Thriving in Engineering

Colette Molteni works in data analytics and AI governance inside a major tech company, where her job has shifted in the last six months from building dashboards and presenting to leadership to building the AI agents that do that work instead. She also writes Empathy Elevated, a newsletter on human skills for tech professionals, and has a decade of experience watching organizations bolt AI onto broken processes and wonder why it doesn't work.

In this conversation, we discuss:
  1. Why slapping an AI agent on a broken process is like hiring an intern with no onboarding — and what to fix first
  2. The 80% problem: why AI gets most of a task right, but the last stretch still needs a human — true for code and dashboards alike
  3. Whether empathy skills transfer to working with AI agents — and why a calm, clear communication style gets better results
  4. How AI note-takers are quietly letting people zone out in meetings — and why that's a trust problem
  5. Why team politics don't get smaller as teams get smaller — they just change shape
  6. Her actual prediction on the AI transformation timeline — and why she doesn't buy the four-hour workday by 2027
Colette's details:
Empathy Elevated Newsletter - https://empathyelevated.substack.com/
LinkedIn - https://www.linkedin.com/in/colettemolteni/

What is Thriving in Engineering?

Software engineering isn't just strategy and sprint plans. It's ambiguity, pressure, trade-offs, and people.

Thriving In Engineering helps you navigate all of it.

In the newsletter, Alex Ponomarev distills 25 years of engineering leadership into practical frameworks to help you lead with confidence, stay calm under pressure, and build thriving teams.

The podcast goes deeper, talking to engineers, leaders, and builders across the industry about what AI transformation actually looks like in practice, past the headlines and hype. One guest at a time.

Hello. I'm Alex Ponomarev and welcome to Thriving in Engineering Podcast. In each episode I talk about

people from across the software development industry about AI. We're not talking about the

headlines, the hype, the doom mongering. Instead, we focus on honest conversations about how AI is

actually showing up in day-to-day work. What's working, what isn't, and what others can learn

from it. Today, I'm delighted to be joined by Colette Molteni. Colette is in data analytics

and governance, which means she's one of the people thinking hard about how AI gets deployed

responsibly inside organizations, not just how it gets built. She also writes Empathy Elevated,

a Substack newsletter exploring human skills for tech professionals. And she's a familiar face here

on Thriving in Engineering. She wrote a brilliant guest post for us earlier this year about what

happens when a no isn't really a no and what it takes to get to the truth of it. She also

just launched a paid tier of her newsletter, but I'll let her tell you about that. Colette,

welcome to the podcast. Tell us about the changes in your newsletter. What's what's new? Yeah. Hi,

Alex. I'm so glad to be here. Thank you for the introduction. Yes. So, the Empathy

Elevated Substack newsletter. I've had it live for roughly about a year and a half. Very recently,

I did adjust it. So, we now have a paid tier available. There's a few benefits of that. One

of the major benefits is access to the inference engine which is essentially a vault of resources

documentation to take you play by play essentially decoding the human side of working in tech whether

you're an engineer an analyst product manager you name it that part I know can always be a bit vague

and ambiguous I don't have fluff in there I have steps that you can take for real results. That's

amazing. I think empathy is something that becomes increasingly more important as we get

surrounded by AI and some robots and so on. Tell us how is the reality of working in AI day-to-day

affecting you? What does your day-to-day actually look like and what changed in the last few years

as things exploded? Yeah, great question, Alex. And I'll be honest with you, my work, the nature

of it day-to-day has changed. Not so much even in the past few years, really the past six months,

it's changed drastically. And my role was more so more of a traditional sales operations role

within a major tech company working in the SaaS division and I was doing more of the usual in the

background working with stakeholders for building out of dashboards working through troubleshooting

within various programs presentation to leadership about what we're innovating what we

are implementing. My work has transitioned though drastically in the past six months where instead

of talking about how we are going to provide direction to our stakeholders often I work with

leaders, customer success managers, sales etc. It's more about okay how can we translate this

information into the agent that we are going to provide for them to save them this time versus me

just talking at them. I'm giving them a resource that they can leverage in real time based off the

feedback I've received. So my job has transitioned from traditional building out of resources and

building the resource the AI resource that they can leverage. That's that's very interesting. Do

you feel the pace of change in your day-to-day work because something new is coming out every

week, every day really? Yes, actually and I had this conversation quite recently even with some of

the vendors that supply one of the programs that I work in very frequently how there are changes now

it seems month to month even halfway through our deployment of one of our AI agents recently the UI

changed for the better but we did have to go back and make some tweaks in the background to make it

applicable to this new UI I changes I remember early in my career and I've been in the world

of data analytics for over a decade now where you saw changes in deployments. It seemed every year,

every two to three years a new version of something. You're now looking at every month,

two, three months max in some cases. Yeah, we just live in this constant change. Because

before you could say, okay, there's this new technology came out. whatever it is it was

blockchain right before that we had big data and you know mobile but you could say okay

it's going to take us I don't know maybe a year to adapt maybe it feels like this now that's

how I think but now never ending change we just live in it we we get try to get comfortable there

