Welcome back to AI in Action. I’m Eric Housh, chief growth officer here at Fast Slow Motion, joined as always by FSM director of AI, Zack Terry. Zack, how’s it going? You know, it’s good to be here. I almost just had an AI avatar join instead, but I decided to show up. So glad to be here today. We love the real life Zack.

Nothing against the avatar, but it’s great having you. Also great having a very special guest, FSM Chief Delivery Officer Jeff White. Jeff, welcome to the show. Hey, thanks for having me, guys. Looking forward to the conversation as well. Today we’re going to be diving into one of the most important, but often overlooked parts of implementing AI in your business.

The first part of the scalable business framework and that is process. So if you’re asking how do I actually start using AI in a way that sticks? This is the episode for you. So a little later, just going to walk us through a simple, clear framework for AI implementation that actually works. But first, let’s talk about what’s happening in the world of AI this week.

Nothing, right? Nothing. Yeah. It’s a slow week, right? I mean, first headline. Google just slashed LM training data by 99.5%. They’re now getting comparable results with fewer than 500 labeled examples. That’s kind of insane. And it’s insane. Yes. So typically it’s 100,000 data points, right? Like, they’ve reduced this by 10,000 times. So they use something called high fidelity labels to reduce the amount of training data needed.

I can’t pretend to know exactly what goes into that. They have kind of a whole white paper on their website that you can check out, but the key takeaway here is that it dramatically lowers the cost and time required to train models because of just what a massive 99.5% reduction in the actual data that’s required. So, I think it’s going to be interesting to see where this goes.

It means businesses could potentially build domain specific models without needing massive data sets or huge compute budgets. So, yeah, I mean, I’m excited to see what this actually means. It’s brand new. They just kind of came out with this. So I don’t think there’s been a lot of applications for it yet, but excited to see where this goes.

Probably not as impactful, but also pretty noteworthy. Oracle has officially rebranded their flagship event to Oracle AI World. What do you think about that? Well, I guess everything is still ringing true, right? Everywhere. Everyone’s adding AI somewhere and it’s no wonder with all the P and VC money going into these AI companies, it’s really common to see companies leaning into AI.

But with a brand as big as Oracle, a big legacy computer brand tech company doing this, really just positioning AI as the core of their future enterprise stack, not dissimilar to a company like Salesforce, renaming all of their World Tour events to Agentforce World Tour and as opposed to Salesforce World Tour. So I think we’ll see.

Basically, we’ve already seen most tech companies lean heavily into AI, but this trend just continues to push forward in that direction. Speaking of big money, perplexity making a $34.5 billion bid to buy Google Chrome. Yes, that’s B as in billion and Chrome as in Chrome. What is going on? There is perplexity even worth 34.5 billion. I think this is probably the most insane story of the week.

So it makes you wonder where they’re getting $34.5 billion because they’re reportedly only valued at $18 billion, at maybe 100 million to 150 million in revenue. So you’re talking 125 x multiple on their valuation, let alone the ability to fund $34.5 billion by Chrome. Now there’s ways they can do this right. Investors can come in and they can get the capital to do this.

And if they did do something like this, I mean, their whole play is essentially to replace traditional search with an AI native experience. So if they’re able to, let’s say, integrate conversational AI directly into Chrome as a feature, that’s definitely interesting. Now, I do think that this is, you know, if you’re on the side of the fence that says we’re in a bubble, this is just like a massive amount of lighter fuel that’s going on top of the fire, because this is such a crazy thing to look at when you look at the numbers.

But it’s also really showing just how big this market is and how much it’s growing and how much money is going into it. So, you know, we’ve seen crazy multiples in tech before. This is way up there, kind of with a lot of the stuff that’s happened in the dot-com era. So, I don’t know if this will actually go through a lot of sort of pieces have to fall into place for this to work out.

But it’s definitely a big swing aimed at redefining the interface layer of the internet. Yeah. Me personally, I’m a little bit of a cynic. I think it’s just a play to get some headlines and keep perplexity out there, in the conversation about AI, but regardless, it’s a signal, right? Companies are making big bets, making big changes.

And that’s why it’s so important for companies at every level to have a process for bringing AI into their own business. So with that man, I’d like to get into it. Jeff, white again here with us today. Let’s talk through it. Jeff, we’re excited to have you on. Hey, thanks. Thanks for having me. I’m looking forward to talking about AI.

