meaning the technology sector 
possibly the impact as being very strong  yeah I was telling I was telling that people 
my age uh or people that started working only   six to eight years ago debugged every time and 
debugging was also an exercise of learning and   uh right now I haven’t have a single trainee that 
debug manually somehow but it’s not necessarily   a bad thing um I think but I I think shifting 
a lot in terms of productivity and the amount   of things once one one single good individual 
can achieve but still uh if we think about it   as a support for productivity absolutely I’m 
in I’m 100% in uh when we’re speaking about   big corporation uh really transforming the 
way they’re working today I think there is I   think that it’s easier to say than to do with 
the current landscape of offer at the moment.  Yes, absolutely. I think you uh so I’m 
actually very happy you brought up the   debugging part because back in the day and this 
is I’m talking I think u approximately 17 16 17   years back my first project was uh there was this 
time in 2006 78 that there was this rich internet   applications concept right so you had these 
you didn’t have react and all those things but   world was advancing kind of in that middle stage 
where it was going away from flash and trying to   do animations and all that stuff in JavaScript 
so there was this language called open lazlo   and uh I uh I was I was working and there’s a 
project that came in. It had about 300,000 lines   of code and an Excel file with 150 bucks and that 
was my assignment. So, so I spent my entire summer   uh practically working 18 hours a day just you 
know opening a Firefox, opening the debugger   and pressing F10 F10 and kind of debugging the 
entire software. But when I talk to you know   uh when we hire people here and they kind of talk 
to junior engineers or uh you know even five six   years people I tell them that was the biggest 
lesson of my life because I now whenever somebody   anybody used to talk to it’s been a while since 
you know I’ve engineer I’ve done programming but   um when anybody talks to me I can you know just 
just break down the problem very easily because   it’s just natural right and I think to make use 
of genai and that’s the strength you need if you   don’t have that strength in your engineering 
capacity to to think of a problem, break it   down into steps and then leverage AI, you’re in 
trouble. Uh otherwise, because you can give a   prompt, it will code and it will put together 
a decent enough prototype. But you know, if it   goes into production and there’s a bug, you’re 
stuck, right? So because it’s a code written by   somebody else. U so that that that definitely 
kind of resonated that example. Absolutely.   Absolutely. So all right so moving on u let’s 
talk about uh the product that you mentioned the   city forward platform that you made in general 
um can you you know kind of highlight what are   the use cases that you worked on it and how did 
it impact the overall organization the business  so um when I joined general and um I need to 
acknowledge the fact that everything of this   was possible especially in a very big financial 
institution like the ones I work with also   because uh my CEO is a very visionary person. 
Mr. Madzoko has always strived for technology   excellence and also the idea of having this team 
inside and eventually treat us like a startup and   let us fail iterate and start again and it was 
it was essential to came out with good ideas.   Uh long story short um when I joined I wanted to 
make an impact that was really visible from day   one and I wanted also to build a product that 
would make generally somehow special. So this   was my my goal eventually it and I spent the 
first year and a half finding trying to find   something eventually because it wasn’t easy. 
I needed to understand the business model,   how we were making money, how we were managing 
the properties and so on. So I had this idea in   my mind. Then I couldn’t really find the the bug 
where technology may be useful until I studied   with more attention the investment processes 
and in particular the way we were assessing   the performance of an asset over time. Okay. And 
what what we discovered is that roughly 60% of the   change in value over time wasn’t explainable using 
the traditional real estate metrics prime meals,   prime rent, capital value and so on. This was the 
first intuition in telling us that maybe maybe   uh some of that explanability was somewhere else. 
So we started looking at some very very very very   embal research that was done in this in this 
field and we started really roughly collecting   I actually started because I was alone at 
the time um roughly 20 I focused on Paris   um we had a decent amount uh we have actually 
a big bunch of our portfolio data but but I had   that I had that a decent amount of sampling that 
I couldn’t find and then uh I started gathering   from open sources is um roughly 20 GIS uh type 
of data. So uh geoloized data okay in a latitude   and longitude format and it was some POI that 
I extracted from Google. Then there were some   um some uh information about the city that were 
provided by the statistical office of the city.   Then I had s some social demographics. Then I had 
some open information about traffic about uh green   areas and so on. and I built the first model. 
