Can journalism teach us how to trust AI? with Dom Nicastro, Editor-in-Chief at Simpler Media Group
About the show
Hosted by Nikola Mrkšić, Co-founder and CEO of PolyAI, the Deep Learning with PolyAI podcast is the window into AI for CX leaders. We cut through hype in customer experience, support, and contact center AI — helping decision-makers understand what really matters.
Never miss an episode
Summary
In this episode of Deep Learning with PolyAI, Michelle Schroeder sits down with Dom Nicastro, Editor-in-Chief at Simpler Media Group and longtime journalist covering CX, content, and enterprise tech.
Dom has spent years watching AI reshape newsrooms — and he sees striking parallels with what CX leaders are facing today. Journalists and CX teams share the same challenge: cutting through hype, spotting “AI slop,” and building trust through accuracy, governance, and human judgment.
They explore:
- Why journalism’s skepticism is exactly what CX needs right now
- How AI accelerates good work — and exposes weak processes
- Why bad AI (in news or CX) destroys trust instantly
- What Dom hears backstage at CX events that rarely makes it into press releases
- The role of governance, vetting, and human oversight in customer-facing AI
- How CMSWire built “Roberta,” their internal RAG system to support real reporting — not replace it
- Dom’s take is clear: AI won’t replace humans, but it will reveal who’s doing the work well — and who isn’t.
👉 Watch the full episode to learn what journalism can teach CX leaders about building trustworthy AI.
Key Takeaways
- AI should support, not replace, human judgment: Dom argues that journalism’s goal — making sense of the noise — hasn’t changed. AI can speed up workflows and analysis, but the “human once-over” remains critical to maintain credibility and context.
- CX and journalism share the same challenge: Both fields depend on trust and governance. Whether you’re publishing a story or deploying a voice agent, success comes from vetting, context, and ensuring technology represents the brand faithfully.
- Noise vs. narrative: With 650+ AI vendors vying for attention, both reporters and CX leaders must filter hype from substance. The best insights come not from product pitches, but from customers actually using the tools and sharing honest results.
- AI can support using institutional memory: CMSWire’s custom RAG system, “Roberta,” uses generative AI to summarize decades of editorial archives — helping journalists connect today’s stories with historical context, proving that AI can deepen, not dilute, quality reporting.
Transcript
Michelle Schroeder
00:03 – 00:47
Welcome to Deep Learning with PolyAI.
We’re here to help CX leaders get a window into the latest and greatest in AI. If you enjoy this type of content, if you enjoy us, please give us a like on YouTube.
Leave a five star rating on your favorite podcast app, and please subscribe. Alright.
And with that, I would love to introduce our guest today who is Dom Nicastro, editor in chief of Simpler Media Group, which so many of our listeners will recognize as the publisher of CMS Wire. Dom’s been with Simpler Media for twelve years, and he’s been a journalist for over four decades.
Sounds like starting back in fifth grade. I cannot wait to hear a little bit more about that.
Dom Nicastro
00:47 – 00:47
Yep.
Michelle Schroeder
00:47 – 00:57
I’m super excited to have you on our show to talk about CX, AI, making connections, which has always been at the heart of your work. So welcome to the show, Dom.
Dom Nicastro
00:57 – 01:15
Yeah. Thanks for having me.
And, yeah, I’m glad you brought up the long distinguished career in journalism. Unfortunately, I was paid more at the Fuller Elementary School in Gloucester, Mass in the mid eighties.
I was paid a lot more than I am now. So but I do it for the love of the game, Michelle.
You know?
Michelle Schroeder
01:15 – 01:56
As do we all. I love it.
So, tradition has it that we typically share a new story, and then we will or opinion article, and we’ll get your opinion, get your thoughts on it, and I’ll ask you a few questions about it. And there have been some big news in the AI and journalism world this week.
NewsGuild just started a campaign called News, Not Slop. And it’s meant to call attention to keeping the human element in news publications to protect not just jobs, but quality, the quality of journalism.
So I would love to but before I ask you some questions about it, did you did you have a look? And what do. you think about it?
Dom Nicastro
01:56 – 02:01
I did. Why are journalists always so upset? You know?
Michelle Schroeder
02:01 – 02:02
You tell us. I mean,.
Dom Nicastro
02:02 – 02:03
know?
Michelle Schroeder
02:03 – 02:03
maybe it’s okay.
Dom Nicastro
02:03 – 03:17
people. I I hate I hate the PR people.
I I hate my editor. I hate everybody.
Cranky bunch, crayon. Cranky bunch.
But the the initiative is really good. Excuse me.
It is. It is.
You know, we I mean, I think it’s up to every single news organization to make sure the slop doesn’t enter the AI slop doesn’t, you know, lessen the quality of what you’re putting out. I mean, if you’re you’re trying to you’re trying to make as a journalist, you’re just trying to make sense of the noise, you know, and cut through it.
