Opening
Welcome to the Major Project Podcast. I'm your host, Orion Matthews. In this pod, we are learning from the people that work on projects over one billion dollars, trying to tease out those lessons and learnings from these big endeavors. My guest today is Bruno Caldas. He is a phenomenal individual, an incredible reporter.
He's got an 18-plus year career. 16 of them are in mining, and he is at the intersection of projects, finance, systems, and analytics. So he's pairing up the owner side discipline of governing capital projects with hands-on project controls and backs those with this sort of business intelligence and AI. So one thing about Bruno that's interesting is that we actually, uh, I was introduced to him by someone from, uh, a conference in Houston, and I was like, "I need to meet some of the most incredibly talented people, uh, that are out there in the [00:01:00] data reporting space."
And so, uh, mutual contact friend said, "You gotta talk to Bruno. Bruno's amazing. The work he's doing at Rio Tinto is, is, is incredible." And yeah, so that's how I got to know Bruno, and I'm really excited to have you on. So Bruno, welcome to the pod.
Introducing Bruno Caldas
Thank you so much for this introduction. So yes, we've been talking since this, since we are being introducing, and, it's been a honor to be part of your podcast.
Maybe you can take us a little bit through your career and tell us how you got into project controls. Just a little bit about your background.
Yeah. So just a little bit of introduction before starting my career. Uh, so I was always a nerd guy i- in terms of computer.
So when I was, like, 14 years old, I worked in a cyber cafe. [00:02:00] So I, I built, like, 20 computers. So I started with hardware. Uh, the computers came in, only the piece to assemble. So I help it in, and then I start to work on the software piece. So I was working there, like, not really at work 'cause I was in school at the time.
I was working there Saturdays and Sundays, but it was a really nice, uh, experience. I was in contact with computers all the time, so the digital piece was always my, my passion. Um, I started my career ... So I have a bachelor in, mechatronics automation engineering. At the time, like this was, I conclude my bachelor in 2009.
There was no AI, but you kind of learn everything to, to, to robots. Like, the true piece of a mechatronics engineer [00:03:00] is mechanical engineering, electric and programming. So was, uh, start to have this first contact with programming. I know it's a background, but this is really start to see, oh, this is really nice.
I think this is what I want to, for my career Um, so after I conclude my bachelor, 2009, I went to a oil and gas refinery project. That's my first billion-dollar project. But I was just starting as an intern. Uh, I was responsible for the Primavera P6. At the time, we had a full room of people just taking care of the Primavera P6 that was, scheduled.
I believe it was 80,000 activities. Uh, and the ... was at the time, this was [00:04:00] 2009, it was a local, uh, Primavera P6, so we are able to access all, all the same Primavera but in a local, network. Completely different from today. And we are working, of course, the critical path, all the activities. And I went to the site, uh, to learn more about oil and gas, was really amazing experience.
So I started my career direct in, in, in project service, project controls And I decided to go all in on this journey. And the thing that really helped me since the beginning was the combination of my software skills with the engineering and the technical skills. Um, so I stayed there one year in the same company.
I s- I was actually a, a contractor. I was not, I was not [00:05:00] the, the oil and gas company itself. I was the PCN company at the time. And then they send me to a railroad project. Uh, it was actually a, a subway and a monorail. Was two projects at the same time. So we are building the schedule using MS Project. And I helped them with AutoCAD as well.
So I was hired as a training engineering. The training engineering program in Brazil, you, you stay like six months in one role, six months another role, six months another role. So the o- the goal is to learn all the roles. So I stay a little bit with procurement. I stay in the engineering side doing AutoCAD for the, for the train stations, for the monorail.
Uh, the contract also included the trains itself. So they built a factory [00:06:00] for 25 subways and 25 monorails. Each, each monorail has like six, uh, cars, same for the subway. It was a really good experience as well because I learned engineering procurement, uh, scheduling again but now with MS Project instead of Primavera P6.
Uh, we are also using at that time SolidWorks, which is the 3D drawing. So I did the full training of SolidWorks. I never really draw using SolidWorks but it was interesting to learn the 3D piece at the time. And as I mentioned to you, my, my strong was always- Uh, learn software pretty quick.
I was learning everything by myself. At the time there was no internet, so we are doing [00:07:00] PDF books. I was printing 400 pages and I was preparing the books all by myself and reading and doing the exercise in the computer. So it was a different time. It was still possible to learn by yourself without YouTube, but it was little bit harder.
Like you need to read, I don't know, 400 pages. Yeah. But it was a good experience. If you want me interrupt me for questions? If not, I can- Yeah. So- Yeah.
Major Projects
So that was Petrobras and then you went to Vale and started your mining career, is that right?
One guy that I met on the refinery, he went to Guinea in Africa for a iron ore huge project there.
And then he was looking for someone with Primavera P6 skills and also with English. My English at the time it was not the best, [00:08:00] but it was good enough. So, uh, 2011 I received a proposal. Um, so the proposal was really good in terms of money because when you work, on the FIFO, fly in, fly out, you have a like 100% bonus.
So it was like three times my s- my compensation at the time. So it was a no-brainer, uh, choose, choice. I, I, I didn't search the, the, the, the country in the internet 'cause I said, "I think if I'm gonna search, maybe I'm gonna see the, the other side of Africa, so let's just go and see what happens." This was 2011.
I received like my onboarding July 4th. In July 7th I was already on a plane to Guinea, Africa. So this was... Maybe I can share, uh, my screen.
Sure. [00:09:00]
Just to talk a little bit about this project. This is a $20 billion project. I feel I don't have the option to share. Or I have, yes. Okay. Screens. Yes.
Mm-hmm ...
so this is Guinea, northwest of Africa.
Mm-hmm.
Uh, the capital is Conakry, and the mine was actually here, one, two, three, four. So the Simandou project was running two companies at the time. Mm-hmm ... half was Vale and half was Rutindo, the company that I am today, so it's kind of coincidence.
So that covers a lot of miles then. Those mines aren't that close.
Yes,
yes ... a couple of hundred miles apart for each or, or more?
I think, yeah, it's around 150 kilometers.
Okay.
So we went direct here to the mines. Um, here, of course, there's no infrastructure, so the first thing that we had to do is building some schools, hospitals, roads, even a camp for the people that's gonna work.