I wanted to ask you how the environment around you changed because we definitely all feel that and

this changes the like human-to-human dynamics in the team and I'm curious if you're noticing this

and how how do you think we can help each other to to just not just survive but also thrive in this

new environment? Yeah, it's definitely changed. There are some things that have not changed.

The things that have not changed, we still have to deal with office politics, corporate environments,

different office cultures, whether you are remote, hybrid, in person, working with others depending

different geographies across different time zones. That has not changed. Now having AI in

the picture though has enabled us to do some other things quite well. It's also I think in some other

ways it's made us not participating in other things quite as well either. Let me break those

both down for us. One thing I think that it's done very well in terms of I'll give an example like in

meetings having a meeting notetaker. Now I still actually like to take some handwritten notes. I

just find my own comprehension. I like to have that and even the AI noteaker any program I've

used will almost always warn you make sure to validate for accuracy and I've seen that I've

seen that where it misinterprets something I know from a very large meeting I was recently a part

of it referenced a program that we haven't even used in over a year but because something in the

conversation had picked up it misinterpreted us and it was like revise this thing in this program

which we had no plans to do okay so where it can come in to be not so good. And this actually ties

into the first point if we are too reliant on it where it is used as an excuse to perhaps zone out

because oh, I'm going to catch up on the notes later or I'm going to rely on this as if it is

100% true if everything got documented correctly and not take a critical eye to it because you look

at something and maybe you're you're using an AI tool. you really do need to look. I am I don't

shy away from leveraging AI perhaps to double check an email correspondence I'm going to send

out. I had the other day though three times in a row where it was referencing an incorrect person

on the chain [laughter] and eventually I just I just had to correct that one sentence. The rest

of it was fine. But something like that can could cause some friction mistrust. You know

why are you referencing this? we could make that human interaction more complicated. So, there's

benefits and also areas where you can fall into pitfalls. It's a balance. I'm also noticing that

if you don't catch these little things, it it kind of hurts the trust because then people think, oh,

you're just blindly copy pasting whatever AI tells you. So, yeah, definitely interesting dynamic. Do

you think at least today AI creates more problems in the workplace for us and by problems I mean

like specific things that I see for example. Yeah, workload is part of it or expecting more

output wi-i which it does help you to move quicker but I think there are still some misconceptions

about depending on the AI tooling you are using to what extent it can help you. For example,

one of the things I always emphasize and we have as part of our framework on our team where we

deploy AI is we have to have the foundation of the process and the tasks and be able to break it down

into multiple parts. I come from a lean six sigma background where we did this day in and day out

even before we were leveraging AI. Without that foundation, you cannot train an AI agent generally

just to go out and guess. I mean, if that's the directive you're giving it to get advice, sure.

But do not expect the AI you are building to do it better than perhaps you could say like an intern

who is just starting on the job without much direction. You have to provide them directives

on what to do. If you have a broken process to begin with, slapping an AI agent on top of it,

if you have not worked through and done that groundwork, which is not as exciting as having

the AI, you're not necessarily going to get the end result you want. So the problem in summary

with that is there being a misconception when a problem arises which perhaps is a process issue

or a documentation issue or a staffing issue or a skills gap where it's like oh well can we build AI

for that what do you want to build the AI for what are you tackling what is the challenge and then

let's explore the AI if you don't get past that you have to get past that. And it sounds like the

misconception is that no like this idea of we're not going to need humans with AI and it at least

so far it looks like we actually do need humans to manage the agents to build a process around

them. Kind of like you need surprise surprise managers who seem to do nothing. They're just

moving between I don't know between the raindrops don't really do anything. But somehow everything

falls apart without managers. I mean I certainly as a manager myself sometimes I look at my day and

I'm like where did my day go? On that topic though with AI and the human intervention you bring up at

this stage yes. Will we see more autonomous AI in the future? I believe we will and we already are

in some places. What I interpret more as I look to the future though is and we hear about it the

creation of more jobs as a result of AI jobs we cannot even dream of today. I mean truthfully the

work I'm doing now with working and coding AI agents and building this is something I would

have never imagined growing up in my teen years. I would have never even I had no concept of this

back then. I was teaching myself code and how to build websites and I had no assistance other than

giant books that I bought at the bookstore and testing. I had no vibe coding type technology

available to me and I think that's only going to exponentially increase what we are able to do.