You know, people are telling you things all day in your inbox, wherever social, and they’re trying to get your attention. And you’re just trying to sort out what’s what’s real.
You know? What’s really having an impact. And, yes, if if AI if we need this concerted effort, lawsuits, and all that to to help move towards a more genuine I mean, I I don’t even think AI is the biggest problem that journalism faced.
You know? It’s it’s bias. You know? And they had AI had nothing to do with that.
You know?
Michelle Schroeder
03:17 – 03:19
Yes.
Dom Nicastro
03:19 – 03:29
It’s it’s the Adventist the Internet and social media and, you know, who’s who’s out there? There’s people out there saying a lot of things before AI. Right?
Michelle Schroeder
03:29 – 03:30
100%.
Dom Nicastro
03:30 – 03:41
And and and what’s genuine, what’s not. And that’s our job as journalists to take what’s been said and and filter and make sense of it to the end user, the consumer.
That’s it.
Michelle Schroeder
03:41 – 03:41
Totally.
Dom Nicastro
03:41 – 03:45
So it’s a fine effort that that that news guild effort.
Michelle Schroeder
03:45 – 03:45
Yeah.
Dom Nicastro
03:45 – 03:46
Yeah.
Michelle Schroeder
03:46 – 04:11
I mean, I I love how you said, you know, that the job of the journalist is, like, filtering through noise, you know, in service of the the reader, in service of the person who’s consuming it. And it’s AI definitely feels like it’s been adding to that noise, and a lot of it is slop.
And so, yeah, I agree with you on the initiative. Speaking of.
you know, we started this by by you saying, like, it’s a it’s a cranky bunch. I’m curious about the bunch that you work with.
Dom Nicastro
04:11 – 04:12
Are.
Michelle Schroeder
04:12 – 04:16
they using AI? How cranky are they about using an integrated?
Dom Nicastro
04:16 – 04:48
Yeah. Yeah.
Oh, they’re all curmudgeons. They’re horrible.
No. Of course.
Absolutely. I I 100% support use of AI into our journalism circles.
Now I think with the new that that issue and the lawsuit you brought up and and that effort is more about, you know, AI slop, meaning content that’s just generated from AI 100% only and pushed out into the world. That’s to that to me is slop.
Michelle Schroeder
04:48 – 04:49
Yeah.
Dom Nicastro
04:49 – 05:23
As a journalist, how can I do that and tell my readers if they actually had a question on it? Like, you know, like, I don’t know about this customer journey mapping tool that you ran. And what’s Dominic Castro gonna say? Like, I I don’t know.
Ask ask the LLM. I got it that that made it for me.
Right? So it’s I support the the weaving, the integration of AI into journalism 100%. And right now, it’s a matter of making us faster.
Michelle Schroeder
05:23 – 05:23
Yeah.
Dom Nicastro
05:23 – 05:29
I I won’t say it’s making our content better. You know what.
I mean? That’s up to us.
Michelle Schroeder
05:29 – 05:31
Yes.
Dom Nicastro
05:31 – 06:23
It’s made us faster, more efficient, and if we aren’t given it the the classic human once over, then that’s on us if we’re gonna roll something out and and and not do that human in the loop check, you know, because it’s I mean, we’re we’re we’re taking 9,000 word podcast transcripts and weaving them into some takeaways. And then.
you’re not reading those takeaways before you put them out there and saying, did did the LLM capture what we we me and this other dude talked about, then, that’s on you. So.
so I’m a 100%. behind it.
I am not like, I mean, I think some journalists in the beginning were like, no. No.
No. I don’t need that.
I’m, you know, I’m gonna interview the person and run it. And, hey, if you wanna edit a 9,000 word transcript, go ahead.
But I’m using an LLM 100% of the time on those.
Michelle Schroeder
06:23 – 06:49
Totally. Yeah.
It sounds like you’ve found a way to integrate it well where its job is to support speed and support the great sort of things that are already happening with your team. But.
that they’re still that their job is still primarily around quality. And I’m I’m curious if you think any of those lessons that you’ve learned, map onto how CX leaders might be thinking about where AI is gonna add the most value.
Dom Nicastro
06:49 – 07:04
I think so. I I think so because they’re also trying to communicate with people, right, with customers.
So what what is your I mean, if a if a company hires or implements a chatbot, right,.
Michelle Schroeder
07:04 – 07:06
Yep.
Dom Nicastro
07:06 – 07:11
that is a twenty four hour representative of your brand.
Michelle Schroeder
07:11 – 07:12
Mhmm.
Dom Nicastro
07:12 – 07:52
If you were to hire a human being that was like a twenty four hour representative of your brand that answered every question that every customer ever asked, would you just, like, kinda roll out that person and be like, alright. Alright.
Cool. You can you don’t have to sleep or anything? That’s all.
You know? You’re good? So, the vetting, the governance, and, you know, I have so many lessons I’ve learned from practitioners in the CX world from conferences I’ve been to, and I’m like a vessel of what they say. I’m, this Dominic Castro wasn’t the expert of how to weave CX and AI.