So we need to develop the cities around because we also need the locals to be trained to work with us, to have a good salary, and I think that was the biggest thing that I saw. Like, the first month that I arrived in Africa, the, all the houses are not developed, and after almost two years there, we saw, like, the community growth.
They were building houses. Everyone was growing 'cause at the time, I think w- uh, 70% of the labor was local, so we are developing a whole country. And [00:11:00] this is here, we also have the full railroad until the port in Conakry. So we are actually almost in, I don't know, 50% of the country, we had some constructions, and this, the railroad also has tunnels, bridges, so it's very complex.
Uh- Did
you... So you built the railroad then too?
Yes.
As part of the project? Yes. Wow. So-
The railroad and the port ...
for a mining company to build a railroad, is that common,
It's, it's- ...
out of...
Yeah.
Yeah.
So when we talk about iron ore, we are talking about million tons per year. So the four, one, two, three, four combined is 120 million tons per year.
Mm.
So you need to build a railroad. So for iron ore, it's kind of, when you are always doing an iron ore project, you're always doing the logistic as well. And that's includes also the port. [00:12:00] Yes.
So you built the port- And- The railroad goes all the way through that country, basically.
Correct.
Yeah.
And the railroad, um, there's also the option to, to...
for people, the passengers. It's not just iron ore. So they are using the railroads for other things as well.
So this was a 20 d- billion-ish project. Like, what was the hard part of the project when you showed up? Like, how did you intersect this, and what do you think were the big challenges on getting a project like this one done?
My... I think the biggest challenge for me at the beginning was the malaria mosquito.
Mm.
So because here's malaria is so strong, like, if you have malaria, you can die in 48 hours. So we are in a country very hot, using long shirts all the time, [00:13:00] passing the repellent every two hours. We also are sleeping in what we call, uh, net, uh, just to protect against the, the, the malaria mosquito.
So the biggest, the first big challenge was the malaria. So if you have a, a headache, you need to do malaria test, because malaria can kill you in, in 48 hours. But we had it... If, if you identify malaria, you just take the correct medicine, you are fine. So the first big challenge for me was to understand this new disease.
Um, and this was the strongest level of malaria that can kill in 48 hours. In north of Brazil, there are a little bit malaria, but there is a weak malaria that it's, it's not a big problem. So here, this was the first big challenge. And now talking about the construction, the biggest challenge [00:14:00] was if you need an extra equipment, like a excavator, usually- Mm-hmm
takes three months to arrive on site. Yes. Oh,
wow.
So we had to do all the prep. The, the primavera was with three months to arrive a new excavator for me. So we had to do a different preparation. We had to start ordering everything. We had to start the procurement earlier- So is that
like everything is a long lead item basically, uh-
Yes, because we didn't have a railroad.
We didn't have a good, research. We didn't have EPCM companies close. Like we ha- um, we actually brought two EPCM companies from Brazil, the, the number two at the time, and they had other projects in Africa but very far like s- [00:15:00] like Mozambique. So it was very hard to bring equipments.
So we arrived early because we had to do the planning three months in advance Uh, and a good thing about MB, I was able to learn everything about, I don't know, about civil work since the beginning because we started with civil works, earthquake, uh, earth, earthworks. And then we start civil, mechanical, electrical.
And I was... As I was responsible for the primavera, I was, going to the site, uh, twice a week with some civil engineer, uh, with, uh, 65, 50 years old just to learn everything. It was a really good experience. And as we are in the middle of the jungle, there was nothing to, to socialize. There was no bars, restaurants.[00:16:00]
There was not even an option to spend money. Like, all the money that we are gaining, the only way to spend was, a station, a gas station that you can buy, um, Heineken, Coca-Cola. So I was... So we are living there. The s- still flying fly outs was 40 days in site, 10 days in Brazil. So the four days site we are spending $0- Wow
because, because there is no place to spend dollars.
So then just sort of fast-forwarding a little bit here. So you're in Vale for 10 years. You started in Guinea, then USA- Brazil ... uh, got, Yeah. And then USA, Brazil, and then Canada.
Yes.
And through that you were basically a project controls engineer, project control specialist, then FP&A analyst- Yeah
um, and then interim manager. What... How did [00:17:00] your... what job did you like the most there, and what kind of other projects were you working on? You know, maybe just give us- Yeah ... a little bit of a, a run through of Vale. Um...
Yes. So I started my career on the major projects. After this one I have the S11D.
The S11D, very similar project. Um- So we had the mine. The D is actually the body D of the iron ore mine, and we have a full railroad until, uh, São Luís, 1.5 kilometers port. Very similar. So here I stay, like, four years, fly in, fly out as well. But this was, like, three weeks in the headquarter in Rio de Janeiro and one week here.
Sometimes two weeks here and two weeks in Rio. And I was walking the full railroads, the 1.5 thousand kilometers. So every night [00:18:00] I was sleeping in a different city just to see the progress of the railroad. And these are something that we implemented here in Africa as well. So every Friday we are doing a drone picture from the same site, like
So we have the full railroad, a lot of pictures Friday. So we are able to update our Primavera just looking into these high, high-quality pictures. We are building reports from those pictures. So it was really nice that we implemented this. We actually started this in, in Africa, so I brought this to this project as well.
Was that your idea then, to use the
drone
footage? Uh,
actually it was my boss in Africa. At the time in Africa, the dr- one drone was, I don't know, $50,000. Was really expensive. And like four years later, we were able to pay, I don't know, $10,000. Was really cheap. Today, I don't know, you can buy for $500 maybe. Hmm.
So at that time it was, was [00:19:00] way expensive, and we did some training. We trained the local teams to learn how to, to pilot the drone just to take the pictures every Friday. We are also doing videos, uh, once we com- 'cause this has 50 bridges, 1.5 kilometers. So every time you conclude a new bridge, we do a drone video and then we show this video on the headquarter.
So it's a really nice experience. I like this part. Here I was doing the performance reporting. But my role here, I was the eye of the CPO. So as the CPO was in the headquarter, I was going here, seeing everything, doing my own reports. So he had his own reports to compare to the EPCN reports and, and make sure everything was correct.