Well, let’s start right here. Everybody’s hearing about tools, AI, everybody’s heard about ChatGPT. Most people have heard about Gemini, but most businesses are still asking, how do I actually use this stuff in my company? Yeah, that’s a great question. A great place to start. And, we don’t we don’t need to overcomplicate it. There’s there’s a proven formula when you’re looking at introducing technology into your business.

And so, you know, let’s let’s talk through that. The first is like you said, we want to identify a use case. And we can unpack that further. But we want to find a good place where I was going to add a lot of value into the business. And then what we want to do is pilot that use case, find out, find a small group or a, or a small section of the workflow and, and try using AI in that area.

So we want to pilot the use case and then we want to validate the output we’re getting. Are we seeing the improvement we were hoping for are the responses is the way I interacting with our business process what we would have expected? And if that works then we want to expand that. We want to further integrate it into the workflow, spread it out to the rest of the team, etc. and once that’s done, we want to continue to monitor and adjust like you guys talk about every week, a new changes coming out.

And so I as a constantly evolving technology where we’re seeing improvements more rapidly than we have with other technologies that we’re using in our businesses. And so we want to monitor and evaluate, are we still leveraging AI the best way and, and that change cycle maybe a little faster than we’re used to? Yeah, I love that last point too.

It’s it’s so important to monitor how these models are performing. And this is really true for any area of the business. We should always be monitoring what’s going on and be willing to make adjustments, because our business is always evolving them. The market’s evolving with AI. These models can drift over time as data changes. So really if you’re not reviewing, you’re ultimately regressing.

Yeah, we’re seeing that time and time again for sure. Jeff, walk us through. You came up with this really cool model called the step model, your framework for thinking about AI. And, the maturity as as it, implies. And businesses tell us what that step model is. Yeah. Happy to dig into that, Eric. A lot of times we want to over complicate the situation.

And so I try to really think about how do we think about identifying these use cases. How do I identify where I it’s going to really help us in our business. And for me personally having some labels or categories that I can go look at really helps me, as I look at AI, Fast Slow Motion. As I was thinking through this, this concept of a step model.

So let’s just let’s just high level talk through it. So, so first there’s support, then there’s think, then there’s engage and then there’s perform. So let’s start with the step. This the support aspect of AI is probably the most common use case today in businesses. It is activities like having AI help me write an email. You know, most people have asked AI to help them write something, whether that’s an email, website copy, you know, a text to their friend.

Most people have used ChatGPT or Gemini or Claude or some other tool to to ask it, hey, can you help me do this thing that’s that support model and it can add a lot of value to a business when you when you have to say, send a follow up email for collections or you do have to, you know, write a complex email or produce web copy or, you know, or a script or, you know, any anything that you might need in your business, you can say, hey, can you help build this for me?

Like, have you seen businesses using, AI in this way? Absolutely. And I think it’s easy to look at support and say, all right, that’s, simple and, hard to look at. Well, how is this really going to bring value? But, a really great example is our own marketing team at Fast Slow Motion. They are able to scale a really small team and do things that a small team would not be able to do without the assistance of AI.

So they’re able to generate so many high quality marketing assets across multiple channels, with just a select group of people that previously would have taken an entire team to do on a regular basis. So I think it’s a really good example of how something relatively simple, as far as AI goes, can really bring a big impact. Great.

That’s a that’s a great, great example that the next step in the process of stop is think. And more and more. This is where I personally find value with AI. One of our client directors, Jason Ramsey, framed it to me this way. If you had access to a mentor that had 20 or 30 years experience solving the problem you’re looking for, what would you ask them?

And so if we frame AI as a thinking partner, it changes from asking if to write an email to asking it help you identify and solve problems. There’s a good book. The AI driven leader by Jeff Woods, and the intent behind the whole book is really using AI as a strategic partner, asking it to help you identify and solve strategic problems related primarily to your job.

And so that’s really helpful to say to the AI, hey, I’m having this challenge in the business. Whatever your role is, I’m having this challenge in the business. What are some ways I might consider solving that? You know, a simple, you know, you can get you can get really complex with the problem. But but being very conversational, asking it like it might be a mentor because it can really help refine your thinking.

Yeah. A couple of kind of specific use cases come to mind there. For me personally, I love to use this to help me compare and contrast, let’s say, different tools. If we’re looking for a solution, it’s a great way to go out and save the time it would take for me to do that research on my own, but collects a bunch of different tools, identify things maybe that I would not have identified on my own, and then draw conclusions around the pros and cons.

