AI in Action is brought to you by Fast Slow Motion. Our team helps growing businesses AI to work with practical, scalable solutions. To learn more about how we can help you implement AI in your business, visit fastslowmotion.com.

Hello and welcome back to AI in action. I’m Eric Housh with Fast Slow Motion joined as always by our director of AI, the one and only Zack Terry and FSM newly promoted Chief Delivery Officer, Brandon Landes. Thanks for joining us, guys. So good to be here Brandon. So glad you’re here man. Happy to be here. Happy to be here.

Brant has been with FSM for several years leading delivery client success enterprise implementations, now overseeing our on the ground delivery team to ensure we’re bringing these platforms to life in the right way. Across Salesforce and HubSpot platforms. Today, we’re diving in to HubSpot’s Data Hub. This is the data foundation that powers accurate AI, cleaner CRM operations, smarter, more predictable growth.

But you know how we do things before we jump into the geeky stuff, let’s talk about everything that’s been going on in AI this week. Zach, OpenAI declaring code red. What does that even mean? Well, I just got to come in on the geeky stuff. Like, I think this is all pretty geeky until I podcast about. Yeah, yeah, yeah, it’s geeky around, but yes, of course, OpenAI Sam Altman last week is is quoted as declaring a code red.

And this is coming directly on the heels of Google releasing Gemini three and Claude releasing opus 4.5. And those are great models. They’re so good, in fact, that Sam Altman is declaring a code red, basically saying, hey, we’re going to pause everything that isn’t hyper focused on the next model. So they’ve announced that they’re pausing advertising plans, which of course was going to be a big revenue stream for them.

I think a lot of us as users are glad that they’re pausing this. Hopefully it’s an indefinite pause, but, they’re working on a new model that they’ve codenamed Garlic, which is kind of fun. That’s apparently a successor to an earlier project called Shallot Peat, which, you know, these codenames are where to where do you guys come up with this map?

Yeah, I don’t know, man. But, you know, apparently Garlic will be the next sort of big OpenAI model that’s coming out. So, as always, I mean, early tests are, you know, apparently suggesting that it can help them reclaim the top spot, but I just kind of see that as the cycle of continuous foundational model improvement.

So as soon as one model makes an improvement, then the next group has to come out and make another improvement. And so I think this is good. This is healthy competition. This is having, you know, multiple companies that are out there creating these models and making improvements to them. And so, I think it’s it’s interesting to see that OpenAI is publicly talking about how much they are shifting their focus to make sure that they can go and reclaim that top spot.

It is an all out arms race. It feels like and good for the consumers, right? We’re getting better models. We’re getting more consistent results. Let’s talk Cod’s been in the news. Pretty concerning about, cod being used in cyber espionage. What can you tell us about? I think it’s important to remember that that these are tools.

And so, like, with any tool, you can use them for their intended purpose, or you could subvert them and use them for nefarious purposes. And so that’s what we’re seeing here. So a Chinese lead group, a state linked group apparently quote unquote, misused Claude code to help create attacks that can bypass guardrails inside of Claude and so using, you know, different techniques to essentially get it to do what you want it to do, even though it has guardrails designed to spot those type of activities and prevent them, they were able to circumvent those guardrails and use Claude to automate 80 to 90% of the reconnaissance of creating this exploit.

Development, and they were able to succeed with several breaches before these accounts were shut down. So Claude did eventually find these accounts, and they did shut them down. But it is just sort of another story indicating how these tools can be used for, bad actors to accelerate their ability to cause harm. And, you know, specifically when we’re talking about generative AI, it’s it’s able to accelerate both the, offensive.

So people actually going out and try and cause harm, but then also the ability to, to catch those people. And so we’re seeing not only attackers using these tools, but then of course, people defending against it, using these tools to help spot and shut that down as well. So just another development and another reason to stay security conscious in this world of very rapidly developing technology.

Absolutely. You don’t really think of the FDA as being an AI forward organization, but the FDA this, this recent week giving all of its agency employees access to generative AI. What can you tell us about the FDA did decide to step into the AI era, so they are deploying an agent AI tools. I’m not sure which one.