I was also doing, Primavera, but on the macro level, not going [00:20:00] into the, the task was, was really nice experience as well. So here we are starting the Amazon forest. Okay. So there is a, a, a particularity here. So when we are start, we are joining the Amazon forest, the railroad needs to be elevated, 'cause you cannot mess...
you need the- the animals need their own space. So the railroad starts to be very elevated here. But it's a small piece that is on the Amazon forest.
So for a project- like, 15, 20 billion. At that scale of CapEx, like, how does that change how you plan, communicate, control, and then how do you deal with the executive pressure for a project that big?
So here we have a, a particularity that is possible to do in a project like in this size. So we have like 1,000 kilometers of railroad. [00:21:00] You can divide it in 50 kilometers scope. Sorry to say kilometers, because I'm not used to miles. 50 kilometer scope by EPCM company, and then the companies that are performing better, we, we give more kilometers, we give more contracts.
So because of that, we are able to be- have a really good performance on the railroad. Uh, and this was one of the benchmark projects that we, we have. Uh, one of the projects here, there is, we call the, a railroad from the S11D to the old railroad. This one was the benchmark project that I, that I worked with the, in my entire career.
So we delivered on budget, on time, and we actually got all the lesson learned from this project and start to apply in other projects. And of course, this EPCM company that
did- Tell me about- ... a really good job. All right. Tell me about lessons learned a little bit, because that's [00:22:00] something that sometimes gets skipped.
You said that- Mm-hmm ... you had lessons learned and you managed to apply them. How did you actually do that?
So because railroad projects are very similar It's not that hard to, to do benchmarking. Like you are doing 100 kilometers of railroad here. What were the, the, the best things that you applied that you can apply to the other, uh, sessions?
For instance, the management. Like at the time the dashboards was Excel, and we started doing some dashboards for the, for the railroad and for the port as well. So we are able to see the performance that was outstanding, and then send a team there. It's why this performance is outstanding. Ah, because we did this and this.
So, so let's bring this to the others. [00:23:00] So-
Data, BI & Reporting
And you were kind of one of the dashboard wizards if-
Yeah ...
they were lucky to have you. Um- There's so few people in the industry that have that skill to bring it all together. Would you say that was... You know, how did you do that for a project this big w- working in Excel?
Yeah, I have a funny history. So when I was in SLN in the port, which is in São Luís, I s- I met a guy that was doing a dashboard in Excel, and I saw that fascinating, fascinate. It's not in the picture, uh, São Luís. Let me see here. But São Luís port to SLND port So when I was in the port, so here, here is the port.
Uh, I saw a really nice dashboard. The Excel at the time was really heavy and slow. But I, I sit with him the whole afternoon. I asked him to explain. He sent me this Excel later. I was able to do reverse engineering on that Excel, and that's how I started to do dashboards [00:24:00] Um, so I started Excel, you know, it's really hard.
Like, if you start adding more data, start to get really slow. The computers at the time, it was not that fast like the ones that I have n- now. So I remember that the dashboard was really heavy and hard to share. But we are using some pictures from the Excel to build our monthly reports, and that's how we, we start to do dashboards.
And then-
Um, yeah ...
do you think that that made a difference for establishing trust with leadership? Or, you know, how did, how did the reporting fit in to the overall sort of project trust-building with executives?
I always like to be very visual. Uh, when I start to do the drawing thing, [00:25:00] I saw that people in, in the headquarter really wanted to see my weekly reports with the high-quality pictures on the side.
And then I start to grow this report, like the monthly version of the report. I start to do some dashboards, some charts. Uh, I was doing Excel and printing on the... At the time it, we had two monthly reports. Was one monthly report with 200 pages Word, PDF that will be shared with the whole team, and had one more executive version with three pages, just what the leadership needs to see.
So in, have only three page, you need to be more visual, adding some charts. Then you saw that people were enjoying, so I started to, to like. And I'm very visual as well. I, like, it's easier to see some dashboards instead of [00:26:00] reading 200 PDF pages. At the time we didn't had AI to summarize a 300-page PDF. It was all, yeah.
Don't know if everyone was reading that report at the time.
Portfolio Manager: FP&A & PMO
So how do you deal with bad data then? You're bringing in all this information and it's in Excel and you were generating the reports. Was it at this point you were QA-ing it yourself or, you know, how did you kind of sort through the signal to noise on all that information?
Yeah, I, I think to answer these questions maybe talk about my next position. So after almost two years in Africa, four years in north of Brazil, I was ready to go to a, a portfolio level. So I, I, I said, "I've already lived in Africa and north of Brazil. I would like to stay here on the [00:27:00] headquarter a little bit."
So I got a position on the FP&A, and I was responsible for the whole CapEx, uh, management for copper and nickel. That's wh- when I started to work with Vali Ca- Vale Canada, 'cause at the time, the copper and nickel, still today, headquarter is here in Toronto, like three blocks from the CN Tower that you see here.
Mm-hmm.
Uh, so when I started to work in the FP&A, the main source was SAP So I start to go really deep in SAP, and I was responsible to teach the things in the copper and nickel mines how to use SAP. And as soon as I was got more mature, my SAP knowledge, start to do trainings. So I trained around three hundred people for SAP.
At the time it was SAP [00:28:00] PS, product system, FM, funds management, MM, which is the procurement. And once I was got more mature of SAP, I thought, "Oh, I think it- it's missing some fields here. Can we customize SAP?" So I, I had st- start meetings with IT, how can improve SAP. So w- I was part of a group that was constantly improving, SAP customizing.
And we also did s- at the time the company did some merges and acquisition. Not, not merge, just acquisition. So we had to implement SAP on the, the new acquired companies. I was part of this as well. And then twenty seventeen came and Microsoft launched Power BI, and that's changed everything. So I was, was already with a good, a good data knowledge.
I was already doing some dashboards in [00:29:00] Excel, but I was a- already doing some PowerPoint presentations with charts. And then when Microsoft came with Power BI, and was already included in our Microsoft, Office three six five, so it was already free for the whole company. So, uh, I started to look some videos.
At the time it was, like, one professor in the whole YouTube. So it's good that, that I, uh, speak English because in the beginning it was just English. And then some, uh, training in Portuguese was starting as well. So I... Since the beginning, I was the SME for Power BI. I started to implement, uh, in this F&A team.