I think that’s a great use case. That’s something I find myself using pretty often. And then one that’s really, I think, interesting and novel is one of our team members actually uses it like a focus group. So he goes into Gemini and he prompts it by asking it to simulate a focus group. So he’s got this gym that he’s saved where he can easily pull this up, and then just apply whatever context of that specific situation that he’s working on.

And it will simulate a focus group with multiple people and multiple opinions. And essentially it’s like they’re having that conversation around what’s going on, and it allows him to go and review that feedback as if an actual focus group was looking at it. Now. Yes, it’s a it’s an LLMs. So like, can it hallucinate? Can it come up with things that might not be 100% realistic for sure, but the value proposition there is really the ability to bounce ideas around and use it to help you think about things that, you know, maybe as a blind spot for you.

Right? So it’s a really easy way to go in and use these tools to really help you process through some things. It’s a great example. I love the focus group concept, and really, if we zoom out from that specific concept, the principle there is helping you get another perspective. It it provides AI provides a lot of value getting that perspective because it’s hard as humans, we’re so biased.

They obviously has biases of its own, but it at least can help you get another perspective. Another example, you know, the might be relevant for really anyone in their in the role is when you have some communication upcoming, you’re going to present an idea to your boss. You’re going to talk to the board. You know, you’re you’re writing something, putting something together for the customer.

You can describe those people on the other side of the conversation to the AI and say, help me understand how this communication will be received. You know, what weaknesses, what do I need to shore up in the presentation? Or what questions should I expect from the CEO? Those kind of use cases for AI really helps you think strategically about the work that you have in front of you.

Yeah, so it’s the concept of AI as a sounding board, right? So you have the ability to use something that’s trained on, you know, billions of data points to bounce ideas off of, which is, you know, we’ve never really been able to do that until this point in history. And so I think it’s it’s really interesting. And a great way to use these tools.

Excellent. So after we move from start to think the next element of the setup program is engage. And engage has elements of support and think built into it. But the idea with engage is that we’re now starting to introduce AI into the workflow. And this is a really exciting, exciting use case because you can start to, you know, get efficiencies and the ROI is a little easier to see.

It’s not actually measure. And so, Zack, I know we’ve done some work for clients or, you know, you’ve actually built some prototypes yourself. You know, how do you see businesses using this engage aspect where they’re where they’re pulling AI into the workflow? The human is still very involved, but it’s creating some efficiencies for the team members involved. I’ll give you a simple but powerful example, and this is something we’ve implemented with clients.

And this is specific to Salesforce, an Agentforce. But this concept is applicable across pretty much any AI tool. And that’s using prompt templates. So this is creating a template that allows you to take information from other sources and inject that into the LM to produce an output. And so an example here is a service company. You’re using cases in Salesforce.

You’ve got, you know customer support tickets and things like that. And maybe you have a lot of different types of customer support categories that can come in. Well, we can take a prompt template and we can say, okay, your job as a customer service analyst is to go in and try to figure out the results of a case based on the subject and description.

Right. We always get a subject and description. Someone sends an email and we know what the subject of the email is, the descriptions, whatever is inside the body of that email. And we can tell the LM all right. Based on these rules, I have certain business rules if you see this word or if you see this sort of sentiment or if this type of response comes in, you need to categorize it this way, and you basically just kind of write out in plain language, just like you would tell an intern or a new hire, hey, this is kind of how we think about this, right?

This is our thought process. And that prompt is running essentially every time a new case comes in, and it’s using those business rules to analyze the content of the message and then automatically categorize that case. Now, where the human in the loop kind of comes in here kind of twofold. One is we are not having the LM or the system automatically update these records.

This is being surfaced to the customer service rep as essentially a shortcut. So they’re able to look at the LM predicted response on what type of category this case should be. And they can accept it as is, and that’ll save it to the record. Or they can go and modify it. So if there is a need for them to say, well, you know, I see why you came to that conclusion, but actually it’s this they have the ability to go and intervene before it actually becomes a system of record in their system, which I think is really important.