I was developing in the beginning with a PCA   u methodology. So uh the first model eventually 
performed extremely good. So really I started   working on these things I would say a couple of 
weeks not not more than that we were in lockdown   and uh I can remember very well and I was really 
surprised really like very simple setting in my   notebook this things would would work so well not 
only in terms of accuracy but also the diagnostic   of the model itself. I come from a statistical 
background. So the diagnostic of the model itself,   the robustness of the model itself was really 
really convincing and uh and that was really the   start. So I came back to my leadership and I said 
I may found this may this be of interest meaning   every time we invest in something new every time 
we disinvest in something or when we are looking   at the or where we are looking for example at the 
evolution of a city where we don’t have commercial   presence in we don’t have people in places we 
don’t know yet where we don’t have offices yet   this is this possibly useful for us to also think 
outside the box to fight sometimes the convention   of humans that are based on our own convention 
that sometimes is not really data driven and   thanks god my co was agrees agreed with me so 
we started on this journey of course the thing   became quite more complex because we now we’re 
speaking about 800 variables uh we also added   the mobility data we added a lot of data regarding 
uh consumer spending uh so we have transactions   uh we have real estate data we have social 
demographic we have point of interest we have all   of it for the last 10 years we have paired it with 
more than 30 machine learning models. The idea of   the usage of of city forward inside the company 
is pretty straightforward. An asset manager or   an investment manager can use city forward on a 
descriptive way. So exploring the city looking at   the trend understanding the composition of that 
geographies where the asset is in in a very very   micro look at a very very micro location uh level 
or they can use it to predict the performance of   the building in the next 5 to seven years. So 
to understand how much that particular asset   will grow in the area and most importantly 
why so which are the the variables that   are contributing more to the performance of the 
asset both negatively and positively positively  and and and this is I think is very interesting 
because in the beginning the product wasn’t   really like you know loved by everyone 
but I think the the biggest reason why   is that it challenged a lot the convention.
Yes. um it challenged a lot the expertise   uh but I think the tool like this especially when 
we’re looking at tool that predict that support   the decision process what we call decision science 
are usually less easy to adopt because they don’t   solve a today problem but they will help you think 
for the future which I think it’s a very different   kind of objective and so I think that um this not 
always intuitive results you will get from city   forward but of course the more interesting ones 
but also So somehow the most discussed one for   example we have relying a lot on the ideas that 
retail strives in places that are very touristic.  That’s the gut that you that 
you know that speaks that hey  the gut yeah the gut tells you that of course 
places that have a lot of tourism usually have   retails that tribes. While this was true for many 
years, after COVID things actually shift and data   are telling us that things shift actually overall 
income of residents became a lot more important   than the tourist influx within certain areas. 
Yes. Which is really counterintuitive but is   eventually important to understand for future 
decision. Overall we need to understand that   our overall objective is to build resilient 
portfolios and even if this business is really   about the scarcity of the geographies we’re 
interested in my co always say that you don’t   have two shams in the world you don’t have two 
montapon there is one and that’s the space but   even if this is true cities are changing people 
are moving the way of we work is changing and this   affects all over the asset class, the retail, 
the residential, but the offices as well,   hospitality, industry, everything. Yeah.
Everything. So, um, all this phenomenon   needs to be taken into account when looking at the 
future, we’re looking on the medium and long term.   And I think that a tool like this has been able 
to give this kind of support in the discussion,   in the decision, in the brainstorming, in 
the revision. And also what came almost   immediately that this tool can be of 
interest for a lot of other industries   retail for example where should I open my next 
franchise where should I open my next popup  but also physical ads industry where are the 
places of the city where I’m is most likely   that the target I have in mind will see my heart 
will see my campaign how much return I’m going to   have from that campaign or for a public sector for 
example understanding where should I open public   services based on how the citizen are moving 
around the city where families are going where   single people are living and so uh very soon uh it 
was clear that the potential of city forward was   a lot bigger than just the real estate and right 
now city forward in its public release phase where   we are serving also other clients outside the real 
estate business with customized products based on   the location intelligence provided by city forward
that’s awesome Actually that’s I I’m sure you know   when you started it you never really kind of 
thought that hey this might end up pivoting or   this might end up going in this direction. We 
have a saying going in the office where we say   uh in God we trust, rest all others bring it.