And then the, my director at the time, she was always with her iPad, her tablet, [00:30:00] and she was, "Oh, can you do some dashboards for me that I can bring to my meetings in my tablet?" So that was my first real sponsor. So I did some, dashboards that she, she can use in the meetings. She can access in her iPad and comment, "Oh, this is the numbers."
So this was, like, my, my first sponsor back in two thousand eighteen. Yeah.
You know, this is an interesting point. I'm, I'm curio- I'm gonna loop back to SAP for a second, but, um, and then I wanna ask you a question on the Power BI too. For SAP, people listening to this, you sort of casually mention, "Oh, and then we roll it out, 300 people training," you know, mergers.
Like, these are big endeavors that are- Yes ... very painful for a lot of orgs, and maybe somebody is listening to this going, "I'm about to do this SAP implementation. Maybe I should quit my job." Um, what [00:31:00] advice do you have, if you've done these successfully, for SAP, say, during an M&A deal where you're gonna roll it out?
Like, what made it work for you, if it did at all? Maybe, maybe your advice is not to use SAP
I think mostly they had a lot of manual reporting. I remember a team that were spending the first 15 days of the month building reports. So we were paying very expensive senior analyst specialists to instead of bringing sites from the data, just creating reports.
So this was one of the, the first problems that I saw. And with Power BI it was easy to automate the reporting, and then we, we have those brilliant minds, specialists that can go to the numbers and bring insights how we can improve our operation, or how we can [00:32:00] improve our project management, how we can- Use this data to improve the company, uh, the productivity of the projects and the operation as well.
So I show them the value of, of, what we can do with a standard SAP with all the data coming really good. Yeah. You know, the ex- the garbage in, garbage out. So let's treat our SAP well. Let's focus on SAP. Each team maybe have someone with all the knowledge to make sure all the data here is good. And this is very valuable for the headquarter.
If you are receiving a good data from all of your sites, from all of your projects, you can do amazing analysis in your headquarter. And I think that was the next step. So once we organize it, all our data, [00:33:00] uh... So SAP was implementing the company in 2015. So after 2016, all the data was reliable. So I build a huge historical data from 2016.
Uh, and then we are using the historical data for the budgeting process as well, like the five years rolling forecast. And we actually had the historical data to support us. And was not just billion-dollar projects, was the full, uh, portfolio of projects. That includes R&D, sustaining. So we are talking about 5,000 projects per year, CapEx of $10 to $50 billion per year when you combine 100% of the portfolio.
Well, and so I have a... So this one you were FP&A, so you're on the- Yeah ... finance side and, uh, you're reporting to the board. [00:34:00] But- Yeah ... one question I have for you, 'cause we talk to a lot of people that are on the project side, and project controls is different than finance, right? You have a CFO- Yeah ... and a PMO.
So you were on the CFO side, so but you've been on the project side. Mm-hmm. So how did seeing projects through the finance lens change how you thought about project controls?
So I was hired to be the CapEx side of the FP&A because of my project background. So before, like, like, on the finance side, you have a, a, a variance of, I don't know, a b- a million dollars in a project in January.
So if you, if you don't have a, a project background, you're gonna see the variance. Ah, you're gonna talk with the guy who's gonna explain to you, and you're just gonna put on the report. You're not gonna question- Mm-hmm ... because you don't understand projects enough to question. So it's, it ... Maybe it's gonna
The [00:35:00] difference is, uh, tax rate that you can explain 'cause you are finance. But if the difference is, oh, my EPCM contractor created a claim, a claim because of this and that, and that's the, the difference. So my first role was to question the big variance 'cause this is what the CFO wants to see. Like, why we are spending 10% more in February or why we have spent 20% less.
What's happening? So this was the, like, I, I was on the F&A side, but I was bringing a lot of my project background. And an- another thing completely different that I learned on the F&A was the funds management and also the effects rate
management. Because we had, at the time, projects were in Canada, in Brazil, Tunisia, in New Caledonia, so we have different [00:36:00] FX rates. Uh, you have the funds management SAP, which you load budget by project. And then you have these FX variants that you need to take care. You need to do the, reallocation between projects.
So if your project is not performing, another project's performing really well or maybe there is a change management or a claim that needs an extra $10 million, you need to find a donor. So this was a complete different level of how you see the projects. Because when I was on site, the focus was deliverable, deliver on budget and on time.
Mm-hmm.
And then we are on finance, and we are seeing the whole portfolio. You also need to make sure all the projects that are performing are receiving the budget. And if you have an FX rate issue, you need to fix. [00:37:00] So we are-- I was seeing the other side of the coin. Also really important, and I would say the, the company was very mature in the data, uh, management.
Like, after two years there, I thought they were really mature in their SAP process. Uh, I thought they were really good in observing the, the, the acquisition company. And I think this, this I learned a lot with them. And that's what's really helped me to get a position in Canada. So- Yeah. Do you have another question for Pinney?
Can I move-
Project Controls & Performance
I do. Uh, I'm gonna try and pin you down, on a, on a little battle here. So PMOs have been on the rise, I would say. Some folks argue that the pro- the PMO should report at a higher level. So sometimes PMOs report through maybe the CFO suite, maybe a PMO reports [00:38:00] directly, uh, and peers with the CFO.
Like, where do you think the, the best place for a PMO is within an organization? Is it peering with the CFO suite 'cause they're different but-
Mm-hmm ...
friends, or would it be one level down?
My experience with PMO, it's always reported to CPO or the CTO, 'cause we have some technical aspects.
Uh, when this big billion-dollar projects, the company needs to have a really mature project management system And that's why usually the PMO is with this, or the CPO or the CTO. Uh, all the new projects needs to follow the project management system, in the oil and gas and mine. We have, uh, the projects are by phase, [00:39:00] yeah.
We start with concept, pre-feasibility, feasibility, execution, which is the construction. Then we have the commissioning and the operation. So, uh, the PMO needs to support all those phase. Those phase are complete different. Like we start with exploration, same as oil and gas. Uh, and then after you see, oh, maybe we have some, some, uh, a good reserve here.