I don’t know if this is homegrown or if it’s connected to a frontier model, but they are giving all of their agency employees access to generative AI to help them with planning, reasoning, execution of multi-step tasks. So a lot of the stuff that a lot of us have been using it for over the past couple of years, but it’s it’s great to see this going into a, a big government entity to take advantage of, hopefully some efficiencies that these tools can bring.

The commissioner says it’s all about accelerating more cures and meaningful treatments. And so hopefully we’ll see some, you know, more rapid development, potential treatments and case studies and things like that. So, yeah, just, just another example of not just private companies, but also public entities that are adopting generative AI into their workflows. Yeah, you’ve got a lot of that.

I mean, the government embracing AI, you it was inevitable, right. And these employees are probably using it already even though they didn’t have the rubber stamp. So it’s good that they’re okay. I think as long as, as long as we can, you know, reduce bias and some of the ethical considerations around AI. I do think is a good thing.

But we do have to be careful when we have entities that are a part of the government and have power like that. We gotta make sure that hopefully it’s being used in an ethical way. Correct. A bunch of folks are early employees of Anthropic, probably getting pretty excited right now. Rumors of an IPO. What can you tell us about that?

Anthropic CEO Dario Amadi Amodei I’m not sure exactly how to pronounce it. But he was at an event called the Dealbook summit, and he made some pretty staggering claims about Anthropic. And apparently they have grown 10x year over year. And so they are on track to do reportedly $10 billion in revenue in 2025. That’s a hundred times more than just two years ago, according to their own reports.

So again, it’s it’s self-reported. So take it with a grain of salt. But that is a significant amount of growth. And they are also reportedly preparing for a 2026 IPO with valuations that could exceed $300 billion. So pretty wild to see these very early companies. I’m pretty sure that that Anthropic really came into the mainstream in 2023.

And so it’s a very recent company that has grown very, very quickly. Personally, I really enjoy using Claude. I think it is a great tool and a great model. So it’s just it’s great to see kind of how quickly this is moving and, we’ll see, I guess next year if they end up going public. Exciting stuff.

Wrapping it up. It wouldn’t be a show if we weren’t talking about Salesforce. Right. So Marc Benioff, raising some eyebrows, talking about potentially renaming Salesforce. I think if we were to go scrub some older episodes, we may have predicted this. I think a lot of I think we predicted this. Yeah. This isn’t this isn’t something unique to do.

Our own predictions. But Benioff said in an interview that a rename to Agentforce. And by a rename, I mean the company. So shifting from Salesforce to the word Agentforce, he said that that would not shock him. So to be clear, he’s he’s not making any claims that they are definitively changing the name. But, you know, imagine that you’ve got the Agentforce Tower in San Francisco instead of the Salesforce Tower.

So we’ve seen all of this moving this direction anyway from Dreamforce this year. All of the Claudes have essentially been renamed. And so instead of sales Claude service Claude it’s agent for sales or agent for service. So a lot of this is already sort of taking shape. So the last domino to fall might be the name of the company itself.

But this is also not without confusion. So imagine if you’re a new customer and you bought agent for sales, which is the historical sales Claude. Well, you may think that by purchasing that you now have access to Agentforce and some of the generative AI tools, but you don’t unless you’ve purchased Agentforce for sales. That’s incredibly confusing, especially if you’re a new customer.

And so I think Salesforce is really going to have to think through how to provide clarity on all of the different products that they are offering. If they’re going to inject the word Agentforce into every. Yeah, we’ll probably see some shifts in how those SKUs are arranged after the fiscal, and in January. So big week, mass evacuations, new threats, government rollouts, competition heating up and potential Earth changing.

Name changes with Salesforce. You come in at about the towers. Like, I’m just thinking, man, we need to be in the sign business right now. If we’re going to switch on, let’s let’s, let’s shift to something every business needs, before I can deliver value. We know is data clean, unified data. And so that’s where HubSpot’s data hub is coming in.

And that’s what we’re going to chat about today. So just start with the basics. What is data hub. So this is not dissimilar from Salesforce Data Cloud or Data 360, but it’s essentially HubSpot as a unified data layer to allow for cleaning and blending and activating customer data. So it’s consolidating capabilities from a few different products Data Sync Ops Hub, Warehouse integrations, Data Studio.