All others must bring data.  I have it I have it on top of my closet uh 
behind me in the office. I have this thing.  I have these things over my head.
In fact, there’s one of our we also working with   another company who actually solved this problem 
that hey, if I’m looking to open up a a retail   shop, where do I open it? And that’s also kind of 
a similar conversation. It’s it’s true that in US   this kind of technology is not really new because 
also the flow of information in US are different   and uh there are amazing startups also right now 
like I would not call them startups I call company   uh that are providing similar uh services but I 
think that uh what we did was really unprecedented   in Europe because the difficulties of collecting 
this data in Europe. I was actually about to ask   you guys had to work in that direction as well 
or was kind of publicly available or Okay.  No, we we had for each single country we’re 
present. And so we have 11 uh cities in Europe   plus four for all countries was a very like 
tailored approach to have the same type of data   somehow with the same consistency quality 
granularity but having in mind that every   country as it’s shown for example real estate 
systems I don’t know if you have something you   have like Zillow in America in Europe we have
one for Italy one for Spain and so for example   all the real estate ads needed to be somehow 
compatible but we couldn’t uh source them all   from the same providers. Same with a lot of 
local data. We actually scale with a couple of   uh variables meaning that the POI and the social 
demographics are now coming from a single source   for the uh perimeter of geographies we have but 
it was a very long a couple of years long work to   really build this data infrastructure that was 
robust enough that was uh accurate enough then   powered the algorithms. one of the challenges 
that they you know people or organizations are   kind of talking about is sustainability in this 
urban development and and real estate and and   all of this. How do you what are your thoughts 
on that? When we spoke about uh the propt tech,   I we actually I think I actually missed a point 
that the propt tech of course is made of a lot of   very cool company also in Europe that are working 
on the IoT and hardware part of this business and   um I met many great CEOs working in this space 
and I think that while this is not exactly a   focus of general real estate which is a more pure 
asset manager I think that everything is somehow   connected and synergy Okay, regarding the smart 
cities, um this is a big topic and I think that   we will eventually continue to talk about it uh 
somehow trying to solve it somehow and I think   that even in Europe and I speak in my opinion here 
it’s really differentiated also from the from the   positioning of the state that you are speaking 
about. So France has a certain kind of ideas.   Italy another one then Spain another one and then 
you go to northern country. I have a very good   friend who have recently visit um Stockholm for 
the first time and she saw a lot of improvements   also on a lot of physical part of the city. 
Again I will return on the point that cities   are eventually made of people and uh we need to 
understand why people are moving around a certain   geography or decide to come into a geography or 
eventually leave a geographies this has a lot to   do in my opinion and of course on the economic 
driving of the daria so if London I’m thinking   London Paris land so you know business center 
of interest somehow bras lot to do also with   the kind of experience and quality of life you 
are having into that that space in the research   that I’ve been doing in the space of academia I’m 
in because of the industry I’m working in we we   very often speak about the manatan effect so where 
you have this cities that becomes so expensive and   somehow so not built for citizens that the center 
is emptying emptying emptying and all the People   are then living in the suburbs and the center 
eventually becomes a big retail park for tourists   which is in fact right now Manatan if you think 
about so really central New York it’s eventually   hotels and retail and really no one lives there 
and and then with different degrees of course of   the different part of the city New York is a giant 
city but uh eventually if you think about when   uh a city becomes what they call a megalopoly so 
the really the influx of people of businesses of   interest of money is certaining this is 
an effect that it’s really physiological   so everywhere you are experiencing these things 
everywhere it’s not specific or any city London   is the same Paris is the same Milan is the 
same Madrid is the same and it’s really and   um which in a certain way in my opinion again is 
not really something if I would say the services   and the personalization of the experience of 
the city you have it’s around what you need   Here again when we’re speaking about people uh we 
are speaking about families we are speaking about   single people young people older people and it’s 
important right now to detect where these people   are moving within a 5year time Italy 5 year times 
10 years time where they are going where they’re   more likely to move within the same geography 
city so that you are really concentrating the   things that are of interest for these people 
in the places we’re leaving them again this is   really also my thoughts on the conversation I have 
also with the academia right now and I hope will   see this technology versus urban relationship at 
the moment where we can have so much information   to provide better services and serve better the 
communities of that specific part of the cities  cities at the end of the day are people right so 
absolutely so from your experience what are the   key factors in creating a team that can deliver 
such a product or can work on these ambitious   problems that we’re trying to solve.
