Latest / 10X AI with Julius Neil / Will AI Replace Project Managers? What Changes in the AI Era

Transcript

Will AI Replace Project Managers? What Changes in the AI Era · 10X AI with Julius Neil

• Julius Neil Buenconsejo: There are different universities and institutions in the world that you can apply to when it comes to AI. So why University of Oxford, not Cambridge, or MIT, or some other Ivy League schools in the US? MARIAM PETROSYAN: frankly saying, like you mentioned about the MIT. Julius Neil Buenconsejo: What's the biggest misconception about AI right now? MARIAM PETROSYAN: perception is this halo effect the strongest PM will not be the person who produces the most documents, but the person who asks Julius Neil Buenconsejo: Good day, everyone. I'm here again with another episode with Mariam Petrosyan. She's from Armenia. She has years of experience in project management. And I met her because she also applied at the University of Oxford diploma in AI for business. And we might be in the same cohort. So we will be in the 2022-2027 cohort batch of this program. So Mariam, can you introduce yourself to our viewers and listeners and tell them about your background and why you are not pursuing AI? MARIAM PETROSYAN: Yeah, thank you. Thank you, Julius. so my name is Mariam, as Julius mentioned, I'm based in Armenia, in the capital city of Armenia, Yerevan. I have the educational background in both economics and engineering, which has helped me look to business challenges from different perspectives. So I started my career in HSBC Armenia and spent almost fourteen years There. I moved through several roles from banking operations into technology and project delivery, and eventually I was leading the software delivery team in HSBC Armenia. and currently, it's about one year, I'm leading the operational risk management division in one of the Armenian leading banks. And my career has always been connected with change, learning and transformation. So I'm especially interested in how AI can help organizations work more effectively while still mean maintaining strong human judgment and responsibility. So the keywords are the judgment and the responsibility, which I believe that should remain with the humans, not with the AI. So Julius Neil Buenconsejo: Yeah, yeah. So you lead teams, lead companies for 15 years for transformation. So inside some of the most complex and also the regulated environments in banking. today, for example, if I'm a project manager, I can essentially have an analyst, strategist, writer, researcher, or problem solver sitting inside my laptop. What has changed from the project management perspective? Or what do you think will change? Because from project management 10 years ago, and now you have an AI. MARIAM PETROSYAN: Hmm. Yeah. So very, very interesting question because AI is moving the project manager away from being the person who just manually collects and reforms the information. It can prepare the first draft of a plan, summarize meetings, consolidate status updates, identify some inconsistencies. Suggests risk and tailor some information for different stakeholders. The real opportunity is not to remove the project manager, it is to give the project manager more time for judgment, relationships, negotiation, and decisions. So AI can accelerate the administrative layer, but the PM still has to understand the organization. Challenge the output, so the the most important part, and bring bring people together and remain accountable for delivery. So this is something that should be done by the project managers in 2026. So maybe for some 10 years ago or 15 years ago, the project managers mainly were doing mainly the administrative part. But now the project managers are accountable for delivery. So this is little bit different from what we have some years ago. And in other words, like the role becomes more strategic than it was. And the strongest PM will not be the person who produces the most documents, but the person who asks the best questions, recognizes weak signals and turns information i into coordinated action. like in a large banking system integration, application inventories, dependencies, migration waves, testing evidence, vendor actions, cut over plan all had to be connected. And AI could help consolidate those inputs and flag the gaps so while leaders still decide readiness and whether to Proceed. In operational risk, so where I'm currently working, AI can draft an incident summary or group recurring issues. The risk professional must still validate the severity, control effectiveness, regulatory implications, and the ownership. So those are the most things that yeah, for the project manager should we Julius Neil Buenconsejo: Okay. So you, yeah. MARIAM PETROSYAN: know, yeah. Julius Neil Buenconsejo: Yeah. So you mentioned about the operational risk. Currently you are doing that, right? So where do you think AI comes into the operational risk in terms of whatever, you know, you're in banking or regulatory environment that you are experiencing right now? Where do you think should AI first focus on? MARIAM PETROSYAN: Yeah. so the main risks you know by the having the AI in the operational risk and overall the role of the operational risk persons little bit different. The main risks are not only like hallucinations, we also need to think about pure or the outdated data. Like confidential information, hidden biases, weak traceability, inconsistent outputs, over reliance, like we have a lot of like unclear accountability and the risk that a polished answer appears more reliable than in it release. that last point is the like hollow effect. if an yes, AI tool Creates an excellent presentation or summary, people may assume its financial analysis or the risk assessment is equally strong. But it couldn't be the reality. So at organizational level, a successful customer service pilot may lead people to believe the same solution should be used in compliance, recruitment, operations, and risk, but each use case has different data controls and consequences. So Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: for the operational risk and in the era of the AI, it is very important to have the correct controls, have the like to assess the consequences that AI can have, mainly the places where it is wrong used wrongly. So Julius Neil Buenconsejo: Okay, so currently MARIAM PETROSYAN: In an operational like yeah, yeah. Julius Neil Buenconsejo: in your current role, for example, since MARIAM PETROSYAN: Mm-hmm. Julius Neil Buenconsejo: you wanted to integrate AI, so where do you think in your current role, whatever you're doing right now, what are the things MARIAM PETROSYAN: Mm. Julius Neil Buenconsejo: that you think should be integrated in AI? What are the implementation use cases? MARIAM PETROSYAN: Mm-hmm. so you know it's in real world examples that I would say. Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: in my current role, and considering that my current role combines like the operational risk and also this transformation and project management. So it is somehow the combination of the risk and the project management. Julius Neil Buenconsejo: Yes. MARIAM PETROSYAN: Yes. so it is very important you know to have the the prompts so that can help you to to assess the risks, to assess the situation, and it is I would say that the one of the challenges to have these correct prompts because it is not about just asking the question like what are the risks. for this process and AI says this, this, this, and you say, okay, those are the risks. So Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: it is not the case. So you should be it sh shouldn't be like the prompts should not be very generatic. You should have Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: ask very exact let's say questions and the you should exactly define for the prompt From which role you are you want the AI to present the situation. So if Julius Neil Buenconsejo: Okay. MARIAM PETROSYAN: it is for the yeah, so like act as a senior banking transformation, cut over and operational risk advisor. So the role will be very exactly defined. Then the context. So a regulated bank is like Preparing weekend migration and systems integration. The scope includes customer facing services, multiple applications, so the there should be very detailed context. What you want the AI to produce for you. Yeah, and and this Julius Neil Buenconsejo: Okay, okay. Yeah. MARIAM PETROSYAN: is I guess one of the challenges for for my current SR role to Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: have this prompt very specific. very let's say to have exact inputs and to have the output what I want to receive so that I will just have the final output review it and have like let's say the final decision whether I'm okay with the result what I am receiving or it should be somehow polished or somehow changed because sometimes I feel like when you are just asking a question, because like in i in this era like AI is just collecting all the information, they know like what is your name, what is your role and everything about yourself, but sometimes they forget about it. And you should Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: be very careful just to assess the answer that the prompt provides you. And Julius Neil Buenconsejo: Okay. MARIAM PETROSYAN: The other thing, so you should always have these updated prompts. So it is not the case that you have some prompt, it's okay, it's working. So you should every time go update it based on the situation, based on the project, based on the environment that it should work. Julius Neil Buenconsejo: Okay, okay. So currently in your role, are you using AI already or individually or your companies or the companies that you're working for already MARIAM PETROSYAN: yes, so Julius Neil Buenconsejo: integrating AI? Yeah. MARIAM PETROSYAN: yeah, so at at this moment we are very actively trying to integrate the AI. So it is really very I would say that very long journey for the company current transformation. So even I recall just w two, three days ago we were having a catch up with our CRO and like Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: is very like like this AI in initiatives and we are always like encouraging like you should start to use it and it was you know we were just even discussing like how we can use it in our real life considering that like in Armenia we have Armenian language so our official language is Armenian which is like AI knows Armenian for sure, so this is not something unknown for for it. But anyway, Julius Neil Buenconsejo: Mm-hmm. Mm-hmm. Yeah. MARIAM PETROSYAN: it needs some time to have collected all this information in Armenia, and it is more sometimes like we are using the English version than translating it into Armenian. But anyway, it it is something that the company already starts to think and Basically from the risk perspective. So because it is very important when you are starting this transformation. it is very important at the early stage to involve like all the parties, like the risks, compliance, legals, because this is not something just technical. We are not just buying the AI tool, installing and start using. So this is something. that need to be adapted and this is something that need to be very clearly understood by the management and have the exact plan how we are going to translate to transform the processes and overall the workflow of the bank from the current situation to the AI. So This is something that we start, but I I guess that it will be a very long journey for all of us. Julius Neil Buenconsejo: Yeah, so I understand. So when you said CRO, is this a chief revenue officer? the chief revenue officer? Risk officer. So it's very MARIAM PETROSYAN: Risk officer. Risk officer. Yes, risk officer. Julius Neil Buenconsejo: interesting that your CRO is already thinking about using AI in your organization when he knows there are a lot of risks when using AI, right? So he is a very forward-thinker person. MARIAM PETROSYAN: Yes, yes. Like yeah, yeah, yeah. He he's very yeah, yeah. He's very like modern guy and I guess he's very well aware about all these new tools and techniques and for him it is very important also to correctly integrate the AI into the company's processes. So this is one of the key things. Julius Neil Buenconsejo: Is it only the CRO? Is also the C levels already aligned that we wanted to implement AI? MARIAM PETROSYAN: So I would say that on on a sea level, Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: I I I didn't speak with all of them, so but overall what is my perception is that they are ready to adapt it, though Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: it's it depends on the like say function to function, because like for business level sea levels, like were mainly like working on the sales or that stuff so it is a little bit different but mainly for the ITC levels so this is I I guess the the first point that they have on their table