# AI Strategy: Decision Quality, Trust, and Practical Implementation

**Podcast:** Product Momentum Podcast
**Published:** 2026-04-29

## Transcript

Hey everybody, I'm Pete Sullivan, producer of Product Momentum, a podcast powered by ITX.
For more than 180 episodes, I've had the best seat in the house, learning right alongside you from some of the brightest minds in the software space.
Marty Kagan and Teresa Torres, April Dunford and Richard Mironoff, Jared Spool, Petra Villa, Matt LeMay, Nazreen Changwell, and so many more.
I'm reaching out today.
to let you know that the Product Momentum team is launching a new bit of content for you.
We're calling it the Topics and Products Series, or TIPS for short.
With TIPS, our goal is to engage with the same important products-based issues that we do with Product Momentum, but this time through the experiences of ITX product managers, UX researchers and designers, engineers, security analysts, and the whole rest of the team.
For nearly 30 years, ITX has been building custom software to help our clients solve complex business challenges.
With tips, we provide a new forum for us to experience these many challenges through the lens of ITX experts.
I invite you to have a look.
Let us know what you think.
If you like what you see, be sure to like, comment, and share.
And of course, if there's things that you think we can improve upon, I hope you'll let me know that too.
Thanks for tuning in.
Sean and Andrew were recently in New York City at a small product leader's breakfast run by Perna Singh and hosted by ITX, focusing on AI.
I'm really excited to hear you.
What did you talk about?
What surprised you?
So first, Perna put on an excellent event.
We're going to go back in April.
Really looking forward to doing it again.
I think the thing that hit me the most was really, we're talking about AI.
But the conversation was very anti-hype.
It was very practical.
Everything that we were talking about was like, what are we actually doing?
Where are we trying to go?
And how are we trying to make small leaps versus we are going to automate all of these workflows and here's what we're doing that's changing the world.
It seems like you guys had a really safe space to have, Brianna.
a real conversation, right?
Tell me a little bit about like, how did this event come to happen?
Yeah, ITX has been hosting and sponsoring these types of intimate events over the past few years.
But we found this intimate setting, if you think of 10 to 15 people, all kind of at the same level of their career journeys, you know, those are where we find the best discussions kind of become unleashed.
So this is really in the spirit of experimenting.
We've been talking to Perna for a long time and she's like, I...
put together this breakfast, we talk about these things, and I have a great audience.
We were just like, can we come along?
Can we help you out and see who else we can bring?
And that's pretty much it.
It's like an experiment that went really well.
I think people were excited for, especially if you're running an organization, if you're a leader within an organization, empowering the people below them with the right tools to do more, but not necessarily to go out and build new products right away or build improvements to products right away because they still want to make sure, again, we talked about...
core product skills.
They want to make sure that we're actually solving the right problem and not just going straight to, I have an idea, let's build it.
Biggest differentiator moving forward is going to be decision quality, right?
Your ability to consistently make good decisions, right?
Have some agency within your organization so that you are not the sole person making these decisions, right?
You've got the weight of an organization behind your decisions.
I think that's one reason.
Sorry, Dan, I got to do it.
Got to sell just a minute again.
But I think that's one reason why when people partner with...
Companies like ITX, we're bringing in a layer of outside experience and decision making that might not exist in your internal team.
And that's OK.
These things are compounding.
They help each other.
They grow with each other.
But that judgment and decision making, I think Sean nails it.
The biggest piece of advice for me coming out or from me coming out of that product leaders breakfast is for those of us that are getting that direction kind of top down of do AI, right?
maybe even a little bit more refined of like, where can we use AI?
I think the question back up or the question that you need to ponder and really respond with is kind of like, what work do we want to go away?
All right, welcome to the pod.
Today is a special one.
Sean Murray is leaving the co-host chair and going into the guest seat and joining us is Andrew Knobloch.