The other piece is every interaction that we have within LM, and this goes beyond the prompt and into the actual workflow that’s been integrated with it. We are stamping every record that gets touched by an LM so that we can audit everything that’s going on. Right. So even if we were to automate this and not have that user intervene to where they’re having to approve essentially what that prediction is, it’s always going to stamp it with, hey, this was updated by AI and I can go and I can report on that, I can have a subscription, go out to a customer service rep every day that says, hey, here’s everything that was updated by AI.

Let’s do a quick, you know, cross-reference and make sure you agree with that. But the time savings there, it doesn’t seem huge. But if you’re if you’re talking 30s to a minute on every case and you’re a service team where maybe your service reps are dealing with, you know, two, 300 cases a day, that compounds over time to create a lot of, savings in terms of both time and money.

Because time is money, right? And it’s also something that people don’t like. Do it. It’s it’s sort of a pain to have to go through and figure this out. It’d be great if customers could just categorizes themselves. Right. So this both reduces time. And it is a pain point for actual employees, which is a big part of adopting tools like this.

Right? If it makes my life harder, I’m not going to use it. But if it makes my life easier, whether that is tangible in terms of the amount of time it’s saving, or if it’s just a mental block where it’s it’s saving me from the cognitive load of having to do this 200 times a day. That’s a huge when you think about time savings.

And I immediately think about the account summary. Right. The ability to combine emails, assets, cases, even field history, to prep a wrap before a call that used to take several minutes or an hour to kind of go through and see what the status of an account is. Now, an agent can give you their estimation of that.

Just in a few seconds. Yeah, you can you can do it in an automated way to have they hit my inbox, you know, 15 minutes before the meeting. You know, the, the other, use case that, that I’ve seen, it was a prototype you put together, Zack, where you took a picture of a receipt or invoice and used AI to help process, categorize, add the records in the CRM.

And again, it’s one of those pain points. It’s also hard to get right for, you know, it’s a lot of typing and I get the numbers right and I just reduces the load on the, on the human. So as part of engage they are is not acting autonomously all by itself doing a bunch of stuff. But it is saving a significant amount of time.

And, and usually, is adding value in these pain points that are important for the business, but don’t add a lot of value to the client. So as a business, we got to figure out how to classify a case. We got to figure out how to prep for these calls. We got to figure out how to process receipts and track our expenses.

But it doesn’t necessarily add a lot of client value. And so I can really help reduce the load, on those activities that this the perfect type of use case to dip your toes into I. Because if I’m creating expense reports and I is going to do it for me, there’s no way I’m not going to review it because I want to get paid right, and I want to get paid the amount that I’m owed.

And so it’s a great way to get this type of a workflow into the organization, while ensuring that you’re going to have the necessary human oversight, because employees are not going to not check their expense report. Right. They’re always going to check it for accuracy. So it’s a really good sort of, use case to, to add to a business.

That’s a great point. I love the sort of incentivized, review there. That’s a great point in tech. And it’s, it speaks to, you know, this type of processing, say, of a receipt or invoice has been possible with technology for a while. I just reduced reduces the time to value for you to put in a solution that works for your business.

And so it makes something that was really complex, require a lot of development into something that’s very manageable for a business to take on. It’s absolutely I mean, that’s that’s something that takes a few hours to set up a workflow in Salesforce that can do that beforehand. You would have had to have something like optical character recognition, you know, an OCR tool that could go out, analyze a document and pull it out.

You would probably have to define those mappings individually. If you got variations in those documents, it would be subject to breaking. Not to say that it isn’t still subject to breaking. There’s always room for error, right? But the ability to just process natural language and process images through these multimodal models, it just lowers the barrier to entry.

It lowers the cost to implement a tool like that, which I think is a really compelling, reason why you might want to look into dipping your toes into AI solutions. Exactly. Once. Once we engage, once we’re thinking through the use cases, really to engage. The next area that I think about AI is perform. So it’s the P and the step model.

And this is where AI is performing activities on your behalf. You’ll hear a lot of conversation in the AI space about a genetic use cases or a genetic AI and the fullness of that phrase really means that AI is working on your behalf and more and more becomes autonomous in its activity. So there is less human in the loop or no human in the loop.

So in the engage model, we everything we talked about it was improving the activities. But the human is still very involved really by the nature of the work that’s being done. And that a genetic model allows the AI to go and run a little further or a little longer on its own. And so I think about those in that perform category.

Zack, where have you seen clients or businesses really start to look at leveraging perform? Yeah, we’ve seen some use cases with, let’s say, an SDR type of agent where you’re getting inbound leads or you’re you’re loading in leads from, let’s say, a trade show or something like that. And you want to have the ability to automatically reach out to them and manage that conversation.