And the idea is to provide a single source of truth across all of your different hubs. So if you’re on marketing hub, sales hub, service hub, operations hub, the idea is you’ve got a single source of truth across all of those that unifies it all together. And the end result is it helps your generative AI models have the context that they need to perform their best, and to pull the most up to date, most accurate data.

And it also reduces dependencies on things like spreadsheets and other tools outside of your hubs, so that you can have all that data in one central location. So, Brendan, from when we talk about delivering excellent technological solutions to our clients, why does Data Hub matter so much? Yeah, I think I think the biggest thing is it gets you it allows you to control your data, the cleanliness of your data, the extensibility of your data.

You’ve got a single, you know, set of data that you can unify within HubSpot. Clean. That’s huge, you know, lets you not have conflicting numbers across different teams. So everybody’s running off the same, information. It really streamlines implementations because you can leverage the capabilities within Data Hub, to really manipulate the data, however you need to, to ensure it’s clean.

And that’s huge for two main things. One, adoption of the platform. Right. Making sure that your teams are have confidence within that data. But then two, if there’s no if there’s confidence in the data, you can leverage, I, you know, AI’s only going to be as good as the data that’s underlying or that it has context for, so this lets you get it clean, get a good foundation for adoption, extend to other platforms, and really get some confidence within that data to then layer AI.

Yeah. Let’s kind of double click on that just for a second. Zach. Why why is that unified data all in one place so critical for AI in a business? We’ve talked about this a lot. And and we’ll continue to talk about it, because having good data is really critical to the ability for an AI model to be able to perform well.

And that is because it’s grounded on that data. And so if your business data is scattered or it’s inaccurate or it’s incomplete, the AI is going to be grounded on that data. And so through really no fault of its own, it’s going to reflect the quality of the data that you have. And so if you want the AI to be able to reflect accurate results and meaningful results, that data needs to be really good.

And we’ve talked about, you know, you don’t have to clean everything, right. You’re not going to just from day one, use generative AI across every single functional area of the business. But if you’re going to pick a process, whichever process you’re using for AI to empower that process, it’s got to one be a clear process. We’ve got to be able to articulate it clearly and then to the data that powers it has to be trusted in a way that when you get insights out of that data and when you’re running a model on top of it, it’s going to be able to output good results.

Let’s, let’s talk a little bit about Data Studio, because a lot of this is where the power of this shows up. What can you tell us about Data Studio? Yeah, I may I may have to lean on Brandon a little bit on this one, but but my understanding is this would allow teams to blend multiple functional areas CRM, billing, product warehouse data.

If you have it connected, you don’t have to use SQL to hook everything together. So it’s is a nice declarative interface that allows you to, piece that data together. And it can enable things like, all right, I want to see my campaigns, but I want to see it broken down by revenue so that I can understand which campaigns are performing the best.

I want to look at lead fit and intents, intent scoring across my leads and how they’re engaging with our content. Customer health scores based on business specific attributes that we define so that we can understand a customer that’s in a really good state versus one that we might want to put a little bit more attention into to make sure that they’re happy and doing well.

So, it’s really allowing data sets to flow into some of the things in HubSpot that we’re familiar with, like reports, lists, and workflows. Yeah, I’ll jump in on that. I’d say the biggest thing is it reduces the reliance upon having to, you know, request information from it. Right? It really puts the power of the data in the hands of the users.

You know, the people actually in there trying to answer those questions instead of having to send out some requests so they can, you know, query the database or ask a question and get that those insights, surfaced very quickly to answer those business questions in the context that they need it, and then they can action on it.

So it’s not just the visibility, Zach mentioned. You can see it that you can report on it, but then you can also action on it with workflows and other automation within HubSpot. So extremely powerful tool. And the biggest thing, again, is you’re not waiting for somebody else to help answer the question for you. You can get the answer to it yourself.

Yeah, I think teams are going to love that right? For sure. Bringing in that data from every part of the business, making it actionable, making it visible, to the folks, on the front lines, making those decisions. And we’ve talked about this a lot, but, it’s a big pain point. The CRM is dirty data and data hub can really make an impact here.

What does that look? It really can be a few things. Monitoring duplicates is a big one. So making sure that that you don’t have a bunch of different versions of customers or companies or deals or things like that across your different hubs, changing formatting errors. And so sometimes there are objective things about data structure where we can know that an email address isn’t formatted like an email address as an example, or a phone number that isn’t formatted like a phone number.