So let's do an engineering to see if makes sense. So we start with financial modeling, capital location. If the project has a good MPV and IRR to go to, to the next phase. So this is something that I learned a lot in the FP&A phase. But I think the, the CFO needs to be really involved.
I think that my next position kinda showed this a little bit. So when I was on the FP&A, and I was doing the budget cycle for, uh, Copper and Nickel in Canada. They sent me here [00:40:00] to Toronto to stay like, uh, one month here, helping them to do the, the full budget cycle for the next five years. And then when I arrived here, I start to do the bud-budget cycle with Power BI.
So I create some scenarios. We had a cap of budget, so I, I was able to create a capital location scenarios. What would be the best projects for the next five years, and what would be the next project for the next year, which would be the focus. And then during the meeting, I was with the CFO here in Toronto, the CFO of Copper Nickel.
Then I showed him the presentation that I did for the, the budget cycle. I opened the Power BI on the screen. At the time it was physical meetings, was not, uh, was in-person meeting. Then I start to show the scenarios, and then I come back to my presentation and [00:41:00] recommend the scenario for him based on data And then he, he said on the, on the middle of the meeting, he said, "Oh, we need this guy here."
And that's how I was hired for the, for the company here on the headquarter in Toronto. And this meeting changed, it completed my life because I got a, a work permit now to move to, to Canada. I was doing what I loved, which was, uh, PMO with Power BI, and I was implementing Power BI, Power BI on, on, on here in, in Canada.
And all the other things we are seeing, I say, "Oh, I would like this in my team as well. I would like this in my department as well." So it's kinda like w- went viral in the company. Everyone was seeing the dashboards, [00:42:00] uh, the presentation, and they were really interesting. So my position here in Toronto was to build a PMO for Corporate Liquor.
And a different type of PMO, like with everything on a dashboard, a one-stop shop thing so you can see the risk management of your projects. You can filter your projects and see, the cost, the schedule. So we are building the full one-stop shop, like a project control tower for the whole portfolio, and this was really nice.
That was my first experience as a manager, so I had a team around seven people. Uh, had some consulting as well that, that I was leading Um, and then I started to implement a, a investment committee, and the main [00:43:00] client for the investment committee was the CFO. So coming back to your question, I think the
At the time, the PMO was reporting to the CTO. The CTO in mining is not technology, is technical.
Mm-hmm.
And we had this, uh, investment committee every month with the CFO, and we had another investment committee on a quarterly basis with the global CFO from the headquarter in Brazil. Uh, and then I implemented this investment committee.
Investment committee we are bringing, the performance of the projects. We are talking a little bit about the historical data, the budget cycle, project approval. So it was a very successful investment committee and I was responsible. It was around [00:44:00] 24 meetings per year at two per month. Was a lot of work just to run this investment committee.
But I think that was the time when the whole company was seeing my work, uh, and was, was amazing to be recognized like everyone. They asked me to implement similar to other products. And, uh, the company at the time, the iron ore was always the top one. So everything that iron ore was doing, the other products was copying.
So for the first time, the iron ore was seeing what's, what the base metals copper, nickel product group was doing, and I was doing a benchmark project to teach them how to do the same. So this was one of my biggest accompl- accomplishment, accomplishments. And this was a whole combination 'cause I was using some slides from the Power BI in [00:45:00] the investment committee.
The investment committee was way based on data. Of course, we had some other topics, but I usually start bringing a lot of data and graphics and charts. Uh, that time I started to study storytelling, so how I can build storytelling in my presentation. And this was one, uh, was really good. That's how I got my, my next job as well.
So, what was, would you say, you set up a PMO from scratch. That's no small feat. Yes. Uh, let's say someone's listening now about to set up a PMO for a multi-billion dollar org, um, what was the hardest thing about setting one up from scratch, and what's a tip that you would have for the audience on how to do it?
So because I, I have a big system background, I started with the [00:46:00] system piece. So as I explained to you, the SAP was very mature at the time. So the first piece of the PMO was to do the cash and cost performance, and that's includes, as I explained to you, the historical data. So we have like 10 years, uh, on the past plus 10 years in the future.
So it was like a 20 years. My dashboards you can s- you are able to see 20 years of data. It was really good. That's why you have a really strong data foundation. It's really important. It's really hard nowadays to have 20 years of data available, but it's, it's really good. So we s- I started with my... I would say where I was more in my comfort zone, which was the cost and cash.
And then we started to go piece by piece. The company already had a project management system in place. [00:47:00] So the first thing is we simplified because the, it's m- was more item R oriented. And as I explained to you the beginning of the podcast, item R is half logistic. So it's, it's railroad and port as well.
Mm-hmm. Copper and nickel you don't really have railroad and port. We are talking about, uh, kilotons of, of nickel and copper. We are not talking about million tons- I don't know, it's like four hundred million tons per year. The whole three fifty. And like it's, it's complete different commodities. So the focus was the mine and the, the big difference were the smelters and refineries and the mills.
So we need to learn more about the processing plant and also the byproducts. When you extract copper, you have gold as byproduct. When you extract nickel, you have copper and, and cobalt as [00:48:00] byproduct. So you are generating multiple products, multiple commodities in a single mine. So it's little bit different.
You need to, to learn. So the project management system needs to be adapted for, for the, the product group. And this is, was also an opportunity to simplify. So we had a team doing that as well. So first got the project management system in place. Simplify. We did the
systems with Power BIs, SAPs. Then we were able to get, the physical data as well from the Primaveras and bring it to this Power BI. Uh, we had-- We got our risk management team to standardize the risk management process, and then we are able to also create a Power BI that can see all the risk and opportunities for the, for the, the portfolio [00:49:00] projects.
Uh, what else? I can't remember. Uh, the next step was to create, uh, prioritization projects for the, for the future projects. So the level... When you have underground mines, like the nickel mines are two point five deep. One mine here, two hours from Toronto in Sudbury. So there is also this complexity of underground mines as well.
Um, and this makes a lot of sustaining projects. So the sustaining projects and the growth budget was pretty much the same. I remember one billion dollars per year to be on this. In growth projects, you have like five to ten projects.
In sustaining, you have two thousand. R&D, another two hundred. So we are able to create a process to capture new projects to, uh, they need to answer some questions and then create a prioritization process. [00:50:00] This was one of the big things of the PMO, the prioritization process, 'cause we are the benchmarking the company at the time, and we implement it on the other products as well.