Michelle Schroeder
07:52 – 07:52
Yeah.
Dom Nicastro
07:52 – 07:55
Dominic Castro gets experts to tell him how to do it,.
Michelle Schroeder
07:55 – 07:56
Hello.
Dom Nicastro
07:56 – 09:13
and then he writes to other folks trying to do it. So, you know, I’m filtering what they’re saying, but the the things I’m hearing from them, you know, about CX and AI and and and and what are the challenges.
A lot of the people I talk to are getting more internally I wins, much faster than customer facing ones. This is like a small sample because we have we have survey data that says different things, but my one on one outside the keynote in the hallways of these conferences, you know, this is the things, you know, that they’re telling me.
They want it’s much easier to deploy AI where governance risks are low. That’s what a huge concern is.
So if they have that governance locked down, they’re they’re on board for coaching, QA, hiring, agent assistance, workforce management, analytics. But the because because the customer facing AI really requires, like, a heavier compliance, security, legal review.
So those are kind of some of the things that might hold them back from rollouts. So it’s it’s it’s in our surveys that we we’re we’re sending out, we’re seeing much more confidence in AI and CX, much more,.
Michelle Schroeder
09:13 – 09:13
Yeah.
Dom Nicastro
09:13 – 09:16
than, say, last year even.
Michelle Schroeder
09:16 – 10:22
Absolute we’re seeing the same. We’re definitely.
seeing the same. But going back to what you said about, you know, journalism and AI not being the problem, but you have this underlying bias issue, I think the same applies for, like, enterprises thinking about how to incorporate CX and AI and CX, which is that if you have a great theory of your relationship, you have a good underlying foundation for who you wanna be at scale and what you want that twenty four hour person to say and sound like.
You know, AI just allows you to do get to that quality faster. It allows you to get to that relationship building faster.
But we’re definitely seeing a trend where the fear’s kind of going down about a lot of those things. Because a lot of these compliance and governance things, those those kinks are being ironed out and have been ironed out over the past few years.
And I guess to that point, part of it was with the advent and introduction of Chat2BT, you know, three years ago. And just like them kind of helping to open the floodgates and get consumers comfortable with interfacing with AI and getting enterprises to say, okay.
You know? It it’s time. Like, almost like the, the tail has wagged the dog in that sense.
But.
Dom Nicastro
10:22 – 10:23
Yeah.
Michelle Schroeder
10:23 – 10:40
do you still see CX leaders in an experimental phase, you know, with with things like ChatChiPT and and AI? And what do you feel like is holding them back? If if compliance and governance are starting to be like, fall by the wayside, what do you think is still holding them back?
Dom Nicastro
10:40 – 10:56
Well, let me just tell one of our reports, shall I, and what they told us. The digital customer experience report.
And I actually you could Michelle, you’re you’re you’re free to roll your eyes too because when I am interviewing someone and they start bringing in their stuff, I’m like,.
Michelle Schroeder
10:56 – 10:56
No.
Dom Nicastro
10:56 – 10:57
alright.
Michelle Schroeder
10:57 – 11:00
I respect it. I will always respect the hustle.
Always.
Dom Nicastro
11:00 – 11:22
No. I did because I I think it’s important because it’s actually data that, you know, a lot of digital customer experience leaders have told us.
So this is kinda like speaking for the masses. You know? Here we go.
So number one concern in in the last year’s report, not the current one, was the same, data privacy.
Michelle Schroeder
11:22 – 11:23
Mhmm.
Dom Nicastro
11:23 – 11:38
So 58% in 2023. It went down a little, so the concern is getting less, but it’s still up.
It’s still number one. 49% now.
So data privacy, number one concern, back to back years. And this is post.
chat GPT.
Michelle Schroeder
11:38 – 11:39
Yeah.
Dom Nicastro
11:39 – 11:46
Number two, same deal, going less and less but still number two, cybersecurity problems.
Michelle Schroeder
11:46 – 11:47
Sure.
Dom Nicastro
11:47 – 12:08
49 2023, forty two percent twenty twenty four. And then I’ll just kinda go down, the list Protecting intellectual property, losing authenticity and content, copyright.
issues, like plagiarism. I know there’s lawsuits there, like New York Times and stuff.
Adhering to brand standards, that’s another,.
Michelle Schroeder
12:08 – 12:09
Yep.
Dom Nicastro
12:09 – 12:36
issue with with our audience and our CX leaders. Job loss, displacement due to automation, so there’s some fears there, lack of transparency, deep fakes, bias due to data bias and underlying models, losing employee trust, in in loyalty.
And, 3% have, no issues, no concerns whatsoever. I’d like to meet that person who.
just.
Michelle Schroeder
12:36 – 12:36
think me too.
Dom Nicastro
12:36 – 12:42
care. Throw it out.