so and the the thing that they are always thinking and trying to understand like how we can adapt it easily in the current organization structure because anyway it is it is a big big challenge and Julius Neil Buenconsejo: Okay. MARIAM PETROSYAN: yeah. Julius Neil Buenconsejo: Okay. So how is your, for example, in your department or specifically for some, some of the sections that you're saying, so there are sales, this is different. So in your, you know, in the risk department per se. MARIAM PETROSYAN: Mm-hmm. Julius Neil Buenconsejo: So how are you integrating AI now? Is it only individual, you know, individual, know, individual education or do you have a training program, you know, upscaling of the people using AI? MARIAM PETROSYAN: yes, I would say that yeah. Mm-hmm. Mm-hmm. So considering that we are just starting all this transformation, at this like some I would say months ago, it was just on an individual level. Everyone just have his or her personal chat DVD or copilot and using Julius Neil Buenconsejo: You MARIAM PETROSYAN: it just for daily activities. but for now what we are also want to do We also want to more integrate in our like the analysis parts. Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: and for not to have some agents because at this moment like well we are I would say that we are not fully ready to have just the agents and to do some work because for having the agents and develop correct agents, I wouldn't say that there is operation operational risk is the f correct place to start. but from our perspective, so we have some operational risk management tools which where we also want to integrate this AI that it will just analyze the incidents, analyze the issues, maybe it would find some connections between the issues which for for the human it will be very difficult just go and remember or somehow connect with some old information. So at this point we we are starting to implement it, but not on an agent level. It is more on a case by case level, so and on some specific processes like this operational risk management tool where it is possible to somehow integrate the AI and have some help from the AI where it is possible. Julius Neil Buenconsejo: Which model are you using now? Is it Copilot or ChatGPTR Cloud? MARIAM PETROSYAN: Yes, we are yes, we are we are using like officially we are using the co-pilot. but like personally for me like Chat GPT was one of the first tools which I was using, though during last month I also somehow transferred to the co copilot because like for for work I'm using the copilot, but for my personal use I still using the Chat GPT. Julius Neil Buenconsejo: Okay, okay. So now I have an overview of what you're trying to do in the company now. So is this the trigger point? That's why you applied for the University of Oxford diploma in AI for Business. MARIAM PETROSYAN: Yeah, yeah. You know this this is something like I was thinking about some I don't know some executive education and about I don't know one year I was looking for different programs and what I feel that like I have two master degrees from economics Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: and from engineering and I feel that I want just to spend this time to learn something new and in this era I feel that AI is some something that need to be in your everyday life so it it is not something that we can say okay I'm not using AI I'm against the AI so it it it can't work. So th Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: that was one of the reasons that I feel that like I don't know finance innovation management marketing so this is all something that I have learned for many years. I have work, I have experience on this and AI mainly for business. I I'm like very excited that there like I will learn something that really could combine my understanding of the AI and overall Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: the business processes and the activities and it will make me more you know think more structured and what is the important thing like the ethical part because for now what Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: we learn about the AI we are just okay like I don't know some articles, some use cases but overall I'm more for academical education and for me Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: it is important to have this academical education and to learn from the roots and learn from the correct sources what it is, how to use it and where to use it. Julius Neil Buenconsejo: Okay, okay, that's nice. So how is the, what is the, you know, the reaction of the CRO when you talk to him about, know, going to study this at the University of Oxford? MARIAM PETROSYAN: So you know, overall i i it was something like I guess like everyone like when learning or really want to because I have graduated like fifteen years ago, my last that's diploma Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: from from T T T years, like are you ready to start from the from the I don't know, from the university to have all these exams and all that stuff, but I feel that this is something that I want to do and overall the my management was like very supporting my decision and Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: yeah so for me it's and for them also it is very important to have this continuous education so th this is something that you can't never stop and think that okay I have completed one degree it's enough now let's go and work so what I was always saying that like the like learn learn and keep learning this is I feel that one of the key models for me so for example ten or fifteen years ago Being strong on some Excel or some traditional IT tools could make someone much more effective. So this is something or Julius Neil Buenconsejo: Okay. MARIAM PETROSYAN: you know Excel, you you know some tools. So it's good. But then digital transformation changed the skills we needed. And today we are in a AI era. In another 10 years it may be something different again, and my own path from banking operations to technology delivery. Transformation, governance, and operational risk leadership showed me that earlier experience is not lost. So it be it becomes the Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: foundation for the next skill. the best future PMs will know how to coordinate people and AI while keeping purpose and accountability clear. So y you always should learn. And now it is time for learning Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: AI. Julius Neil Buenconsejo: Yeah, so I'm just curious. There are different universities and institutions in the world that you can apply to when it comes to AI. So why University of Oxford, not Cambridge, not King's College, MARIAM PETROSYAN: Yeah. Julius Neil Buenconsejo: or MIT, or some other Ivy League schools in the US? MARIAM PETROSYAN: Yeah, frankly saying, like you mentioned about the MIT. So this Julius Neil Buenconsejo: Uh-huh. MARIAM PETROSYAN: is the this is one of the universities where I dreamed to study, I don't know, 15, 20 years ago when I was on my undergraduate degree. So I I just thinking that maybe one day I will go to studying in the MIT. but now when I was just looking for different programs and different universities, like I was looking some university in the US, in the China, in Singapore, so in in all over the world. So the Oxford is one of the like they providing one of the programs that can be combined my current life space. So I can work, I can study, I can travel between Yerevan and London. So like I calculate all those logistics things as well because like it is Julius Neil Buenconsejo: Ha ha. MARIAM PETROSYAN: very difficult to travel to US like for several times during the year. It is very long distance, you have to do some I don't know, there are no direct flights, and the same with I don't know with China and and that part of the world. But in the case of the Oxford, so they are providing the this course which combines only four modules so it is not something that you should just quit your job your your life in your country and move to the Oxford and just study there. So it was just very good combination of my current lifestyle and what I want to learn. And from the other side they also provide this AI for business course Because I was mainly looking something related to the AI. Like not all the universities have this type of the programs. So mainly for the executive education, it is like the MBA. So this is something that all the universities Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: c can offer. but like MBA, it is very long journey. and from the financial perspectives, from the logistics, it's it is more difficult to pr to have parallel in your work and in your life, yeah. So I I just find Oxford Julius Neil Buenconsejo: Okay. MARIAM PETROSYAN: that just matched with all the criteria that I was looking for. Julius Neil Buenconsejo: Okay, that's nice. That's nice. So, going back to the project, project management and prompt engineering. So for example, if you are going to give an AI assistant a problem, a project problem tomorrow, what would separate a mediocre answer and an answer that you could actually take into an executive meeting? MARIAM PETROSYAN: you know it's it's it's interesting question because like AI is very useful for translating the same project reality into language of different audiences. A delivery team may need detail, an executive needs the trend, impact, decision owner, and timing. AI can summarize very lar large volumes. Of information and identify recurring themes and draft a concise update. The project manager still has to frame the message. I would ask AI to separate facts from interpretation, state what changes Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: since the last report, identify the top like free risks, explain the business impact and make the required decision explicit. I would also provide a word limit and ask for. plain language rather than technical details because like for mainly for the executive communication you don't want them to just read some technical details and the PM must then validate Julius Neil Buenconsejo: Yeah, yeah. MARIAM PETROSYAN: the message so a concise report can still be misleading if it removes uncertainty and hides disagreement or turns an assumption into a fact so it it should be validated by the PM. before providing to the this message to the executive committees. In operational risk reporting, we have these incident trends, key risk and control indicators, and they can be converted into an executive narrative, but the accountable leader validates the severity escalation and the message to govern to these governance committees. so this is something that I guess we should always keep so use the AIs correctly and transform Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: and validate what we are giving as a message to the executives. Julius Neil Buenconsejo: Okay, okay. Do you think, example, in prompt engineering, so people are thinking that they need to be an AI engineer to be able to understand AI or to be able to use it effectively, but you're talking about it also specifically for a project management or project manager's perspective. So what does a good prompt engineering actually mean in the context of a project management? MARIAM PETROSYAN: you know for a project manager the the prompt engineering is structured briefing. We already know that a team cannot deliver well from a vogue request. AI is the same. The quality of the response depends on the quality of the brief. So a strong prompt explains the context. Objective, audience, available data, constraints, and desired output. It also states what the AI should not assume. If information is missing, the AI should flag the gap or risk question instead of inventing an answer for you. So good prompting is, you know, it is also iterative. The first response is always just draft. project manager. should challenge it. what did you assume? What evidence like support this? What could make the conclusion wrong? Which risks are missing? Then the PM adds context and improve the next version. So it it is the flow that I I believe should be kept to have the Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: right prompt and to not just have let's say wrong assumptions just because AI Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: said it. Julius Neil Buenconsejo: What do you think is the biggest mistake that most project managers are making when they prompt AI? MARIAM PETROSYAN: what I would say on this, so I I would just bring one example from from my own example when I was a project manager and it was the era of the agile project management. Like my agile experience is is a good parallel for this AI and PM transformation. When agile became popular, some companies just replaced the waterfall without analyzing the project type, culture or delivery needs. They applied it incorrectly, failed to deliver and later moved to hybrid approaches. The