Don't you think AI is taking the fun out of what we're doing day-to-day or maybe it's the

contrary and it actually makes us do more creative things and you know does the routine for us. Yeah,

I think it depends on your role truthfully. It depends on your role and what you personally

extract your joy from and what you enjoy doing. Like personally I've never like the experience

of going circles with my code and continuously getting an error and it being vague and not being

able to figure it out and AI does not always resolve it for me. But as of late I have found

working with AI to bounce my code off of and hey what am I missing or maybe it's just like a syntax

error something so silly because my coffee has not kicked in yet in my system and I'm

not looking at it clearly. I find it very helpful for those types of things. The parts that I have

found with AI that have really helped me taking some of the effort out of having to make like a

slide deck that I have to present. I have no issue presenting. I actually like presenting and getting

up. I'm one of those odd people who actually likes to do it like being able to share the message

to a room and be able to talk about the work me and my team is doing. But I've been able to get

rid of some of that drudgery of being able to put together the slides like, "Oh, did I put these

slides in the right order? And I know this is my audience. I know this is the room. Am I stacking

this in the right order?" Is like a person who knows nothing about the topic, does this make

sense what I put together? And leveraging AI as a partner for that. I found that to be quite

helpful. Yeah. Given I have the ultimate say, I'm the human. I know the emotions that the room and

the personalities and people will face. But I have found that that that has helped me bring

actually more joy to what I'm doing in my daily work. Being able to take care of those mundane

tasks and have those handled, it frees up my time. Yeah I'm also in this camp. They remember in this

in the early days of ChatGPT there were stories of AI model breaking out of the researchers computer

and taking over something like taking over the internet something like this like what do you

think is real and what what is hype today? Now I don't have a crystal ball but I will say just for

me personally I am enthusiastic about the future. For me I view what is happening with technology

the next level of us leveling up as humanity. When we have these tools at our fingertips to take care

of these mundane and more tactile tasks. We will free up more time to build community,

go back to our roots of connecting with others and being less secluded, which I can be, I work

from home most days and sometimes I don't interact with other people in person. But I have found with

AI it being able to free up my time. Will that positive net benefit though come immediately? No.

And right now, I'd honestly say we are in a bit of that growing pain stage where truthfully there

are jobs being lost. There are people who are going to have to re-skill. Those new jobs are not

being recreated just like that with a snap of the fingers. It's going to take some time for us to

fully I think realize the benefits. We're seeing some benefits but we're also witnessing loss and

those new jobs that are going to come about that we can't even dream of. The new time savings,

the innovation with it, we're going through that growth period. In terms of worst case scenarios,

I mean, given I'm a big sci-fi person. I'm a huge fan of the Terminator series. So, that too would

give me some nightmares. And I I do make jokes about it sometimes with Skynets or anything else.

Do I think that's what we are facing? No. Here is another question that I'm I'm sure you will

appreciate. In the beginning you mentioned that yes there a lot of things are changing but are

a lot of things are staying the same like the team dynamics, the politics in the corporate world,

and all these things they're not going anywhere there is this idea that organizations will become

smaller that because of AI there will be larger number of companies, of products and the teams

will be smaller because one person will be able to do 20 things at a time and so on. Do you think in

smaller teams that we'll see more of the topic of politics and interpersonal relationships is less

is less important than for larger organizations, larger teams. Yeah. Having worked in both small

teams and larger teams, I've worked in a company of only about like 30 people all the way up to

like 30,000 people. I've worked in both extremes. So, I'd unequivocally say no, it it's not less

relevant. It just changes how you approach it. Whether you are talking to an audience of perhaps

one person more like a one-on-one versus talking to five people or you're presenting to 500,

it changes the dynamic of how you approach empathizing with your audience and kicking in

your other emotional intelligence skills. But the skills are still relevant. It's just knowing how

to adapt for that particular environment. That is why your library that will be part of your

of the paid tier on your newsletter will always be valuable and interesting. Do you think that

we can draw parallels between how a manager works with their human team and with how you would

manage AI agents? And of course agents they don't have emotions and empathy per se is not probably

not applicable, but for me empathy is about understanding how other people view the world

and adjusting your message to be received the best way possible. Do you think skills like that

are relevant when working with AI agents or LLMs? Yeah, absolutely. And here's why. As much as the

AI agent does not have feelings, of course, it only has like a textbook definition of empathy or

emotional intelligence or how to read the room per se. But for you as the person who is programming

the agent, well, this agent is going to be people facing in the end, right? So it you have to

train that in to be able to at least take in the information even though it's not feeling anything.