So maybe they’re having huge wins. I’m not sure.
Michelle Schroeder
12:42 – 13:13
It’s super interesting. Like, when we actually go through sales processes with with companies that are on their way to embracing AI for automation, AI for automating customer experience, we see a lot of that data privacy and compliance and governance things as proxies.
When we actually get down to it and we ask more second order questions and, like, really, what’s going on? What’s the fear behind the fear? A lot of it is about one thing that came up lower in your list, which was about adhering to brand standards. It’s just like,.
Dom Nicastro
13:13 – 13:14
Yes.
Michelle Schroeder
13:14 – 13:45
are you is my job gonna be protected when something goes wrong and that gets broadcasted? Am is am I gonna be okay? Are you gonna be able to do this while still protecting my relationship with the customer? And that has been fairly universal. It’s it’s, like, easy to kinda put it behind the compliance, you know, front and, like, the cybersecurity front, but we haven’t heard a whole lot about AI hacks.
We have absolutely heard a lot about AI fails, you know, where where it’s just like the the bot does not do what it was supposed to do.
Dom Nicastro
13:45 – 13:46
Yeah. Well, I think.
Michelle Schroeder
13:46 – 13:46
I.
Dom Nicastro
13:46 – 13:49
on I think that’s on the that’s on the human behind it, though, I think. I mean,.
Michelle Schroeder
13:49 – 13:49
Mhmm.
Dom Nicastro
13:49 – 14:07
if you’re rolling it out and expecting it to just answer everything in the world for your customers, without any guardrails, without any, due diligence behind that system and putting smart engineers behind it to to. make sure it doesn’t go it doesn’t do that.
And I know it’s not a 100% controllable.
Michelle Schroeder
14:07 – 14:08
Yeah.
Dom Nicastro
14:08 – 14:09
That’s I think that’s on the people,.
Michelle Schroeder
14:09 – 14:10
Thank.
Dom Nicastro
14:10 – 14:17
you know, behind I mean, and. and and and you’re not gonna be able to blame, what was it? The famous one, Air Air Canada, I think,.
Michelle Schroeder
14:17 – 14:17
Yes.
Dom Nicastro
14:17 – 14:18
was,.
Michelle Schroeder
14:18 – 14:18
Yep.
Dom Nicastro
14:18 – 14:33
you know, it was Air they tried to present that in court like it was. the bot’s fault or something like that.
No. No.
It was you know, you’re speaking for the brand. I don’t.
care where it was, who it was, how it was. It’s the brand is on the hook.
Michelle Schroeder
14:33 – 14:40
I I don’t disagree. It’s the same thing about management.
It’s like You complain about somebody on the team, like, the manager takes the takes the fall.
Dom Nicastro
14:40 – 14:43
Yeah. I’ve said it’s on me a lot of times this year,.
Michelle Schroeder
14:43 – 14:43
so.
Dom Nicastro
14:43 – 14:45
That’s on me. That’s on me.
Michelle Schroeder
14:45 – 15:06
I. We have we share that.
We share that. So you guys must spend a ton of time, like, going through vendor pitches and claims and just, like, information, like, a mass amount of noise from this space.
Like, how do you parse through what’s actually meaningful, what’s actually worthwhile, what’s worth following up on?
Dom Nicastro
15:06 – 15:20
Yeah. You get out there and you talk to people.
You know? These these conferences, if, are are so, enjoyable to anyone who does, a customer conference, and you you guys must do one.
Michelle Schroeder
15:20 – 15:20
You,.
Dom Nicastro
15:20 – 15:21
Paul, yeah, get out there.
Michelle Schroeder
15:21 – 15:21
do.
Dom Nicastro
15:21 – 15:24
Yeah. It’s.
Michelle Schroeder
15:24 – 15:26
gonna invite you to the next one. That was our bad.
Dom Nicastro
15:26 – 15:43
yeah. No.
No. No.
It’s all all good. We we we actually had conferences, and as much as we think that they’re coming to to go to our keynotes, You know? They’re coming to go to this session, that session, and they’re all important, and they’re all valuable.
They’re coming there to talk to each other,.
Michelle Schroeder
15:43 – 15:44
Yeah.
Dom Nicastro
15:44 – 16:12
customers to customer. And that’s where I get all the magic.
So so I I you know, when I hear claims from vendors and, you know, oh, 85% of such and such is, you know, I’ll I’ll try to if I can’t really corroborate it, I’ll I’ll I’ll leave it out. Right? But if I if I have corroborated it by talking to a lot of folks and getting a consensus that it makes sense, yeah, I’ll run with it.
Michelle Schroeder
16:12 – 16:13
Yep. Love it.
Dom Nicastro
16:13 – 16:58
You know, and and and and the one thing I am really skeptical of is a company that’s, privately a private company that’s talking about financials. I’m like, no.
It’s gotta be public. I will never print, like, a company’s unless they talk about it openly with me and stuff.