lesson was to fit the method to the context. And the same situation with the AI. So you should just not install the AI tool and say so starting from today we are using AI in everywhere. So the company's needs, the company structure, the overall you know the culture should be assessed and the AI should be and what I feel now. So many PMs just have a quick question to the chat GPD taking some milestones which are not correctly linked with the current situation of the project and they just put in some I don't know weekly updates no one is on a understanding what is going on for Julius Neil Buenconsejo: Ha ha ha ha MARIAM PETROSYAN: yeah and it is one of the biggest things so the PMs and the mistake that PMs are mainly doing in this transformation period so they are just taking what the AI gives as an as an let's say I don't know the items for the project without having the exact culture of the company, without analyzing the exact type of the project. So I I'm always paralleling it with this agile transformation which was I don't know ten years ago and everyone was doing agile. even in the banking so it was very interesting like the banking was all always working with this waterfall approach so we have the project we have big plannings and all that stuff and one day Julius Neil Buenconsejo: Mm-hmm. Mm-hmm. Mm-hmm. MARIAM PETROSYAN: everyone just we are working as an agile but in the regulated environment not every project can be delivered with agile methodology. So this is the thing that many PMs fail when they choose the wrong methodology. The same with the AI. So it is just you should select the correct tools for the for each situation, not just we are using Chat GPT. Okay, so everyone using the ChatGPT and and the PM and the developers, I don't know, and everyone and finally no one could take the ownership and the accountability. What do we receive? Julius Neil Buenconsejo: Mm-hmm. Mm-hmm. MARIAM PETROSYAN: So yeah, it it is one of the biggest mistakes. Julius Neil Buenconsejo: Okay. Okay. All right. Thank you. Thank you for that. Let's put a hypothetical scenario, a hypothetical situation so that our viewers and listeners will be able to understand. So for example, you are in a bank, you're a project manager, you're in a bank, you are implementing a new enterprise disk platform. So it will be in multiple countries, there are regulatory deadlines, legacy systems, different business stakeholders, data migration, and so on. MARIAM PETROSYAN: Mm-hmm. Julius Neil Buenconsejo: So if you're the PM, walking into this project on Monday morning. What's the first prompt you'd give AI? MARIAM PETROSYAN: so what was I i i you you know it's it's really difficult questions. So really difficult question. what I would say Julius Neil Buenconsejo: Yeah. MARIAM PETROSYAN: that like act as a senior PM for for banking transformation project. So the the the first thing that I would say to the prompt, so I should define my role, like how to act it. The other thing will be just the detailed content. What is it about? Like the project content in very details, like starting the banking working hours, like the bank will work Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: from Monday to Friday, from 9 p.m. 9 a.m. to 6 p.m. ending with the organizational structure what type of functions it have because it is very important to have just very correct content for the prompt to understand like what type of bank is it is it just a retail bank is it a corporate bank is it just a I don't know, it it is the central banks, so it it is yeah, it it it Julius Neil Buenconsejo: haha MARIAM PETROSYAN: it's very important just to have it. The other thing will be the objective on the prompt. So what is the objective for this? So help, for example, the steering committee decide whether the program is ready to proceed or identify the top 10 risks. maybe some critical dependencies or decisions that are needed. Then what we need to have it is this input. So use only the material I provide. Like I would provide some milestone plan that was already prepared, some application inventory that we have in the bank, some dependency map, all these migration waves and I don't know some rollback plan or result action, so all this should be provided as an input to the prompt. And then to have very specific constraints. so what constraints mean like do not invent evidence. So if we didn't provide you any evidence, don't invent it. Just yes, separate Julius Neil Buenconsejo: You MARIAM PETROSYAN: facts because sometimes you really feel that AI is just inventing something. the that not your use case and it is important to to be specific do not invent evidence treat ai output as analysis for human validation not as approval to proceeds so those are the constraints that should be defined when you are designing the prompt and finally the output so you should be very specific on this like provide I don't know, five bullet executive summary. A table with risk, cause, impact, early warning indicators, critical actions for the next 72 hours, questions the steering committee must resolve, or some red flags that could justify a no-go decision. So the prompt should be very, very specific. And all this context, objective, input, constraint and output should be very clearly defined on the prompt to be sure that the answer that you are getting, the thing that you you was requesting from the AI. It is not just something that was invented by him. Julius Neil Buenconsejo: Okay, okay, okay. Okay, so for example, if AI could sit in five different chairs around your project table, so you are the project manager, and then there are five AIs, what would those five AI roles be? MARIAM PETROSYAN: it's I would separate them like the PMO assistant, like more administrative Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: part. So for sure this is this is something Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: that even maybe I would dedicate two AIs for this because we Julius Neil Buenconsejo: Hahaha MARIAM PETROSYAN: have a lot of administrative part which which is the thing that can be 100% done by the AI. So this is something where we can just give them and say okay you are doing this meeting administrative part you are doing the documentation administrative part the other thing it could be the business analyst part I guess so this is something that good AI agent can do. So this is more on AI agent level, but this is something that I guess this analysis part can be also assigned to them. I