And do you believe... I mean, it sounds like you did it by the book. You started with data as your foundational piece before really taking the practice and governance model forward. Mm-hmm. And with proper data and good storytelling, that sort of gave you the cover to then go in and build out the rest of the function sort of one at a time.
Is that kind of what I'm hearing in terms of your- Yeah ... the approach?
Because the data, the SAP was very mature, we are able to show incredible dashboards to our sponsors. So we are able to say, "Oh, look, this PMO has a lot of potentials. Can I have more people to help me?" So I [00:51:00] kinda used what was mature to build something that...
to get more sponsors, 'cause to build a PMO, you need sponsors. Otherwise, you have a very small team and you focus in a very small scope. So to be bigger, you need to, to show some work. So my recommendation is start with your comfort zone. Build something really nice that people will like, will see and say, "I would like this in my department as well.
Can, can you help? Can you support with that?" So my recommendation is start with, with what is more mature, and then you go to the piece that needs more work. Yeah.
Mining Industry
So most mining orgs, I would say, I'll probably upset some people, and tell me if I'm wrong, but I think they're generally laggards with technology.
Um, this sounds like it might be a bit of a rarer case. Um, why is that? Why do you think mining is behind on the [00:52:00] technology adoption curve compared to, say, like healthcare or other major industries?
I would say for operation, they are very mature. They are very ahead. Because when you increase the technology of the operation, you can save million of dollars.
Just give an example. The iron ore, let's, let's round it to 400 million tons. So if you reduce $1 on the whole process, on the whole cost, you are increasing $400 million on the EBITDA.
Wow.
So we have, in iron ore, a lot of
technology products, autonomous AI. I would say that the operation piece of the mining components are very mature. They are using AI since 2023. They are having an AI team trying to find solutions. There are some [00:53:00] mines that use autonomous, uh, trucks. Uh, and those autonomous trucks, you can gather a lot of data from the sensors.
And with this, you can use an AI to give insight and increase productivity, decrease the diesel if the truck is diesel, if the truck... But now we have some electric trucks as well. I think the future of the mining will be electric. Another one, I would say the Brooklyns project that I've been. I would like to share my screen again.
Mm-hmm. '
Cause I think this is amazing. So that's 11D. It opens- I think I need to share my whole screen 'cause there's multiple
Green chill. Okay. So this project[00:54:00]
Is truckless. Okay? What truckless means? Means every time the mine, uh, develop, the conveyor belts walk together.
Wow.
And this is 100 million tons per year capacity, completely truckless. So no diesel, no CO2. Um, and this is amazing, like you can see the conveyor belts. Here on this picture you see the conveyor belts, going to the processing plant.
And this was the project that I worked on North from Brazil Uh, so look, this is the conveyor belts coming from the mine to the plant
Wow.
So this is an example of [00:55:00] how technology influence the operation, and I think the operation is maybe five years ahead of the corporate, let's call us corporate. 'Cause they have the dollar return very quick if they invest in technology on the operations.
Mm-hmm. That's the big difference. From the project side, as we use a lot of EPCM companies, we kinda rely of their technology. That's one of the, major aspect of your question. The second aspect is the security, like the company need to be 100% sure, like if you implement a new AI, new... I don't know why the big companies doesn't have cloud yet.
'Cause they have a six, one month year process just to make sure it's safe, data will not leak. So I would say [00:56:00] the EPCM companies is one thing. The second is the security, the data. Um, but I think the companies are maybe one year behind because of the steps, but I think it's catching up. But when I talk with my colleagues from the operation side, they had an AI team with 10 people since 2023.
Yeah. So it's completely different from corporate. Wow. Yes. 'Cause they are more focused- But how do you- Yeah ... on, on increase the operation.
AI Applied to Project Environments
So basically, the areas where there's the highest ROI investing technology, that's where you're seeing it. But where do you see sort of the business intelligence layer going in the mining sector?
You know, we have, like you said, Power BI is an amazing tool, and now AI is kind of entering into the space. What are you seeing over the next few years? Is there a transformational wave coming or business as usual [00:57:00] maybe, just optimizing?
Yeah, I think the AI is changing completely the way you work. We, I, I am doing AI immersions every month, and sometimes what you are studying January, it's outdated in March.
Like things are changing so quick that you need to keep doing, keep studying. So now it's part of my routine to study AI. I would say I'm trying to be an, a subject matter expert for AI, my company. Uh, I would say I'm putting a lot of hours this, and I'm very confident- I'm organizing some AI training sessions with the team, so I, I will be teaching them the fundamentals.
'Cause today we see a lot of people doesn't understand how a token works, uh, [00:58:00] doesn't understand the difference of ChatGPT, SOL, Terra, and, Luna. Same for the three versions of Cloud, Fable, Summit, and Op- Op- Opus. Opus, yeah. Sorry my pronunciation. Sorry my pronunciation 'cause I, I was pronouncing in Portuguese.
Uh, so I've been studying not just the... I also built a full team of agents. Uh, remember that we were talking about the, the project management systems. Mm-hmm. So what if you have an agent for risk, another agent for, uh, social and communities, another for health and safety? So this is something that I've been building at this moment.
Not fully implemented. It's more about me training, and deciding the next steps. The companies are creating a lot of AI positions. If you filter on LinkedIn you see that this is [00:59:00] completely changing the market, I think. And the good thing is my background in Power BI is really useful right now because I have the data background.
I learned, uh, metadata architecture, which is really useful. Uh, and I know all the data foundation. Like y- if you don't understand, if you are really good in AI and you are not good with data foundation, doesn't mean anything. You will not deliver anything.
Yeah, the CEO of Microsoft, I think last year, said that your AI strategy and your data strategy are the same strategy.
Like, they can't really be separated, and I, I believe that. I think some people, if you throw AI- Yeah ... at bad data, it will give you beautiful reporting. It's almost worse because bad data with maybe Power BI- Yeah ... or Excel, it kind of looks bad. Mm-hmm. You can tell it's wrong, whereas AI will [01:00:00] smooth it out and make it look like it's correct.
Yeah.