But you know? So, yeah, it the vendor claims, I think I think an audience that you’re trying to attract, you know, the CX leader, when they’re getting messages from a company like yours, I think they’re doing the same thing. I think they’re cooperating claims with other customers.
Michelle Schroeder
16:58 – 17:03
That’s what? I was gonna ask. How are they getting your information that.
they can on the stage? Like.
Dom Nicastro
17:03 – 17:31
Yeah. I think those lunch those lunches at conferences are so important.
You know? They’re just talking, letting it loose. Their boss isn’t with them.
They can say whatever they want. They’re free.
And I think, the the more that vendors like you, the more anybody trying to push some kind of agenda can put more customers in front of folks that are actually using the tool, doing the work. I think that filters out a lot of the nonsense.
You know?
Michelle Schroeder
17:31 – 17:32
Yeah.
Dom Nicastro
17:32 – 17:53
When I get a pitch from a a vendor and it’s just something that’s a product, that’s fine. That has value and it’s a nice product update.
We’ll cover it. But even better, if there’s a customer in there, I always say, can we pull out that customer and talk about high level CX contact center trends? You know?
Michelle Schroeder
17:53 – 17:53
Yep.
Dom Nicastro
17:53 – 18:21
Because that’s where the real awesome, like, fun, genuine conversations are happening. So so that’s my little little advice to you.
Get the customers out there front and center. And that’s not that’s not easy.
That’s not easy. If, if the if the vendor of the CMS that I use said, Dom, could you do a huge case study with us? I’m like, I don’t have time.
I you know? But it’s hard. It’s hard to get those customers in in front of, your audience, but definitely smart.
Michelle Schroeder
18:21 – 19:01
No doubt. That said, you know, a couple months ago, we had our customer advisory board in person in San Diego, and I was sitting next to somebody who said, hey.
This person from, you know, x company reached out. Should I take the call? And I was like, x company.
They are we’re selling to we’re trying to sell to them right now. They’re calling you for a reference.
Yes. You should take that call.
And it happens all the time. So to your point, you know, the best way to parse through the sea of noise that is AI for CX is to talk to actual people who’ve deployed.
it before and can. give you an honest, you know, take on what it was like, what it really felt like to deploy.
Not the marketing language, not the messaging that, that, you know, my team’s writing, but, like, what, is.
Dom Nicastro
19:01 – 19:01
have?
Michelle Schroeder
19:01 – 19:02
actually happening.
Dom Nicastro
19:02 – 19:21
Yeah. There’s value for that, and it needs to be out there.
It needs to be a story. It needs to be told.
You don’t always have to have a customer in there. But, man, it it it it’s really good when you can do that and hearing from the folks.
And then and then even even journalists get even skeptical on that sometimes too. Like, oh,.
Michelle Schroeder
19:21 – 19:21
Yeah.
Dom Nicastro
19:21 – 19:26
they’re only putting their top three out there. I wanna talk to the ones who had trouble.
Michelle Schroeder
19:26 – 19:27
True. Yeah.
Dom Nicastro
19:27 – 19:54
And you too. And and I think vendor pitches, the more the more you can listen and and not tout your services and, like, what can we do for you? Like, how how is your unique case how how can we support your unique case? And one of the most honest things is like, you know what? We might not be the best fit.
You’re not gonna bring that back to, like, a board meeting and say, yeah. We’re not the best fit for anybody.
But it’s sometimes a good way to to be honest and maybe open up a door later. Yeah.
Michelle Schroeder
19:54 – 20:22
Absolutely. Or get better if it’s something that you do want to handle in the future.
But getting that honest feedback, you know, shapes shapes a business, shapes a brand. We say the same thing about our advisory board.
It’s kind of like how we opened up the conversation, which is like, we’re not here to make a commercial, and we’re here to make a roadmap look better. We’re here to, like, make it make a great experience for you and for your customers.
So, going back to your experience, you you’ve probably been to more CX events than anybody on the. planet.
Dom Nicastro
20:22 – 20:23
Yeah.
Michelle Schroeder
20:23 – 20:28
What. are the things that you do hear as people are walking off the stage? I’m like, what are they saying to you?
Dom Nicastro
20:28 – 21:18
A lot of stuff. You know, I’ll I’ll give you one example.
There was a health care industry leader, like a like a like an engineer software developer type, basically, like a marketing technologist, you know, somewhere someone who sits in between, like, the CX front facing folks and the the back end team. And one of the things, you know, there was a demo and a keynote from a provider with AI in implementations into the the digital customer experience stack, and it looked impressive.
You know, there’s 200 agents, you know, which, by the way, I wanna I wanna know, and I I might come after you guys.
Michelle Schroeder
21:18 – 21:18
Please.
Dom Nicastro
21:18 – 21:31
Like, I wanna know I wanna know the adoption that’s happening with. AI tools.
I want I wanna know as a journalist for CMS, why I wanna know where that your, your customers are using the tools most the most.