would also assign more not senior but more junior developer and QA roles. let's say more junior. technical part to the AI. But anyway, I would Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: keep this seniority on the technical part to the human. But more technical and let's say administrative parts could be also provided to the AI. So what else? I guess at at this moment I wouldn't replace any other role because what I have from the other part I mainly these I don't know the final users of the product which I wouldn't for now replace by the AI because this is the place where I think that should be let's say checked the AI work and validate the product. So this is something that I wouldn't Julius Neil Buenconsejo: Mm. MARIAM PETROSYAN: give it to the AI and keep with the human. Julius Neil Buenconsejo: Okay, so among those five AI roles, which role would you trust AI with the most? And which role you will never fully trust with it? MARIAM PETROSYAN: the most I would rely on the administrative part and this Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: periodic, I don't know, meeting updates, some exchates, some just reminders to the stakeholders, all this part I would fully transfer to the AI. And the least I would say that this final acceptance. So I I I wouldn't trust it to the AI and I would keep this final acceptance of the product to the human. Julius Neil Buenconsejo: Okay, yeah, that's a very good answer. And then you also talk about the accountability. So for example, if AI gives you the wrong answer, who's accountable? The AI, the vendor, the project manager, or the executive who approve it? MARIAM PETROSYAN: Hmm. it it it's it's a very difficult question, you know. It I wouldn't say there Julius Neil Buenconsejo: Yeah MARIAM PETROSYAN: is just one correct answer. So it it depends on the situation. For sure, AI is not accountable. I would say that accountability cannot be on the AI. So it is just provide you the information and the accountable person. should be in this case mainly the project manager or depending on the situation so it could be some other stakeholders of the project but mainly I would say that the main accountable person for the project is just the project manager so the AI is not accountable Julius Neil Buenconsejo: Ha MARIAM PETROSYAN: for the for the wrong project flow for I don't know for Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: the mismanagement of the project. So this is something that should be kept with the project manager. Julius Neil Buenconsejo: Do you think learning from engineering actually makes someone better as a project manager? MARIAM PETROSYAN: I would say yes, you know, because it give you more time on like you delegate some administrative part to the AI and you have more time to spend on communication with the stakeholder to more on you know on analyzing or assessing the the situation because from my personal experience what we have some years ago when everything was on with paper and pen let's say so we are just took the notes on the paper during the meeting then go and start to think what they say what like this one said this one and trying to have some this meeting minutes. And this is something that now PM just can even think about it because you have the I don't know meeting on a Teams or on a Zoom. It took the Julius Neil Buenconsejo: Yes. MARIAM PETROSYAN: all the transcript from what was said, you can put it to the AI and will give you just the summary of the actions of all about the meeting. So this is something that make easy the PM life, and from the other hand. PM can spend more time to have this stakeholder management because I believe that the key role of the project manager is to manage the stakeholders, mainly in a big project. So in this case, the project managers have more time on the stakeholder management rather than on the administrative part. Julius Neil Buenconsejo: Yeah, so you mentioned about this is transcribe during meetings and then so many, so many information, MARIAM PETROSYAN: Yeah. Julius Neil Buenconsejo: you know, in papers before. So currently, so one of the biggest jobs of a project manager is translating complex complexity or complex, you know, complex problems, complex reports for MARIAM PETROSYAN: Mm-hmm. Julius Neil Buenconsejo: executives. So how can AI help with that? MARIAM PETROSYAN: So here AI also can help not as I have already mentioned, not just copy and pass the information, but just analyze by correct prompts. Like I don't know, you want top five risks on the project. So you provide a lot of information to AI, it analyzed and took like five point but those five points should should be accepted by the PM before transferring it to the executive. So this is something that I c I guess should be kept and this is something that AI make the project manager's life more easier. So just not reading all the day all these old meeting information, I don't know all actions, what did what, but it can just give all this information as an input and get the correct input. The only thing just to have the right prompt for input and the output. So it will but anyway the output should be confirmed let's say by PM before transferring and Julius Neil Buenconsejo: Mm-hmm. Mm-hmm. MARIAM PETROSYAN: even sometimes it needs to be challenged the the AI the the PM should challenge the output that the AI provides. So it it's it's it's important just to not forget about challenging the AI and not take as Julius Neil Buenconsejo: Yes. MARIAM PETROSYAN: a truth what AI provide you. Julius Neil Buenconsejo: Okay, okay. Just like you, I also work in a highly regulated environment. what happens when an organization, because this is more of the risk, you know, the risk side, when an organization adopt AI faster MARIAM PETROSYAN: Mm-hmm. Julius Neil Buenconsejo: than their governance frameworks can keep up? MARIAM PETROSYAN: you know i it's I I guess it is the current situation for many, many companies. the first thing, maybe in the regulated environment, it is the data leakage. So it when you don't have correct adoption and some day you may find out that your data is just leaked to I don't know, some un Trusted AI tool because when you are implementing the AI tool, first of all it should be from the trusted source. So the AI should be, I don't know, the like the from Google, from Microsoft, so it it should be from the trusted and known vendor. So it shouldn't be just someone wrote some AI tool and you are just adapting it in your regulatory