Or maybe there’s some regional differences there, right? Or I’ve got missing fields on a majority of my records in the CRM that can flag something as well. Maybe that’s not actually needed. Or if you’re going to use that in a model, you need to go in and pay attention and fill out that data in. An AI can actually be used to kind of help do that as well, depending on, the other data that you have in the system.

And so it’s doing things like that, managing duplicates, enriching records with high quality data, looking for formatting issues, things that could cause an AI model to trip up if it doesn’t have the right data. Really just empowering teams to have those tools so to run more or less in the background so that you’re not having to manage that in your day to day.

Yeah, dirty data is one of the biggest factors. If you have trust in that data. Whether it’s duplicate management, whether it’s formatting, your adoption skyrockets. The other thing I’ll say is it really surfaces some of those data quality issues. So brings it to the forefront with you have a process where you’re leveraging data hub to monitor that quality.

It really shows where there’s issues in your process as well. So it’s not just the data itself, it’s is there a business process that we should potentially change to ensure that we’re getting quality data, that we’re not having duplicate entry, that we’re not having, you know, people manually do certain steps, to fill out fields or properties that don’t really have an impact.

So data hub can help you not only manage the quality of your data, but also the quality of your process by highlighting those issues. That’s an excellent point. Let’s talk about, market a little bit. Enterprise teams, larger teams. Warehouse connectivity is huge. Zach, walk us through what this is. And when we say warehouse, we’re referring to things like data lakes or data warehouses.

So, having a lot of data stored off platform. So this would be the ability to connect HubSpot with things like snowflake data bricks, redshift BigQuery, the ability to bring all of that enterprise data into HubSpot so that you can actually segment it and report on it, and give AI the ability to understand that context, because that’s that’s another thing.

We talk a lot about data and we talk about process, but all of that, the purpose of that is to provide really good context to an AI model so that it better understands not only the questions that we’re asking and how it relates to that data, but also the specific situation of a given customer or a given deal or given company.

So being able to connect a bunch of systems where we’ve got essentially different pieces of a customer’s story, they’re stored in different areas. I might be able to tell one story by looking at, a data warehouse or by looking at, let’s say, marketing hub, and that’s going to give me, you know, a little bit of the customer story.

But if I want to see the entire thing, being able to connect all that together for one unified context, that’s where the power really comes in. Completely agree. The more data you get in the in the end, users hands, the more they’re going to be in that platform, and the more they’re going to leverage it. Less swivel sharing between different platforms.

It also helps you with maintenance. You know, when you can integrate right there. It helps with data governance, because you’ve got control over that integration. You don’t have to do some custom middleware, or other solution. So it really, helps you grow on the platform and encourage your team to stay within that platform to be more, more efficient throughout their day.

I’d also say that it, it allows teams to operate with consistent definitions of what fields are. Right. There are there’s the technical aspect of doing some of these integrations. Then there’s how do you actually use it as an end user. This lets you reduce that complexity and operate off of the same definitions of different properties, different objects, to make it really easy for that, and absolutely shattering the silos that that tend to form within organizations.

Getting everybody on the same page. That’s the name of the game. So let’s bring it back. You know the name of the shows. AI in action. Zach. So, guys, what is what is how can we action this data within HubSpot on a day to day basis? Some of the features that you can turn on if you have it, you can use smart lists to basically build some really powerful segmentation based on unified data.

If you’ve brought in data from other sources, there’s, of course, the ability to integrate your workflow automation across all of your hubs. And so marketing sales service, imagine bringing all of those together with other unified data sources to empower automation, not just AI. But now that you’ve got the data available in one central location, you can power a lot of automation based off of that data.

You can look at things like predictive scoring for leads or for your accounts, generating AI insights based on the state of an account. And then again, it’s not just about looking at an account and generating a score or a health metric. It’s understanding that full story. So being able to bring in that unified information and it’s not it’s not just looking at connecting third party sources.