Michelle Schroeder
21:31 – 21:31
Mhmm.
Dom Nicastro
21:31 – 22:05
You know? Vendors are throwing out a lot of things like agents, agents, agents. And there’s an agent for this.
There’s an agent for that. Okay.
What’s the number one use case for the agent? How many customers have adopted? Is it, like, 80%? Is it 8%? Is it zero? I would love to do that sort of investigative, like, report and get that out there. Because I think we need to know where customer.
experience leaders are mostly using and I can give you some stats that we have, but I wanna know what’s happening at, like, 10 vendors. You know? Stay tuned.
Be ready to be transparent, Michelle.
Michelle Schroeder
22:05 – 22:05
We’re we’re.
Dom Nicastro
22:05 – 22:06
the.
Michelle Schroeder
22:06 – 22:10
Yeah. I mean, I can tell you a couple things right off the bat, but just like,.
Dom Nicastro
22:10 – 22:10
Okay.
Michelle Schroeder
22:10 – 22:52
you know, table. stakes are most of our customers are automating the low level, low lying things that their reps don’t wanna do.
Like, their their agents don’t wanna do. Like, a lot of them do wanna actually handle the hyper complex, you know, I I made this order.
I sent it to the wrong address, you know, and yet, actually, it’s a gift, so I need it to do this and, you know, possibly be sent over there. And there.
might be customs. Those types of calls where where they you really need to get a person on the phone and solve them.
Though, we do solve those too, but everybody starts. Like, everybody wants to kind of handle those low level kind of FAQ calls just like, hey.
You know? Can I bring my service dog to this to this hotel or to this restaurant?
Dom Nicastro
22:52 – 22:52
Yeah.
Michelle Schroeder
22:52 – 23:58
You know, can you make a note on my reservation that just puts a little bowl of water on on the ground for them or or something like that? And I’m I’m always saying to the team, you know, where things get really interesting for us is when people when you do that low lying kind of use case really, really well and you start to get to do even more interesting things, and you start to be able to handle, like, not just your run of the mill, you know, like booking a hotel conversation or it’s checking your insurance conversation. But, I have this specific problem, and I’m not exactly sure what to do with it or who to go about it or what even kind of specialist I should be talking to in your in your, you know, hospital practice.
Can you help me? When we start to be able to move from those sort of low lying when literally 100% of our our customers do FAQ calls, 100%, that’s when you start to earn the trust and earn the right to do the more interesting things where sometimes you have the more interesting conversations and you get back the more interesting data about what your customer wants from you and like what they’re really worried about and all that stuff. And I’m sure if you talk to 10 other kind of conversational AI vendors, they’d all say, yeah, we do FAQs,.
Dom Nicastro
23:58 – 23:59
Yep.
Michelle Schroeder
23:59 – 24:23
booking and, and, you know, appointment cancellations and changes, and where’s my order? Like, everybody’s got the same sort of baseline kind of use cases that they do and the journey, where they start and where they end. But where things get really.
interesting, where we can make some interesting recommendations are on those fringes. But, yeah, that’s just like that that we definitely should do a part two, and and get through some of.
those stats because there’s a lot of interesting things. on those fringes.
Dom Nicastro
24:23 – 24:44
I would love I would love, as a consumer, as someone on the other side looking at the technology from behind the scenes as a as a journalist in the space, I would love to see companies move to a place where you have a chatbot and it the first question is, do you wanna deal with a human or an AI. person? That’s it.
Just right. off the bat.
And I’ll be.
Michelle Schroeder
24:44 – 24:45
know?
Dom Nicastro
24:45 – 24:46
like, oop, human.
Michelle Schroeder
24:46 – 25:21
Yeah. You know, you talk about overcoming trust.
The consumer has a preference too. And, like, you take away their agency entirely and just push them towards something.
The idea is that we eventually build a preference for the automated version, like the agent, because it’s going to have way more data, way more speed, way more access, you know, to the things that you need to get what you want. But that might be a gradual process and easing people in.
I I think about that kind of thing too, and some of our customers. allow that.
I’m thinking about the same person I sat next to at the customer advisory board brought up that point, and has a deployment that does open up in that way. So.
very, very cool point.
Dom Nicastro
25:21 – 25:27
That’s how I started a converse. I was like, human, you know, human.
Michelle Schroeder
25:27 – 25:27
Yeah.
Dom Nicastro
25:27 – 25:36
That that’s it. That’s how I start because I know it’s just not gonna work.
I just but maybe I’m maybe I’m too skeptical there, and maybe it will work. Maybe I should try, but,.
Michelle Schroeder
25:36 – 25:36
Unfortunately,.
Dom Nicastro
25:36 – 25:37
just just to.
Michelle Schroeder
25:37 – 25:41
have to start this conversation in that way, which. is.
good.