environment. because it was very quick and it was very cheap one. So for regulated environment it's important to have first of all the trusted tool and then have not to be in a rush just to transfer everything to AI but to start with the things That can be checked and validated. Then you can just go Julius Neil Buenconsejo: Mm. MARIAM PETROSYAN: to the processes that are more let's say complex. but like for example, what I've previously mentioned about this administrative part. So this is something that could be under your control where you have can someone who is accountable for the output. And the AI can just take all this information, analyze and give you some bullet points for the validation. So it is very important to start from the processes that are very well where you have let's say full control of the output and then start to transform the processes which are more complex and require more dedicated time. Julius Neil Buenconsejo: Okay, so do you think AI governance is going to become as important as cybersecurity governance? MARIAM PETROSYAN: Yes, yes, I b I I believe to it. So you know it yeah I I think this is one of the things that make me to to consider this executive education on the AI. because I believe that what we have now mainly on the corporate governance level, it should be correctly Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: transformed to the governance with AI. So it's it is very challenging journey but I believe that even now very big corporations companies like the H S B C globally they one of the first companies who have the executive for special AI. So this Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: is something that they believe that they the Julius Neil Buenconsejo: Wow. So you have a chief AI officer. MARIAM PETROSYAN: Sorry. Julius Neil Buenconsejo: Chief AI Officer in HSBC. MARIAM PETROSYAN: Yes, yes. In in in HSBC globally, yes. I I guess it was some months ago. I was just Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: reading it on the news and it was really very good sign that these type of companies and I guess JP Morgan was and like very let's say big companies now pay special attention to the AI. And what is important that Now you know AI it is not the part of the technology, it is more part of the governance. So in the Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: early stages of the AI, it is more like it was some let's say subpart of the IT or transformation, but now it is more treated as a separate governance role and the in with the example of these big companies I would say that the trends goes to to this special governance treat to the AI. So the AI should be treated specially and the governance should be adapted respectively to the to the AI and the risks and challenges that AI bring with him. Julius Neil Buenconsejo: So I'm just curious about the current state of AI adoption in Armenia as a whole. Is the country already putting up all the frameworks, all the regulations with AI, or people are this... Yeah. MARIAM PETROSYAN: you know, I would say yeah, yeah, yeah. It's I would say that it is in in very early adoption now. So because there are a lot of so on governmental level they are starting to speak in about AI, which is very positive. So we like when we you start to speak about such a let's say big transformation in all over the world, it is a good sign. but from the other hand, I believe that there is a lot of things to be done on this. So in Armenia we have a lot of Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: IT companies, a lot of let's say engineers and a lot of resources who are working on developing this AI machine learning tools and all that stuff. But all these are more technical part. So we also need to have some regulation. But unfortunately, as I know at this moment, maybe I don't know, something something new which which I Julius Neil Buenconsejo: Hahaha MARIAM PETROSYAN: didn't manage to know about it became. But Julius Neil Buenconsejo: Yeah. MARIAM PETROSYAN: at this moment as I'm aware we don't have regulated let's say some laws or rules it is more on a company level but I'm Julius Neil Buenconsejo: Okay. MARIAM PETROSYAN: sure that considering this transformation because mainly in the banking in Armenia we have very big banking sector in Armenia and there are about thirteen, fourteen banks in Armenia, private banks. And they are all in a this transformation era. And I'm sure that in at some point there will be some regulations to somehow regulate these AI agents or AI AI adaptation to the companies. But at at this moment I'm even not sure in around like in other countries, do they have these type of the regulations regarding the AI or no? Julius Neil Buenconsejo: Okay. Okay. So I'll ask you some quick questions. It is a problem or our prompt or problem section. So for example, so AI gives you a confident answer, but no sources. Will you trust it or verify it? MARIAM PETROSYAN: so I will verify it. Yeah. Julius Neil Buenconsejo: Okay, okay. Your executive asks for one page summary in 10 minutes, AI or manual. MARIAM PETROSYAN: I would combine it. Yes. like AI then I will yes. I I would like you know, I I would I would just use the voice Julius Neil Buenconsejo: Okay, AI. And then you just do it, know, check it manually. Yeah. MARIAM PETROSYAN: just to have a quick yes, quick recording to the AI, then Julius Neil Buenconsejo: yeah, yeah, yeah. MARIAM PETROSYAN: then parallel I will start start to manually write something on a on a p on notebook here. Julius Neil Buenconsejo: Yeah. Okay. AI identifies a risk nobody on your team MARIAM PETROSYAN: Mm. Julius Neil Buenconsejo: noticed or identified. What happens next? MARIAM PETROSYAN: I will not believe to him. No. Julius Neil Buenconsejo: Okay, okay. Your AI-generated project plan looks perfect. What's the first thing you challenge? MARIAM PETROSYAN: The first thing I would challenge just go just go and check the input that I have provided. So maybe it missed something from there. Julius Neil Buenconsejo: Okay, what's more dangerous, a bad prom or blind trust in a good answer? MARIAM PETROSYAN: pa pa pa it's it's it's difficult to say. So the blind blind Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: usar I think. No, i it's a difficult question, yeah. Julius Neil Buenconsejo: Good, good, yeah. Okay, and then a junior PM says, AI can do my job now. What do you tell them? MARIAM PETROSYAN: I would tell I would tell that not truly. So AI can do the administrative