Sean and Andrew were recently in New York City at a small product leader's breakfast.
run by Perna Singh and hosted by ITX, focusing on AI.
Guys, you came out of it, both of you independently, super pumped.
I'm really excited to hear you.
What did you talk about?
What surprised you?
So first, Perna put on an excellent event.
We're going to go back in April.
Really looking forward to doing it again.
I think the thing that hit me the most was really, we're talking about AI.
But the conversation was very anti-hype.
It was very practical.
Everything that we were talking about was like, what are we actually doing?
What are we, where are we trying to go?
And how are we trying to make like small leaps versus we are going to automate all of these workflows and here's what we're doing that's changing the world, right?
It was kind of nice that everybody was kind of on the same page of like relieved that we were having this conversation and that people were in the same place.
in terms of their, like, organizational AI journey.
I think The Breakfast really made it clear that it wasn't a trend discussion, right?
Like, what are the coolest things that we were doing?
It was really about, like, actually, this is where my organization is.
Here's where I'm having struggles.
Here's where I'm having some success.
And the conversation was really grounded in that.
Yeah, I think everyone, what I found interesting right from the jump was there was sort of people who are further along in the AI journey and maybe people who are not as far along.
But everyone had some unique value of discussion to add to the group.
And I think that's happening in the industry as well.
There's this building sense of FOMO and trends.
And this group, for whatever reason, came together and said, let's have a conversation about what's really happening in our orgs, right?
How can we actually implement in a smart, strategic way that starts from maybe smaller, less risky and build up?
It was a great event.
And I was laughing with Sean that...
and the group that we started breakfast and everyone goes around and they start talking about the problem space.
Such product leaders, right?
They just dive right into the problem space.
So that's what I knew.
About 30 seconds in, this is going to be a really great group.
There's an important distinction in what Andrew said, right?
There were some people who personally, I think that their AI fluency was very high, right?
And they were able to do things with AI that maybe either the rest of the room wasn't able to do.
But more important, like their organization was just behind their organizational fluency was behind their personal fluency.
So how do I actually pull people forward and how do I pull this organization forward in its fluency so that we can actually start to make some, you know, quantifiable progress?
And now I have FOMO, right?
It seems like you guys had a really safe space to have a real conversation, right?
You know, social media, right?
Things like, you know.
AI, the best, if you're not doing it, you're going to be unemployed by Tuesday.
Tell me a little bit about, like, how did this event come to happen?
Yeah, ITX has been hosting and sponsoring these types of intimate events over the past few years.
We've also, you know, like a lot of companies, we go to larger events like Industry, Byproduct Collective.
But we found this intimate setting, if you think of 10 to 15 people, all kind of at the same level of their career journeys, you know.
Those are where we find the best discussions kind of become unleashed.
So this is really in the spirit of experimenting.
We've been talking to Perna for a long time.
And she's like, I put together this breakfast.
We talk about these things.
And I have a great audience.
We were just like, can we come along?
Can we help you out and see who else we can bring?
And that's pretty much it.
It's like an experiment that went really well.
And like Sean said, we're going back in April.
But that intimacy is so much different than a large conference room, right?
You have people that are looking for jobs.
You have CEOs and everything in between.
So I think the magic sauce is if you get the right people in the room, the conversation takes off.
So you mentioned kind of in the beginning that it was more than an AI feature, right?
You guys weren't sitting around talking about building like the coolest agent.
And there is a conversation about trust.
Sean, I'd like to hear kind of how that went down.
Yeah, it was interesting.
It was kind of like trust became not a technical conversation.
It was more of a...
product conversation, right?
I talked earlier about, you know, building things in small stages.
And so we kind of reinforced as a group that trust with AI was going to be built in increments, not in giant leaps, right?
So people were more comfortable kind of like approving or correcting AI output than like surrendering control completely.
And I think that that's probably where we are as kind of like in the business industry right now.
I think a lot of the direction that we're getting from the top down is do AI, right?