You have to provide that human input. You yourself though aside from AI agents if you're working with

AI technology if you are cool calm collected and you are explaining your work in a manner

that is understandable even to this inanimate technology that does not feel anything. You're

probably going to get better results versus being annoyed and I mean I guess you know writing in all

caps and being you know irritated with it. That doesn't help you at all. It just it will mirror

that back to you and you will not get the type of work quality product you want. I've noticed

this interestingly with somebody that I work with with a very high degree of EQ who is very detail-

oriented and also writes out things as if she is talking to a person and explaining it very clear

and building upon a conversation when she's trying to build out something programmatically and her

results are significantly improved. Conversely, she also recognizes and has admitted, "Oh, I got

so frustrated. It wasn't giving me what I wanted and then it just stopped talking to me working

with the chatbot." [laughter] And that's that's not that's not based out of emotion, right? It's

not it's not that it's not that the AI is angry at you. It's just I I suppose it's overloaded or

your directives are not clear at that point. So, there isn't really anything to provide you back.

So yes, understanding how to engage with other people is helpful and applicable to how you

work with AI technology. It's a very interesting direction. What I see among the engineers is that

a lot of us are either introverted or just not great at communication for all sorts of reasons.

But I'm noticing that how engineers communicate after working with Claude Code, Cursor,

Codex is changing because as you said they get used to explaining themselves very well

providing the context surprisingly you get better results if you provide good context

and they're like oh this is how it works and now they're doing this with humans with their team

teammates. Do you think that the result of of the transformation that is going on that we will see

the result of the transformation in let's say in 2027 like a lot of people are predicting or it's

like we're very early in the process and it's something that already started but most likely

will arrive at where we're supposed to arrive in 5 to 10 years. Yeah, I if I had to bet on it,

I would say probably a few years, not 2027. I think we are getting closer to where we want to

be in the transformation and people talking openly about, hey, these are the limitations of AI.

These are the advantages this is what it can replace this is what it cannot replace and having

that dialogue but having that transformation where we want to see where hey you know what let's say

maybe I'm just saying this hypothetically maybe as humanity we don't have to work as hard maybe

we are down to that 4-hour a day work day right I mean that that would be fantastic I think a lot of

people would love that to have more leisure time are we going to get there as early as 2027 2028

unlikely. No, not at the current trajectory. It will take more movement and transformation and

collective change to see that happen. Could we get there eventually? Never say never. I mean,

we had our ancestors working before us sometimes 12 or 14 hour days. You know, farm work, very

manual labor. Look how far we have come and even a hundred years really blink of an eye so quick.

So never say never. If you could give one advice to to our audience what would be the thing you

would recommend them to focus on to you know to thrive in the age of AI but also to stay human and

to stay empathetic and Yeah. Yeah. Well, you just named one of the words of course just

as the name implies with my Substack Empathy Elevated. Yeah, elevating your empathy. What

I would say is working consciously to do this. 98% of the population, we are capable of being

empathetic. There there's a small percentage that actually is not, but that's for another

chat some other time. It can be learned. It can be developed these people skills even though the

task feels impossible at times with the right tools, the right systems and conscious effort.

You gave that example earlier of some of the engineers that you've witnessed who are a little

bit more introverted or perhaps like the people part is a little bit tough. Very early in my

career, I was silent in most meetings. I could not speak up. I felt that perhaps I was interrupting

people and then when I did speak up, I would often get puzzled looks. People didn't know what I was

talking about. I've been affectionately called quirky before. I'm still sometimes called that,

but only in a fun way because of my humor at times. But the point being that this is not it's

not magic. It's not some far-off concept that is abstract. You can obtain it and having that

differentiating factor with your humanity as you said Alex regardless of the technology regardless

of the era this will continue to be relevant to you as a human as your differentiator as your

strength as we yeah encounter on this new frontier where technology is changing monthly. This is

great advice thank you so much I'm sure our viewers and listeners would appreciate it. Great

but thank you so much for joining us today. It was a great conversation. We're looking forward to

read more of the episodes of your newsletter and we'll enjoy your new library and paid membership.

Thank you for joining. Yeah. Thank you, Alex, for having me on. Thank you. Bye-bye. Bye.