So, when I was a lot younger than this than today,   I’m still young. I’m 34, but like I was younger 
and I heard a speech of uh I I don’t remember if   it was the chief of product or the CFO of Apple. 
I really now cannot remember exactly but the the   whole point was that even if they are very often 
very difficult people if you want to be successful   you need to to hire type A people not type B 
people because these people are really obsessed   about what they’re doing they’re striving for 
excellence are extremely ambitious when I had   the chance to meet this team from scratch I really 
got in my mind this idea of trying to find type A   people so people that eventually were of course 
good but also had the a great sense of ambition   uh in in really somehow being disruptive in being 
new in being really different somehow because I   think that in an industry like this or you bring 
this kind of energy or is very it’s very difficult   to change something my personal experience 
especially having worked in Europe um mainly   fortunately I didn’t have any experience in US in 
Europe you have many people working in technology   that are not necessarily very interested 
in becoming a CTO one day. They just want   to do their 9 to5 being very good slash enough 
individual contributor and not really interested   in all the fuss that comes with you know corporate 
life all the people all the influence you have to   do all the relations and so on. So eventually 
I think I got a lot of luck uh in finding uh my   head of data science and engineering who was in 
US was in US at the time and I don’t know when I   met this guy and he’s still with still in the team 
and is managing the data science and engineering   team I knew from the from the first second that 
he wanted my job and that’s what really was   convincing me that he was the right one and beside 
being extremely talented u and of course he’s like   he’s like one of the best u data scientist I I 
ever met. Uh I think that it’s important that   you have people that are entrepreneurial and feel 
these things as them. There’s it’s not just a job,   it’s something more. And then uh everything that 
came after somehow follow the same I would say   the same story line. So also the head of 
marketing site is very strong individual   young one very strong very strong domain expert 
extremely good international experience type   A person and also the younger engineers are all 
very like interesting people. I think that what   allowed us also to to recruit people even if uh 
I’m not Google I’m honest so when I go with my my   brand outside is not that easy to attract that 
quantity of extremely good talent but uh we’re   working on it but I think that what came out um 
and the rumors going around in the community in   Italy is that we’re like a startup so you have a 
34 years boss that is directly reporting to the   CEO the average age is 35 the younger one is is 
22 It’s really managed like a startup. So everyone   does everyone. I code as well. We we do everything 
like everyone does everything and there is a very   a very very very strong sense of ownership of the 
thing. Meaning that I have a rule that who does it   present it owns it. I don’t care the I don’t care 
the audience. It can be the CEO can be everyone.   The board itself. I I don’t care. You did it. You 
present it. You own it. And also we work. Thanks   God we were all on the same page but we work on 
a delivery based and I think that what position   the technology team that works somehow together 
is I’m not I’m not saying it’s so good and all   flowers and butterflies of course we have our 
own challenges but I think that overall our our   domain what we do it’s it’s a lot easier somehow 
to measure cuz like having a delivery based kind   of work where things are very measurable where 
people really understand what their expectation   are on them somehow it’s easier and I think that 