Dom Nicastro
25:41 – 26:27
Yeah. And and just to give you the illustration of that that one practitioner that I think spoke for a lot of, I think, spoke for a lot of CX leaders.
So he got the the demo was done, and it looked great and everything. Agents, AgenTic AI, marketing, CX, it’s gonna integrate right into the your tools.
There was one word that turned him off, and it was free. So they didn’t have to, the the vendor was just rolling it out, seamlessly without any additional licensing or anything like that.
And that concerned him because, you know, he told me if the model is unlimited and free, where where’s the downgrade happening? Like, how’s how is this going to be in terms of quality?
Michelle Schroeder
26:27 – 26:28
Yeah.
Dom Nicastro
26:28 – 26:47
So I think folks like him are asking these really hard questions to vendors these days. You could you could probably tell me more about, like, token limits, model quality, data retention, even, like, vendor financial, sustainability with with AI.
Like, how is how is the vendor doing with its own spending?
Michelle Schroeder
26:47 – 26:49
Totally. Yeah.
Dom Nicastro
26:49 – 26:55
Yeah. Yeah.
Like, low cost AI often means low quality reasoning. You know?
Michelle Schroeder
26:55 – 26:57
Percent or load load yeah.
Dom Nicastro
26:57 – 26:57
they don’t.
Michelle Schroeder
26:57 – 26:57
Yeah.
Dom Nicastro
26:57 – 27:48
use it. He’s not they don’t they don’t use it yet.
So there’s a lot of, you know, like, splash on the stage, and then you get on the Ground Floor, and you’re like, it looks it looks okay, but we we just need the blocking and tackling from our tool right now. You know? We we know we don’t need this polish and and and and and fancy stuff at the moment.
And I’m not speaking. for everybody, because you do see great use cases, with CX and AI on stage, and you hear about it.
And we have a whole TV series where folks are talking about it. So, it’s it’s it’s there.
And we’re we’re we’re so early. It it feels like, you know, you if you’d ask me what has a bigger impact on journalism so far, the Internet or AI, it is 1000000% the Internet.
AI hasn’t changed the journalism game. The Internet did.
Michelle Schroeder
27:48 – 27:50
Sure. That’s.
Dom Nicastro
27:50 – 28:08
know? That was. true change, and I I lived it.
We were excited, Michelle, when we could post a PDF onto our website. Like, wow.
You know? And we put these horrifically long, ugly URLs in the print edition of the newspaper.
Michelle Schroeder
28:08 – 28:08
Well,.
Dom Nicastro
28:08 – 28:21
that told folks to go. If you wanna read all the city council minute meetings, go to w w, and it was just insane.
Great customer, experience.
Michelle Schroeder
28:21 – 28:32
I guess, of, like, of, like, pulling out things from archives. You guys built a rag system, right, called.
Roberto? Why why’d you build it, and how do. you use it?
Dom Nicastro
28:32 – 31:10
This is my favorite part, of the AI integration into journalism to date. It’s not ChatGPT helping you with headlines.
It’s not, you know, condensing podcasts transcripts. But this this rag system that smarter people than me, more technical people than me in the company built, for the editorial teams is brilliant because I think this is where, AI meets journalism.
It’s the best use case of where AI meets journalism. And I say that because this system scans, combs, learns all of CMS wire’s historical content, and it has it.
It’s learned it. It’s mem you know, it’s there.
What journalists do often is they take a moment in time that’s happening that day, that week, and they wanna give it context and perspective. Like, how long has this been happening? What’s happened last year in this area? What’s happened five years ago in this area.
The past is all journalists are obsessed with the past. This wasn’t always the case.
That’s we always say that. You know? It’s changed.
We always wanna bring people into, like, the current day by informing them about what’s happened leading up to that. That’s.
journalism. So with this rack system, it can scan all of our historical content.
And if I need something on CX and AI, just a really broad topic,.
Michelle Schroeder
31:10 – 31:11
Mhmm.
Dom Nicastro
31:11 – 31:36
it will give me everything we’ve written about it and put it into this nice summary with hyperlinks to the actual article. I can set date parameters.
I wanna know everything that’s happened from last year this day to now. And why is that the best thing for AI in journalism is because it’s our own content.
It’s not scanning the Internet for context.
Michelle Schroeder
31:36 – 31:37
Yeah.
Dom Nicastro
31:37 – 32:50
It. is taking our brand.
We’re staying that that concern I mentioned earlier about being off brand, that’s. on brand, 100%.
And we can do I would I would hate to just go cmsya. com and start searching our archives myself.
Is this article helpful to this no. Is this article helpful? No.
The Roberta, who we call her, the rack system, does it for us. And it’s it’s made like even like a vend like, I’ll give you an example of a vendor story.
Like, if we have a vendor press release, that we feel is interesting to our readers, we’ll cover it. And if we don’t, we’re not blessed with the most amazing deep staff resources.
So we might just kinda write off the press release and not do a demo or an interview. So we’ll write off the release, but we’ll feed what’s going on to Roberta and say, give us some context about what this vendor is saying, what they’re doing.