And we on the ground floor are like, well, that could mean a lot of different things.
And how much do you want AI to control versus how much do you want us to be in control?
Where do you actually want to see your improvements?
What are the outcomes that you're looking for?
And we got back to really kind of going back to our conversation with Teresa Torres, Dan, like, are we doing the things?
are we doing the best practices in the initiation and discovery phase of the SDLC to ensure that when we actually get into building something that's going to have some sort of relevant business value?
And we constantly have stated that, you know, core product skills are more important now than ever.
And that was really reinforced at this breakfast, right?
The ability to define a problem and understand if users are going to buy, right, or subscribe to what you're building.
Like, are you actually going to get some form of business value out of it, right?
And none of that.
involves ai right none of that involves like the actual ai coming in and taking over a workflow or actually being your developer for you right it's still like you actually have to make good decisions before you can before you can go and you know move into that like super small development space now dan if i could take another another angle of what i was hearing from a trust perspective in the room was building trust amongst your organization and the stakeholders that are going to play a part in this AI agent, whatever it might be.
First, there was this conversation about, like, let's not take the highest risk, most pressure-filled problem and solve that with AI.
Like, that's going to lead to problems.
So pick a medium to low risk, but pick a pain point that people struggle with that can be solved and start there.
That was sort of an aligned recommendation from the group.
But then this idea of...
Building trust amongst others in your organization.
We all do this every day with our clients and we do this with our own teammates here is what are people's concerns?
What do they care about?
And what are their success criteria?
Is what is legal care about?
What do they need to know in order to get this forward?
What does product care about?
So that that's really that conversation happened, you know, both on the AI agent side, the product side, but the organizational side to really.
was I walked away super impressed with, yeah, that's a great way to approach it.
Like you have to build trust amongst your team and you have to make sure that their concerns are heard.
And that's something, not to be sales guy, that's what I do, but like that's something ITX does all the time in our work with our clients.
Yeah.
I was surprised about how much of the conversation shifted towards how do we bring in legal and compliance early?
And how do we effectively partner with them to ensure that wherever we implement AI, that we're guarded from a legal perspective?
And as someone who's kind of built a career on asking for forgiveness, it was a reminder that these are huge changes and could have far-reaching implications that maybe I'm not thinking of.
So have some agency within your organization and really you know, shop around what you're doing so that you can have the best result possible.
Yeah, I think there's something like really important there as a product person building trust.
Like sometimes you have to slow down and actually vet that the solution isn't going to increase business risk, right?
It's great if you can get to market fast, but then if you're opening up huge legal vulnerabilities, that's not winning.
Super cool conversation, right?
I think I'm trying to go right to like a conversation of core product skills.
As product people, I...
we're only successful if our teams trust us, if our business partners trust us, right, to have thought through what we're doing, to have done the discovery, to be honest with them on what the solution can do, what it's not going to do.
Super cool.
I'm curious, you know, from the discussion, what were the product leaders in the rooms talking about in terms of like, how are they bringing these changes within their organization?
You know, you mentioned that personally, maybe they're, yeah, a step.
eight in their AI journey, but their organizations are at stage three.
How are they getting it worse at stage four?
I think initially it's finding repeatable processes within workflows, right?
So what are you doing today, right?
That's either a bottleneck, right?
Or is something that you think, you know, with some reasonable safeguards could be automated, right?
Where can they make decisions where the human is really driving, but the AI is assisting?
Right.
So how can I, you know, how can we tighten up repeatable processes within workflows that we have today?
Or where can we get to a point where hopefully the AI is now driving and human is basically just approving that output.
Right.
But nowhere was anybody like AI is going to control this specific portion of the workflow and we are going to handle the input and then we're going to deal with the output.
Right.
It was we want to work together, treat AI as a teammate the whole way through any new workflow.
Again, it kept going back to trust, right?
The biggest mistake that teams seem to make is trying to go from, yeah, we've put an agent right here in this workflow and we're just going to trust what comes out of it, right?