these things has also helped a lot. Many of us   became parents in these years. So we needed to 
manage also the parenting game the few first years   and having and being all together in doing these 
things has been really gamechanging on also on a   prof on a professional perspective. Right now I am 
running with a 100% retention rate from day one.   So no one ever think the the problem that you’re 
solving in itself is very fascinating. Right.   So because there’s one thing looking at solving a 
problem and everything and then there’s also this   element where once it’s solved you’re actually 
looking at decisions being made based on that   and the impact that it has in terms of you know 
however you measure excess right so an asset   manager if he picks up a right asset and he comes 
and says that hey your data was act I mean I was   about to pick A but then I picked B because your 
software told me to pick B it’s up like 10 times   versus I would have been down every time right
this kind of being a witness of your own results   it’s along everything I was sitting in a meeting 
the other day and um we also are responsible for   the data management data engineering of the 
company and even if it’s not rocket science we   are now building the first data lake for the real 
estate information both tabular and documental so   eventually I’m building a vector database and 
even if it’s not like rocket science for real   estate cuz like basically no one has it so if 
if I exclude the very big big player you also   will witness uh this evolution of this company 
with the people in so we have built a lot of a   great community of AI business translator within 
the function they really grew a lot are our first   s source of use cases are the people that that 
are bringing us the ideas where we work on for   thinking about new products and that’s been 
amazing but also with the leadership the other   day I was sitting in this meeting and eventually 
I was I was experiencing a discussion about a   system very senior leader said I really don’t 
care because we then attach it to the lake and   it’s going to be solved and I and I was like I 
was like wow I mean I I’ve done my job in this   case I really transferred the benefit because 
in the end I’m always saying that I’m like an   internal consultancy company you don’t really 
have to worry about how this things is done   how much is cost it’s it’s my job but you need to 
understand very well what’s your benefit how this   impact the choice you’re doing every day the 
fact that you don’t need to worry much about   your P1 or B if they are offering somehow 
the same the same model because in the end   you have a strong system that takes care of the 
data management of the information and this for   me is the end game. This is the satisfaction.
Absolutely. You mentioned uh when you started   and between today you become parents 
and everything. So I was looking at   your profile and then I noticed uh that 
you’re also a chief mom. Exactly. So  which is my first title.
I was wondering how did you become   one? I’m still not a chief dad. So how do you how 
do you juggle uh between both you know the job and   the the factor at home and how has that been?
Every time they ask me these questions I think   it’s probably a narrow divergence of mine but 
I’m not able to to lie. So um I would say that   it’s very very very difficult and then and I have 
I would say if I am the chief mom I have my vice   president which is my nanny ch um we need to 
be very clear that I am not doing this alone   she’s my vice president she’s very present and of 
course my my husband because of course he’s here   he exists and everything but but I I’m not saying 
no you know I can juggle everything I cook dinner   every night for No, this is not the case. When 
I’m alone with the kids on during the weekends,   uh we eat McDonald’s on Sunday. It’s it’s so good. 