Is this normal? Is it not normal? Is it new? Is it innovative? So all the questions that journalists always ask, I can ask those now with the help of my lovely Roberta. And amazing.
Right? And. and,.
Michelle Schroeder
32:50 – 32:50
Yeah. Mhmm.
Dom Nicastro
32:50 – 33:05
as an editor, I can use it too if I feel like a a piece of content is is thin, you know, instead of going back to the journalists like, hey. Go find we’re pressed for time.
I wanna get it out the door. Roberta.
Michelle Schroeder
33:05 – 33:05
Hello.
Dom Nicastro
33:05 – 33:17
It’s it’s been a and we’re always tweaking it, iterating on it. You know, Roberta kinda said something weird.
Can we fix that? And I will we’ll feed it to the developers and you know?
Michelle Schroeder
33:17 – 33:19
he ever won employee of the year?
Dom Nicastro
33:19 – 33:29
Yeah. She’s she’s I I mean, for someone who doesn’t complain, does everything in two seconds, like, I mean,.
Michelle Schroeder
33:29 – 33:29
love.
Dom Nicastro
33:29 – 33:29
on.
Michelle Schroeder
33:29 – 33:30
it. I mean,.
Dom Nicastro
33:30 – 33:30
She’s.
Michelle Schroeder
33:30 – 33:37
nice thing about it too is it’s not contributing to the noise, like, that we were talking about in the slop that we were talking about. It’s AI that’s actually.
Dom Nicastro
33:37 – 33:37
yeah.
Michelle Schroeder
33:37 – 33:44
practical and. built off of stuff that’s already been vetted by journalists that have been thinking about the the quality piece.
I love that.
Dom Nicastro
33:44 – 33:48
Yeah. It’s us.
It’s it’s us. It’s not slop.
upon slop upon slop. Yep.
Michelle Schroeder
33:48 – 34:13
Exactly. I have one I’m gonna leave us with one dumb.
question I’m really curious about, and this is going back to fifth grade for you. So so you wrote about the high school football team in fifth grade.
Did you know about football? Were you passionate about football, or did you have to learn about it to write about it? Well, how did this happen? How did you get that gig? Fifth grade too. I mean, that’s.
Dom Nicastro
34:13 – 34:13
yeah.
Michelle Schroeder
34:13 – 34:14
impressive.
Dom Nicastro
34:14 – 34:26
In 1987, Gloucester High School football team. So Gloucester is in Massachusetts.
It’s the setting for the perfect storm movie, George Clooney, Mark. Wahlberg.
Ever seen that one?
Michelle Schroeder
34:26 – 34:26
Of course.
Dom Nicastro
34:26 – 35:35
I’ve heard of it at least. It’s about fishermen lost at sea.
You know? Very, very tragic. Gloucester is, you know, very blessed to have the fishing community and access to the water, but also cursed in a bit because we we lose fishermen, you know, out out at sea.
And it’s beautiful memorial to them at City Hall. There’s a fisherman at the wheel statue right off the coast in the harbor.
But that newsletter, I was just in love with that football team. I thought they were professionals.
You know? And they were kids, 16 year old kids. They came into the school.
I I remember all this. They came into the school, had a pep rally for us because they were going to the Super Bowl, and that was a big deal in Gloucester.
It hadn’t been done in a long decades. And I just wrote about the team.
You know? Somehow I had an inkling to I just had an itch to to write, And, that team, by the way, went to Super Bowl and broke my little fifth grade heart because they lost on a really, really, really, really bad call. Not that I’m bitter about it still.
You know? Forty years later,.
Michelle Schroeder
35:35 – 35:36
I.
Dom Nicastro
35:36 – 35:37
was a good deal.
Michelle Schroeder
35:37 – 35:37
it.
Dom Nicastro
35:37 – 35:45
It was at the suit it was at, like, the Patriots Stadium, you know, like, Sullivan Stadium that it was called back then in Foxborough, Mass.
Michelle Schroeder
35:45 – 35:46
Wow.
Dom Nicastro
35:46 – 36:01
My parents got a hotel the night before. The opposing team was in the hotel.
It was weird. Right? It was just fun.
So that’s where journalism began for Dominic Castro, right at the the hands of that, Gloucester High School football team.
Michelle Schroeder
36:01 – 36:26
Well, that’s how it sounds like you got bit by the bug, and, hopefully, there are more wins to come. It’s been great having you on.
You have a front row seat to how organizations are thinking about AI, how they’re adopting it. And hearing how you navigate that as a journalist and bring AI into your, like, workflow, was really, really interesting and really, really refreshing.
So appreciate it. Appreciate the time, Dom.
Hopefully, we get to do this again sometime.
Dom Nicastro
36:26 – 36:28
Yeah.
Michelle Schroeder
36:28 – 36:29
Yeah. Thanks so much.