Nobody's really there yet, but they're looking forward to getting to the point where they've perfected the repeatable process enough where the agent can at least get to the point where the human's got to make minimal decisions coming out of that process.
Am I allowed to swear on this podcast, Dan?
Yes.
Oh, perfect.
Absolutely.
Well, that's what Perna sent out after in her recap was, you know, we're 10xing shitty things, right?
And if you pick the wrong thing and you just put all this AI juice behind it and you never worked through this model, this co-pilot model that was shared, you're going to end up putting a ton of juice behind maybe the wrong thing.
So that was enlightening to me.
And I think to Sean's point, it's almost like the discussion kept going back to first product principles.
And human drives AI, then AI drives and the human approves.
That final stage, full automation, it was the room felt like most things don't actually ever get there right now.
At least right now, that's where we're at.
So that was a really cool framework to leave the discussion with.
I think people were excited for, especially if you're running an organization, if you're a leader within an organization, empowering the people below them with the right tools to do more.
but not necessarily to go out and build new products right away or build improvements to products right away because they still want to make sure, again, we talked about core product skills.
They want to make sure that we're actually solving the right problem and not just going straight to, I have an idea, let's build it, right?
So just because you can, yeah, you can equip your organization with those tools and you can tell people to go and build, but now are you going to be validating five bad ideas a day that have already gotten into the prototype stage, right?
That's going to be a real...
It's going to be a real resource drain if we get to that point.
So how do we avoid those things?
Right.
We just have to deliver clear expectations about how we want AI to be used within our within our workflows.
Right.
And there's I think there's a lot of opportunity to potentially damage brands.
Right.
By firing all that, you know, the shitty stuff out there.
And now suddenly your reputation takes a hit because you haven't gone through and done your done proper discovery, figure out like, hey, what's actually important to users?
So, I mean, what does this mean for the product leaders who are out there or in this world of AI changes are coming fast and furious?
There's a lot of organizational pressure to bring AI solutions into the market.
What did you guys talk about and how can that be applied to the folks who weren't in the room?
Biggest differentiator moving forward is going to be decision quality, right?
Your ability to consistently make good decisions, right?
have some agency within your organization so that you are not the sole person making these decisions, right?
You've got the weight of an organization behind your decisions.
But it's not necessarily people that can prompt the machine the best, right?
Or the people who have the highest AI fluency.
It's going to be people who consistently make good decisions, right?
And are able to back up their judgment.
I think that's one reason.
Sorry, Dan, I got to do it.
Got to sell just a minute again.
But I think that's one reason why when people partner with...
companies like ITX, we're bringing in a layer of outside experience and decision-making that might not exist in your internal team.
And that's okay.
These things are compounding.
They help each other.
They grow with each other.
But that judgment and decision-making, I think Sean nails it.
This idea of AI fluency and becoming more fluent in AI is not opposed to still making great decisions and great judgments.
You all had a, I'm blanking on the podcast guest's name.
who the the key quote out of that episode was like we're paid to decide right that's what yeah yeah i love that like i love that you're not paid to be right or wrong you're paid to make the decision that's not going away with all this ai for sure yeah and i think yeah but there is something to be said about having like kind of that trusted outside voice right it's so easy to get stuck, yo, in your own cycles, your own knowledge biases and have somebody come in with a fresh eye and be like, hey, actually, you know, maybe the decision model should move this way.
Yeah, it can be really powerful.
So, Andrew, from your standpoint, I'm just curious to hear you.
You're the non kind of professional product person here.
You know, what are a couple takeaways that you heard that you're like, oh, wow, like this is an interesting way that industry is moving?
I think that.
The idea that the whole village has to move forward and someone's going to have to help it move forward.
You know, the reaching out across the organization and getting that buy-in and understanding concerns, that's not going to go away.
There was a lot of talk also about security and privacy.