part but you should start to challenge and to to verify what AI do for you. Julius Neil Buenconsejo: Okay, so before we go into our last part of the podcast, I'm going to ask you if there is a project manager who is watching right now or listening. So I understand that AI is important, but I don't know where to start, right? What actually, you know. So if there's a project manager listening to this episode right now, what would you tell that project manager he does or he starts doing now in the age of AI? MARIAM PETROSYAN: So, what I would say, and what will be my advice, just start to read. in based on my experience, like how I start and now I'm using this type of the information, like for the project managers, there is a project management institute who provide this certificate for the project managers, and they have a lot of information about the project manager's role and the AI. So my current let's say trainings that I used to just go on to learn something new I mainly use this type of the trainings from this project management institute who are let's say collecting the Best from everywhere and make it special for their project manager role. And this is something that someone who has the project manager experience, even without experience, but wants to start the project management in the era of the AI, I would suggest to start with the sources that provide the project management institute. So this is trusted source, this is something that easily let's say understood so it is not a technical thing. So this is something specially designed for the project managers. Julius Neil Buenconsejo: Okay, my last question before the rapid 10x. Will the competitive advantage still be knowing how to manage projects or will it be knowing how to manage AI-powered teams? MARIAM PETROSYAN: I would say that y you should be ready to have both of them. Because now like to have just one at advantage it is not an advantage. You should be ready to have both of them. So because as I like all all our podcast was about this AI and the project manager and in 2026 just to have one advantage it is not an advantage you should be ready to have a lot of let's say skills a lot of advantages just to be ready to use them on a correct for for each situation to use the correct one so but you should be ready for all of them. Julius Neil Buenconsejo: Okay, thank you for that. So we're going to the rapid 10x round, our last part of the podcast. So my first question, what's one AI tool you actually use every day? MARIAM PETROSYAN: chat GPT and and copilot. So I I'm I'm using like Chad GPD for Julius Neil Buenconsejo: Okay. And go file. MARIAM PETROSYAN: my personal work and the copilot for my working purposes. Julius Neil Buenconsejo: Okay, what's one task AI has already 10x or improved in your life? MARIAM PETROSYAN: I would say that just turning this scattered information into structured thinking, it shortens like the distance between information and action by leaving the final analysis and decision with me. So Julius Neil Buenconsejo: What's one human scale AI can ever replace? MARIAM PETROSYAN: Judgment. Yeah. Julius Neil Buenconsejo: If you could automate one annoying thing forever, what would it be? MARIAM PETROSYAN: I guess everything related to meetings, except actual conversation. Julius Neil Buenconsejo: You MARIAM PETROSYAN: The rest can be automated, yeah. Julius Neil Buenconsejo: Okay, okay. What's the biggest misconception about AI right now? MARIAM PETROSYAN: perception is this halo effect that one impressive result make us just assume that the tool is accurate and suitable for everything and one success should earn our interest, not like our blind trust. So this hollow effect. Julius Neil Buenconsejo: What advice would you give someone who wants to become 10x starting today? MARIAM PETROSYAN: just start with real problem, not the technology. just as you know I already provide this experience of this agile and waterfall. agile needed to fit the project rather than replace the waterfall automatically. Choose the right AI tool for the Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: right problem and adapt it to your context. Julius Neil Buenconsejo: What would you tell your younger self about the future? MARIAM PETROSYAN: learn, learn and keep learning. Julius Neil Buenconsejo: Okay, what's one book, podcast, or resource that shaped your thinking about AI or innovation? MARIAM PETROSYAN: you know it's it's a very good question. but when speaking about the AI, I wouldn't say let's say a book, but I would say the science fiction film such as the fifth element. technologies that Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: once like felt impossible increasingly resemble everyday life today. impossible idea can become tomorrow's normal. So what we see in these fiction movies, Julius Neil Buenconsejo: Mm-hmm. MARIAM PETROSYAN: today it became our real life. Julius Neil Buenconsejo: Okay, if you had to bet on one AI trend that will be mass-savvy in three years, what is it? MARIAM PETROSYAN: so I would say about the agentic AI embedded in everyday workflows. Collecting information, preparing recommendations and coordinating actions with clear limits, audit trails and human oversight. So Julius Neil Buenconsejo: In one sentence, what does 10X mean to you personally? MARIAM PETROSYAN: Tanix is not about working ten times harder. it is about combining technology, j human judgment, continuous learning and strength of people to create ten times more meaningful impact. Julius Neil Buenconsejo: All right. Thank you very much for that. You know, for that podcast is very insightful and learn a lot from you when it comes to the project management, and AI era and also about Arimena. Of course, I will, I was able to learn a lot from you. And then if the people who are listening right now and watching right now, where do they want to connect with you? Where do they connect with you and learn more about your work? MARIAM PETROSYAN: yeah, they can connect with me on the LinkedIn so you can just share my LinkedIn. Yes, details yeah here. So and they Julius Neil Buenconsejo: Yes, yeah, yeah. MARIAM PETROSYAN: can just if they have some questions or some want to have some discussion in about some specific topic, so we can connect it on the LinkedIn. Julius Neil Buenconsejo: All right. Thank you very much. So thanks for watching 10xAI. Remember the future doesn't belong to AI. It belongs to the people who use it. See you next episode. MARIAM PETROSYAN: Thank you.