It's even more important to have your data fundamentals, dialed in, I think.
That's what I'm gonna try to do the next months, to build a data fundamentals training for the teams here so I can make them to think more about to treat all of their tools the way they need to be treated. You cannot think that the tool is a problem for you.
You need to input the, the best data that you have using, following the governance standard, and this is gonna be key. And this is what, what we learned with Power BI, right? If the data was terrible, you see all the graphs and charts will not make sense. So this is something that was part of our routine since 2017 to look at all the reports.
So this number doesn't make sense. Let's [01:01:00] come back to the data. So you are learning and learning how to improve the data foundation. And as I had good, uh, mentors, especially on the SAP side. I had an amazing mentor from Brazil that taught everything about SAP, and then I had to replicate this to copper nickel and business, so I was teaching then data foundation at the time.
This is gonna be key for this AI revolution, in my opinion.
Yeah. So you, you do, you teach about BI at Northwestern University in their MBA program, and now you're talking about rolling out maybe a corporate training plan around data. What's your strategy there? Because my experience in the project space is that maybe data, the data science, the level of data sophistication varies a fair bit.
Some people don't have [01:02:00] any fundamentals, and other people might be a little bit further ahead. But maybe the average is n- not nearly where you're at, I would say. So how, what, what's your approach there, and how, how much do people need to level up their understanding of data, do you think, um, to kind of meet this new AI world?
Uh, what I'm doing right now, first focus on how to use AI token, how the AI works, how you train an AI, the tokens. As my company, we have an internal AI, the data doesn't leak. But you need to understand the limits. What kind of data we need to, uh, anonymate to use the AI. So it start with fundamentals, explaining.
Let's say, let's use ChatGPT as an example. So you have the ChatGPT Sol, which is [01:03:00] similar Fable, which is, let's say, a Ferrari. And you see a lot of people using a Ferrari to go to the groceries. No, you are using the wrong AI. You can use a Sonnet or a ChatGPT Luna just to go to the groceries.
You don't need to use a Ferrari. So it starts with the basic stuff, then people start to see you as an AI SME in the company, and they start to do more questions, get more involved. There are people that are really afraid of AI. I see people think when I open the, the ChatGPT Codex, I use the VS code. They
say, "I, I don't wanna learn any coding. I, I just want to go to this level of AI knowledge." So I try to, to teach them just until the [01:04:00] MDs. I think the MDs, the markdown are very important today. So I explain them the fundamentals, the chatbot, the different of web, desktop, then go to cowork Codex. And then I teach them how to build any skill to create an MDs, because I'm gonna need their MDs to build my agents, to build by...
So I'm usually teaching everyone, I would say level one to level five. Yes. And the level six skills, level seven, build a team of agents. Level eight, I don't know, cowork Codex. Say the l- las- last, last level for AI for me is the VS code when we start to getting skills from GitHub. So this part I keep to myself,
because they don't wanna learn. They are happy with what I'm, I'm teaching them. [01:05:00] But I think this is good. Uh, this is where I want for my career. I want to support everyone with the, until the MD, the markdowns level, and the rest I would like to do by myself. Yeah.
Like, what is an AI use case that is genuinely saving you time today?
So today to build a new Power BI, I usually ... If I am on my personal computer, I connect with cloud. I just do it for my personal Power BIs and I don't ... I do a full analysis. Uh, I talk with the cloud, like dimensions, the factors, uh, how I can optimize the dashboard, what I've been missing.
I also ask him to suggest more [01:06:00] charts and analysis that I've been missing. Today we can add HTML charts in Power BI, so we can ask the AI to support with that. Of course, I do a brainstorming a lot with AI. So I, I have someone to discuss with, like we are doing here. I do my own podcast with my AI just to discuss about the project and see what I'm missing, uh, what I think about my storytelling.
So those kind of things. The storytelling for dashboards products are easy. We start on the portfolio level until we go to the final project level, discipline level. Uh,
so I use a lot for Power BIs. AI is also very helpful for my presentations. Uh, on the corporate side there are little bit of restrictions. [01:07:00] So we need to learn. Sometimes I need to create an ex- an Excel database with anonymous data. Instead of site, it's gonna be site 01. Instead of project name two, it's gonna be project 002.
So this is useful as well. I use every day for brainstorming. I discuss with AI before starting a new task. Yes.
Future of PMO & Project Services
So then, you know, that's a really great example. It sounds like you're using agents to sort of give you advice on how well your end product is going to match. Mm-hmm. Um, so you're kind of using it as a coach or a sounding board, and then also to just build some of these tools-
Mm-hmm
So this kind of ties into a question I have around sort of if you've got a PMO that has AI [01:08:00] sort of increasingly running itself, like, what are people gonna be doing in five years? In this project space, do you think, in the PMOs?
I think for the project control side we will have a kinda new way of managing projects compared to the old way.
Uh, and the new way you have the integrated, intelligent and predictive. Today we have a lot of manual, disconnected and reactive. Uh, today we have a problem that data is in silos. Like you, you don't really know the information of a project South America, another project in Australia.
You cannot do any benchmarking. You cannot say, "Can you help me?" Let's [01:09:00] get it back to an example of the, the, the railroad. So now if I had everything loaded in AI, I can talk to the AI and do quicker. So which kilometer of my railroad, which miles are my benchmark and why they are different? What I can implement it from this benchmark from the others?
So I think the AI's gonna change completely the way we, we structure our projects. We will have AI agents to support you.
Uh, and then once you load the historical data, maybe you load competitor data as well, you'll be able to do internal benchmarks. I think for project estimate would be a big thing. Imagine you do your project estimating, get the data from all your lesson learned of your projects-
Mm-hmm ...
so this is gonna [01:10:00] change completely how we estimate projects 'cause I would say 50% of our projects that doesn't go well, the schedule on the cost, it's the project estimating can be a big factor
Uh, another thing
Is the how we can integrate with Power BI. To be able to, through a medallion architecture, get raw data from different products using the mining as example from different iron ore, copper, uh, different regions, South America, North America, Australia.
And then it will be able to have, uh, maybe the silver layer for our agents and the gold layer for your Power BI dashboards. And if you had a Power BI f- the whole portfolio in all disciplines, you have in a one-stop shop, and you also have the AI chat where you can discuss, you can create [01:11:00] some benchmark templates for future projects.