A lot of, you know, we can't share.
One of the cool things about these events is we can't share who is in the room.
Chatham House rules.
We can't attribute to what we're saying to specific people.
Definitely the security and privacy is still going to be very important going forward as more AI tools are released, more models are used, more data is feeding the machine.
Those are the things I walked away with is how do you reach across the aisle, talk to the right team members, and then how do you make sure that their concerns for security, privacy are met as you continue to grow?
Just two of the many things I walked away with.
It was an hour and a half conversation filled with great coffee.
So I'm probably missing some, but it was, those are the two that stick out right now.
Awesome.
Yeah, guys, it sounds like it was a really great conversation hitting a lot along a lot of really, really key lines, right?
So if I'm hearing what you guys are saying, you know, kind of the big things that came out of the conversation, right?
In this AI world, as product leaders, right, we need to be building trust.
We do that by making sure that we're collaborating, right?
Talking to folks in legal, making sure that security concerns are being met, hearing our business partners' concerns, hearing our team's concerns, and then building that in how we're thinking about the problem space.
And that discovery, the ideas that we're promoting, we have to really sharpen those before we unleash an AI tool into the process.
I think the biggest piece of advice for me coming out of that product leaders breakfast is for those of us that are getting.
that direction kind of top down of do AI, right?
Or maybe even a little bit more refined of like, where can we use AI?
I think the question back up or the question that you need to ponder and really respond with is kind of like, what work do we want to go away?
What don't we want to do anymore?
Because it's not fun.
We don't like it.
It has minimal value, right?
It has some value.
Maybe it drives a process, right?
Or it's like a cog in the wheel of a process.
But like, Those are the types that that's, that's your best opportunity to try to find a way to apply AI, right?
So when you get do AI, that's not put AI in the presentation layer of your product, right?
It's find some way to use these new tools to improve processes, right?
Reduce waste, right?
And improve quality.
I think that was the biggest, that was the biggest thing for me.
That, that direction of kind of like, find what you don't like to do, right?
For me, travel requests, email.
Right.
Those types of things are those are things that can be automated.
Right.
There are there are opportunities to apply AI into those simple situations.
Right.
Or those simple processes that are going to get you back to applying your expertise the way that you want to.
Right.
We have a team member right now that's working on kind of like a pet project for me where I think that email is email should be irrelevant messaging tools like Teams or Slack.
whatever your organization is using is really like the future of communication.
We really need to move there.
So how can I take like the emails that are sent and how can I have like some sort of automated process set up so that that email becomes a Teams message from that person, kind of references the other people that are either CC'd or BCC'd on that email, right?
But at least I get like the full email context in Teams and I never have to open that thing.
I never have to open up Outlook unless I'm trying to manage events.
Also, I found something that I don't like to do.
Just check email.
Right.
And turn it into something that I enjoy doing, which is managing messages.
You just went from you just went from first product principles at the beginning of this to let's kill email.
I love that.
I love that journey.
You know, as a guy who is out.
You're welcome.
It looks like the West Side Highway at rush hour.
I love it.
Awesome.
Guys, this has been a great conversation.
Real quick lightning round.
You guys said you're going back to New York for another conversation in April and you're playing train or automobile.
I'm going train this time.
I'm going train after a lovely sleepover at JFK airport last time.
So I'm excited.
Yeah.
First, thank you to Aperna, not just for the event, but also for telling us about the Long Island Railroad.
We had no idea what that was before we went in and then I ended up taking it three times.
because of our lovely travel experience.
I think I'm going plane this time, but I just got to look at the train schedule again.
If I had to do the train again, if I had the time to do the train, I would do it because having Wi-Fi and a seat, right, and the ability to move around a little bit, you basically get a ton of working time back or, you know, working with Cursor on my own app for a while.
But that was a great experience on the train.
Product guys talking features, Dan.
Do you love that?
I love it.
Awesome.
Shout out to Amtrak.