So the assumption here is you’re starting with high quality leads there past the marketing qualified stage. They’re ready to engage with sales. And you’re starting at the top of that sales funnel. And you’re trying to move them into the pipeline. This agent can do the role, a traditional SDR, which is initiating that outreach, crafting emails that are specific to that lead in the situation so it can incorporate information, probably using prompt templates.

Right, but also information that you might have on that record itself. You can craft that email. It can automatically reach out. It can connect to your knowledge base or, unstructured data for things like, well, what products or services do we sell and how do we sell them, and what types of questions can we anticipate and answer for people throughout the sales process.

And so when people respond back to the agent, it’s able to go and retrieve answers from that unstructured data from that knowledge base, which historically that’s that’s somebody who’s got to go and find those answers and craft that email. Well, now you have the ability to have I just autonomy kind of hooked up to the data that, you know, to be good.

Right. It’s your business data. You’re providing it that data. This is not just searching the open internet, right. It’s sort of a closed system and is able to answer those questions. And then ultimately it can book a meeting with a human sales rep once it gets to the right point in that sales process. And so that is a it’s a simple example, but you think about kind of all the moving pieces there, which is I’m identifying the right people to reach out to.

I’m crafting an email that is contextual and pulling in the right information about the lead. I’m sending that out automatically. I’m handling responses. I’m answering questions from an unstructured base of knowledge, or from different files that we might have in the system. That is specific to my business process, and then I know when I can send a signal to book a meeting with a sales rep, and then I handle the actual booking of that meeting, and then ultimately, it arrives in a salesperson’s inbox and they go and they have that meeting and they move it into pipelines.

It’s kind of seems simple, but there’s a lot of moving pieces and ultimately that’s sort of an entire process that’s been offloaded to I, and so I think it’s just a good example of how you can use this sort of perform step. And that’s something that’s also pretty out of the box. With a tool like Salesforce, there’s a there’s a lot of platforms that do things like that.

But Salesforce has essentially a ready made agent that you can enable and turn on. Just give it access to some of your information, tell it what you want to do, and you can start reaching out to leads really quickly. That concept of using AI to communicate, I think, is going to continue to expand. You know, we’re already seeing it.

Like you said, in the SDR, there’s a customer service use case where it might it might take that initial response to an email for support from a client and then escalate if the client requires more help. And so it deflects that case work for your service team. So there’s that communication you can see possibly in say, the accounting department where previously to build out an automated collection process.

It has to be very structured and rigid with all the data points previously defined. And now I can generate some of this communication on its own with, few guardrails, but less constraints, and less, less, actual dictation of the full process. So using AI to communicate on behalf of your business, it makes me a little nervous.

Ask, is that is that just me or where are we? What are we seeing in the real world right now? Yeah, I think it makes me nervous too. And I think anyone that runs a business should rightfully have some skepticism when it comes to these tools, but I think that probably shouldn’t stop us from experimenting and seeing what these things can do.

Right. So within that sort of SDR use case, you’re putting a lot of trust into an AI system. Which brings up a couple things. One, you need to be using a system that you trust that I think, is one reason why tools like Salesforce and HubSpot integrating AI into these trusted platforms and creating guardrails in these layers of trust around it, that’s a really compelling feature, because you don’t have to go and build all of that yourself.

You don’t have to think about how to, you know, create all of that security. It kind of comes with the package right now. That doesn’t mean you don’t have to pay attention to things like your permissions and everything inside of the system. Absolutely you do. It’s ultimately on you to make sure that these tools can only access the data that they’re supposed to access.

That’s where we run into problems. That’s where we’ve seen some things in the news recently where, you know, I saw someone hacked a third party system that was connected into Salesforce by using prompt injection to override the system prompts and just dump a bunch of information from the CRM. So that’s why it’s really important to not give these tools access to things that they’re not supposed to.

Just like you wouldn’t give an employee access to data that they’re not supposed to be able to see. Same things with these AI agents. You treat them just like those users in the system. You don’t give them access to data they don’t need. So an example of that is if you have an agent that is running on the service side of the house, do they need to know everything about the customer and about the account and about the case?

Probably not. They probably need access to maybe 5 to 10 fields on that case. Maybe they need to just know who the customer is. Their email, and their first name and last name. And they may not need much more than that, right, in order to do their job. Now, yes, we’re giving them some of that information. So we have to make sure that’s necessary to do the job.