You can search by similarity. For instance, as I explained to you, the railroad is very similar, but in the mining, the projects are very different. So maybe they ... You have an AI agent to identify the similars, and then with the similars you can find the benchmarks.
It seems like what I'm hearing is that over the last, you know, since Power BI came out, there was a push to combine your information,
those that did it kind of y- have the story that you have, where you're like, "Look at this amazing stuff we've built." And then AI is now this new forcing function where it's, it's no longer sort of like optional cherry on top of a good organization is really well-integrated data. Now it is sort of that's moved into a state [01:12:00] of, critical.
Like that's go- that is the future. So data needs to get sort of combined properly, connected- Mm-hmm ... and then AI is going to be layered into that, which is kind of what I'm hearing you say. And sort of along those lines then, do you think that the PMO teams are gonna get smaller and just be really m- much sharper because AI and these tools are allowing them to be analysts, or do you think that the, the task of integrating this data is gonna make teams bigger and that they'll be more strategic in the future?
Like looking at where the PMO sits five years and what the staffing kind of strategy will look like, what would you say?
I would say the first step will need a lot of people to make sure, because companies have multiple data source. SAP, Carah, Procore, Confluent, [01:13:00] Primavera. So make sure that the data foundation is there.
It's gonna need a lot of people. Then build this AI, those medallion architecture. Understand the difference between regions, products, group. So you also need a, a big team for that. And once you put all the AI agents in place, uh, and maybe break the Power BI dashboard central tower, of course, you're gonna need people to make sure this is working, and you're gonna need more business intelligence people to add more dashboards as soon as they see the That makes sense, a new one.
I think it will, will be, will decrease in the long term. But I think the short term and the midterm [01:14:00] will probably keep the same or increase until the companies have this AI maturity in place.
Yeah, I tend to agree with you. I think that, like, maybe, uh, I would be a, a little more aggressive thinking there's probably gonna be the need for a fairly substantial CapEx investment in, for some orgs to get their data unified- Mm-hmm
in the PMO space, and then that might be over the next five years, and, and the implementation of AI, and then the sort of ROI will probably come, well, fairly- Mm-hmm ... quickly from that. But I would suspect the PMOs will grow or should grow. Like, if I was running a PMO, I would be trying to hire more people to sort out this data problem.
I would try and not put it on my schedulers and my costies as, like, the next to-do that they have to do. I don't know if that would work. Yes. What do you think? [01:15:00]
I agree with you. I believe in the short term, the companies need to prioritize this AI revolution. They need to achieve this new way of managing projects.
I think the companies has, can save million dollars in, when they put this in place. So in the short term, uh, I, I would be more aggressive and hire more people and make sure you have a strong data foundation. You build your AI meta architecture, AI refinery, your AI agents, then your control tower with the reporting and dashboards, and then you scale this to the whole company, all regions.
So I would say the company should, uh, increase the PMO team to put this in place, and maybe once [01:16:00] everything is in place, maybe I would say one to two years for imple- to implement that. Maybe they can start thinking about reducing the, the team. But at the same time there will be a risk if you reduce the team.
Maybe you're not gonna keep the data foundation the same level.
So this brings up an interesting point I wanna talk about with you, um, 'cause you've said this multiple times. When you're building out sort of a success story, you talk about starting small, proving it out, and then scaling it to the whole company.
I have seen large data warehouse, data integration projects maybe from big five consultancies where they go in maybe at the CEO level, CFO level and they say, "Throw 20, 40 million at us and we're gonna do it for the whole company." Sort of a top-down approach, right? So they sort of try to approach this data integration problem by looking at everything and [01:17:00] then solving it all, and then kinda pushing it down.
You're talking about doing the PMO right starts kind of at this really fundamentally small level. You prove it out that way, have a win, and then that win starts to spread in the organization, so it's a bottom up. Yeah. So in my experience, the bottom up is more successful. It's-
Mm-hmm ...
but what would you say?
Is it top-down initiatives are fine if you can get the funding? Uh-
Uh, yeah. I usually starts with the framework, and if I have a good sponsor, the full CPO, CTO, I... With a big sponsor, I will build an MVP with a sample, like, 'cause the company's too big to do everything at the same time. So maybe in smaller company it's possible to, to do a, like a top-down approach, everything at the same time.
But in a company that are in all time zones, need to work with people [01:18:00] 4:00 AM, 3:00 AM who work with Australia, I would go small, do an MVP in maybe one business, one region, and then extend when it's proven. Even with the top-down approach, think it's too hard to, to do everything at the same time
Closing Reflections
Yeah. Well, so, you know, thinking about our conversation today, what do you think would be sort of a number one takeaway that somebody listening to this thinking about their PMOs would you would like them to think about?
Yeah. So the, the first thing, all the PMOs that I built, I had a data foundation team, some business intelligence people that understand data. Data, data was really important before AI. Now it's even more important. [01:19:00] So focus on these people. Don't have 100% of your data foundation on the IT team. You need to have someone with this background in your team.
There are a lot of people with project management background, project control that also have the, the data management piece. So build a, a strong data foundation. Starts with the system that are more mature, and then go system by system until you have everything. I think the ultimate goal for the company is have a one-stop shop for everything, for the dashboards, for the agents, uh, a chatbot that can help you, uh, new projects benchmark.
So the final goal, be the one-stop shop. And now with this AI, AI refinery, maybe it's would be easier to integrate all of your source. So we [01:20:00] have a big opportunity now. I think the, the companies need to explore this opportunity And be more aggressive as it may go.
Thank you for that. That's, uh, the data foundation I think is the number one thing that you have underscored.
Mm.
And it sort of future-proofs you for where AI is going, and it also sounds like-
Yeah ...
um, probably just makes you more operationally advantageous even in the here and now.
Yeah ...
bruno, I really appreciate you coming on and sharing the story of your career. It's an amazing arc, and some of the things that you've done, you're a very modest person, but they are incredible. And getting to understand how you did them, the approach, and your vision on the future I think was extremely valuable.
So thank you for coming on today, and all the time that you've given to [01:21:00] the
podcast. Thank you very much. It was an honor to be part of your podcast.