Maybe you don’t have those third party sources and you’re doing everything inside of HubSpot was the ability to see the whole story across those different hubs. So, this can apply to really anyone, even if you’re not connecting to third party sources, because the ability to have that unified understanding of all of the different hubs and the data that they provide to really weave that story together gives the AI on top of HubSpot, the ability to generate insights beyond just the record that you’re looking at.

Yeah, I think that you really start seeing the value when you can. You can tell the teams are making decisions faster, right? So whether it’s a rep, you know, that’s that’s paying attention to or surfacing the right deals, whether it’s marketing, leveraging that smart segmentation like Zach’s talking about service teams routing issues more intelligently, you can see it across the board.

And then obviously all that funnels up. Zack was alluding to that. That funnels up the leadership’s visibility into the right things in the organization. So leveraging this across the platform with AI, you really start to surface the right business problems and have people prioritizing the right things across all functions. It’s what we preach, right? It’s process first automation, visibility of the data that that you need to have to make intelligent business decisions.

Let’s just double click there for just a second. Let’s talk outcomes. We’ve got clean, unified, actionable data. What what can different teams within organizations. How can they benefit from it. Yeah I’ll take this high level first and then Brandon, I’ll I’ll kick it over to you and we can talk about maybe some specific, you know, business function use cases.

But if we take AI out of the equation here, there’s no downside to having good quality data and a single source of truth, right? If I have one source of truth, and I always know where I am pulling that information from, I trust it more. I can more reliably use the tools that I have at my disposal, and I’m not having to do as much work to go and double check that everything is correct, right?

So you get some built in benefits just by doing this. And it has the added benefit of giving you the ability to leverage generative AI inside of your organization. But even if you don’t use AI, you’re going to get benefits from creating unified, clean data, right? Also, you’ll you’ll tend to see adoption speed up a little bit more because of it’s usable.

It’s easy to use. All that data is at my fingertips without having to, you know, swivel chair. Like you said. Brandon, I like that analogy. Not having to look through multiple tabs. I’ve kind of got everything where I need it. And then now we bring AI into it. This means I’ve got trusted data, I’ve got clear processes. When I go and I analyze an account and I have it generate some insights, maybe I’m pulling in things like transcripts from customer calls, and it’s pulling in information from marketing engagement.

When I have the ability to go and generate insights using AI, I can trust it more because I know that we’ve built a solid, trusted foundation of data. And so really what that means is it’s it’s tightening up kind of every aspect of the customer journey and how we’re tracking that on top of the platform. Yeah. When you look at we’re talking customer journey, I mean just start upfront.

Start with marketing, right? We want to be talking to the right people with the right stuff at the right time. Well, we can leverage that for, you know, getting accurate segmentation, making sure we’re getting to them in the in the right portion of their customer lifecycle. We got personalization right. Smarter personalization, leveraging the AI capabilities within HubSpot.

You also want to know what’s working and what’s not. So let’s get some reliable attribution reporting on that. Then we want to have clean campaign. Let’s we don’t want to be wasting time or wasting effort or money on the people that we don’t think are a good fit for the company. Right. For our business. And we pass over to sales.

Right. Let’s get accurate forecasting. Let’s look at the right deals at the right time. Let’s do less wasted prospecting. On again, people that may not be a good fit. And let’s highlight those ones that we really want to double down. You know, in the leverage AI to help us bring that to the forefront with some true intent score, to understand where we’re at and what level of effort we should be putting in the sales.

And once we get them to a customer, we want to know all that context, right? We want to understand the 360 degree view of how is that customer interacted with our business. We want to get faster resolutions for them. We want that to be very personalized the way that we, you know, react to, whatever has come to the customer service team.

And we want to have proactive churn detection. So things like do is there some sort of indicator that I can help us, look at or flag or surface so that we can ensure we stay ahead of that churn for our business? The last one, obviously, rev ops, you know, we want to be able to view everything consistently.

We want good, consistent metrics. We want visibility into reliable dashboards across the entirety of the the customer lifecycle. We want governance like we talked about that data quality is key. So let’s get some automated, governance around the way that we’re doing things. And let’s get some stronger cross team alignment. Right. This is all in the same platform.

Let’s surface these things in those hand-off moments between the different stakeholder groups. Instead of running it in silos, we can share the AI capability across the stakeholder groups. Get a lot of power out of the platform. Well, guys, great content so far today. We always, for our listeners, like to help the rubber meet the road. So for folks out there that may have their interest piqued, that may be saying, yeah, this is a problem.