But it is very important that we treat these AIS just like we would a human user, because they ultimately have access to whatever we give them, just like a human would. So those same risks still apply, but then they get compounded because it’s happening behind the scenes, right? It’s happening autonomously. So it is really important to make sure that, we’re paying attention to security there.

And then this kind of leads into another question, when we’re talking about performing in this sort of four step process, I imagine that requires, you know, if you’re if you’re really going into a mature implementation here, you’re going to be bringing in things like APIs, probably using CP to connect to multiple systems. So can you unpack a little bit about where connections to other systems, APIs, things like MCP kind of come into play?

Yeah, that’s a great that’s a great point to get the most value in these autonomous situations. And really some of the engage as well is you want access to your business data. And it’s a it’s a similar challenge you have today with your business systems. As the more data in one place, the easier it is. But that doesn’t mean that if you have data in multiple places, you still do need to solve for that.

And so you can leverage API to pull data in to a system, a central system, and then process it with AI. But you can also make API calls to these LLMs, and ask the AI to do custom things for you. And so you really have a lot of options to move the data around to get value out of it for your business.

So you want to make sure that whatever, whatever perform process you’re trying to set up has access to the right data. So you may use an API to bring it in, but then it may be that where the tool needs to live. And in your business, that tool hasn’t built out that a the AI that you need to get your process done.

Well, it’s possible to build a custom tool and access the alarms through an API and get some perform type value out of the AI use case. It it. Also, as we talked about this communication element, really reminded me when we opened this discussion with around the lifecycle of an AI project. And there were two steps in there of not very many steps.

Two sets. One was to validate the output. So you want to make sure you’re running several scenarios through and the output you’re getting, because it will be different every time, but it’s within the range of acceptable. And then you want to continue to monitor and analyze. So as we look at these perform use cases, those two steps are really critical in the lifecycle of an AI project.

All right Jeff, as we wrap it up today we’ve talked about the AI project lifecycle. We’ve talked about your step model for thinking about AI maturity. Just one last question. Before businesses get started at all with AI, what’s the one thing they should have in place before they’re trying to implement and scale this new tech? Yeah, this is this is not an exciting topic, but it’s a very important one for businesses.

You really want to make sure as a business you’ve put together an AI use policy. And then once you’ve put it together, you regularly review it. Again. We talked about the technology and it’s changing. The models are changing. What might be acceptable for your business today may evolve over the next 6 to 12 months, but it is important.

Your team needs clarity. They’re likely using AI for support and think today. So that doesn’t require, enterprise resources. They don’t. They don’t need a development team or access to administrator rights. They can use AI today for support and think. And if you don’t have an AI use policy, they’re making it up on their own. What what they are allowed to do, which tools they’re allowed to use.

So just a few high level, things elements to include in this I use policy is number one, what tools are approved. So in your business today, what can your team utilize AI for and then what data is safe to use within those tools. And then third win as a human required to review and have that today for your business.

Maybe every time. You know we don’t automate anything with AI and that’s an acceptable stance today. It will likely evolve over the next, say, five years, but it may be an acceptable stands for your business today. This is actually a great use for the think part of AI. I’m a business leader, I am challenged. My team is using these tools this way.

With AI, I really need to put together an AI policy so it’s got some support. It’s going to write the policy for you, but it can help you identify the challenges with those use cases and things you may need to be thinking about. You know, Zack, you were really, wise in how to narrow the fields for, for agent, like, access to your system, you know, so what is acceptable from a tool?

Data and then human involvement. Those are the three elements. And I just encourage business leaders, if you don’t have a use policy or you haven’t reviewed it in the last 3 to 6 months to pull up your policy and ask your team, you know, are they are they using this policy? Does it still apply for the business or does it need to iterate?

Well, Jeff, this has been a great conversation. Thank you so much for joining us today and for all the insights that you had. If you’re just getting started, don’t worry about things like having to build your own AI model. You can just start with the process. So use this step framework to help get started. Yeah, Jeff, thanks so much for being on the show.

Zack. Thanks as always. If you’re out there listening, if you’re serious about getting AI right from that internal alignment piece to actually implementing solutions that move the needle, we’d love to help reach out to this fastslowmotion.com/ai. Really easy way to connect with our team for a conversation.