No, but uh jokes apart, uh it’s very challenging.   I would never never never um give up. I would 
never give up my job. This is this really part   of who I am. As I said before, this passion, this 
interest I have, it would be impossible not to do   it. So um and I think that also especially for my 
daughter seeing her mom on the long term. right   now she’s six. She’s telling his friends, her 
friends that moms is teaching computer how to how   to build how to build and how to sell houses. This 
is the comprehension of my of my of my girl. But   I think it’s important for my kids to show female 
uh success story into this field especially as I   yeah especially as I have this great privilege 
of being a female leader in this field in this   very moment of time for our careers. meaning that 
I will have the privilege somehow if everything   goes goes well to reset the strategy of the eye 
for the future in my own company or in the company   I will work in because the game will be played 
in the next 5 to 10 years no and I think that   uh it’s important uh it’s very important 
for them also to to see this example and so   uh it’s is it’s hard and um having goal is not 
possible you will have less performance at work   and you will have times where you will be a little 
less good mom I actually came from um one part and   family my my dad was were very young and I always 
remembered my mom working but I always had in my   mind sometimes I got a accussions somehow that 
what she was doing and what she was doing it and   um and I’ve never actually asked myself too many 
questions and we have an amazing relationship   right now she’s an amazing grandmother so you 
know because these things pass also through your   sort of generational legacy I come from a female 
family somehow everyone always workked worked.  That’s awesome. That’s great. Kind of wrapping 
up things. Uh there two two questions, right?   So one looking into the future, where do you see 
this field evolving prop tech generally AI and   everything in the next 5 to 10 years? And two, 
what advice would you have for young engineers,   people specifically entering this area 
and you know looking to make a mark?   So first question I’m thinking that what we’re 
going to be experiencing with different level   of intensity deflation kind of landscape of the 
future we going to have cost of production or   cost of services that will go down thanks to 
AI thanks to automation and what right now   represent a competitive advantage will became the 
norm. This will will happen again with different   intensive and timelines also because new players 
will join the game. Players that have been built   AI native or or AI first as other people call it 
and they will somehow disrupt part of the market   eroding some revenues to bigger groups and didn’t 
or who haven’t start yet will be in great danger   of somehow collapsing. So I’m always suggesting at 
the moment to start now to reap the benefit of a   current competitive advantage and then be ready 
to the possibly cost game that will be the next   10 years most likely. Absolutely the next seven 
to 10 years. It’s it’s something you need to   do. And so this is and with regards to the real 
estate my think my opinion is that real estate   possibly will be impacted less and with less 
urgency. That’s because in the flowcharts that   describe these processes, no the ambiguous and 
ambiguous processes, a lot of the core operative   model relies in capabilities. Right now the models 
are not able even not good enough. They are not   really able to do it that are relationship based
that are knowledge based that are intuition risk   takingaking wise. And so it’s uh I think that uh 
even if we’ll go exactly through the same process   I just mentioned probably the timeline is push 
ahead and probably the intensity of the effect   will be less than other asset class and other type 
of businesses and industry. So we need to start   anyway we need to have a very lean and automatized 
and precise and accurate back office. Let’s say   again the activities that can be done should be 
done by machines but then we concentrate on our   5% which still remain obsessively important for 
the success of this type of business. On the other   hand, what I will suggest to young engineers right 
now and I had a call with university students this   morning and I said exactly the same thing is that 
writing good code will sooner or later becoming a   commodity. Meaning that you yourself need to 
find your 5% what you’re good at. I make you   an example. I wasn’t a very good programmer. I 
still am not have the greatest coder ever really.   My code sometime is sloppy, not very elegant. 
Many people will not like it. I have my one of   my engineers reviewing my GitHub and giving 
me grades yesterday took a 7.5 out of 10. So   I was very happy. But I wasn’t really, you know, 
I wasn’t really like the the most sophisticated   quarter. But for example, in my opinion, I’m 
very good in imagining things. I’m very good   in finding products, and finding ideas and then 
scratching the first models. And then the models   can be refined, more sophisticated. But I’m I’m 
I’m actually very good in seeing these things. So   somehow my suggestion to find your 5% your 5% 
may be an imagining architectures for example   for engineering in creating frameworks in having 
methodologies that helps deployment in a way that   it’s yours uh that is special that is proven 
that you have tested because in our job there   are many ways to do things but not all of them 
works. So I think buying the niche you like in   the industry you like and also in the niche of 
our job you like and become a master in it that   is not only into the coding aspects but along all 
the value chain of the product and the solutions   we are building. Absolutely. I think it’s now more 
than important to be a well-rounded person rather   than a very good programmer or coder because that 
that perspective that input can help you do this.  That’s absolutely spot on. Uh, I love the way you 
put it. So, find your 5%. Absolutely. All right.   Uh, so Cassanza, thank you for taking out the time 
and being on the podcast. Your the conversation,   I really enjoyed it. I think it’s very the 
way you own this. I I love that feeling.   Right. So, that’s that’s
Thank you very much.  That’s what I enjoyed the most.
Thank you.  And uh good luck with everything. Good luck 
with City Forward. Good luck with your kids.  Thank you.
And cheers.