I could really benefit. I need a tool like Data Hub. In my organization, What practical steps, some of those listeners can take this week to move toward that foundation. I think the first thing is a non fun activity, which is to audit your CRM for data quality. Right. And so that is just asking the hard questions.

How much do I trust the data that we have? Is our data accurate. Can I see good information about my customers and about our processes? That’s the first thing to do. And then you can use some features to help you with that. So turn on the data quality overview so that you can get a snapshot of how your data is looking.

If you’re not using duplicate management already, it’s one of the first things that generally we recommend across any CRM, whether that’s HubSpot, Salesforce or anything else. Managing duplicates is critical to make sure that you understand that you’re looking at the right customer with the right information. If you have an external system and you have data hub, you can look into connecting an external system using data sync.

If you don’t have it, that’s totally fine. You may not get any benefit out of that if you’re already kind of looking at everything inside of HubSpot. But if you do, check out data sync and see if there’s an opportunity to connect those systems together, you can build a data set in Data Studio. And of course, make sure that you’re cleaning before you scale.

So again, it goes all the way back to the first step, which is auditing that data. Make sure you have that clean data before you attempt to scale that across the entire system, so that you can expect good outcomes and reliable AI use cases, and then you can experiment with things like AI enrichment to go out and get information about customers or leads that you have.

Maybe you don’t have all of the data, and you can rely on HubSpot’s enrichment process to go out and fill in some of the missing gaps, which leads directly into better quality data. A fuller picture of the entire story, which will in turn feed AI to give you better results and brand. And from the delivery standpoint, what’s move number one companies need to be making?

Yeah, it’s a great it’s a great question. So one we started this out saying this is the AI geeky podcast right. Isn’t that where we started here? That’s absolutely right. Well we can build the coolest solution, you know, cool CRM with tons of automation and that’s great. But it has to solve a business problem, right? We have to be showing ROI and we have to start with a business problem.

So I would start there. What’s the what’s the business problem. That is most prevalent right now that that you want to make sure you’re going through and having something to solve. Don’t start with the technology. Then work your way into how the technology can solve that business problem. So I’d find what’s the business problem? What are the data points that help me solve for that problem?

Or at least get visibility into how to solve that problem right from there to like, was Zach mentioned. We can then connect the system if it doesn’t already exist within there. Bring that data in and start small, right? Start easy. Get that quick win. Show the value. Prove the value of that use case. Get some adoption of that and show the vision and then expand upon it.

So to put it a little more succinctly, start with a business problem. Look at the data. Do you have it? Do you not? If you don’t connect it when it’s there, ensure you have visibility into solving that business problem, get the adoption and buy in of the stakeholders, and then continue to expand upon it with the other capabilities.

So putting a bow on everything today, guys. HubSpot’s data hub is the backbone of smarter growth. AI is pretty useless without clean connected trucks. Trustworthy data and data hub can be the foundation that unlocks all of. Yeah, and I would say it’s not useless without clean data, but it is hard to get good ROI in business application without clean data.

Yeah, you can always use these tools as a sounding board and and use the training data to help you iterate and do things like that. But if if you’re expecting your business to be reflected accurately in the AI responses and insights, you absolutely have to have good, trusted, clean data. And so keep that in mind. Understand and have clear processes.

Have good trusted data. Data hub as a tool can help provide a path to achieve that so it can help you unify that data, create more reliable databases, and continuously improve it with things like enrichment, duplicate management, which helps tremendously with the prerequisites that you need to have a high performing AI, right? Solve the inputs so that you can get good outputs.

Yeah. When that foundation is strong with your data quality, IT delivery is about giving value, providing value to the, stakeholders or end users that are leveraging the platform instead of having to worry about cleanup all the time. Right. That quality is key. Teams see results faster in that that AI implementation becomes more sustainable. It’s something that’ll last versus being fragile if it’s built on poor data quality.

Great points. Anybody listening? If you want help mapping how data hub fits into your business, hit us up at fastslowmotion.com/ai. Zach, Brandon this has been fun man Brandon thanks for joining us. Been great. Yeah, thanks for having me.