AI-powered Audience Simulator

 

476.Jeremy Greenberg

SPEAKERS

Will Bachman, Jeremy Greenberg

 

Will Bachman  00:02

Hello, and welcome to Unleashed. I’m your host will Bachman. And I’m delighted to be here today with Umbrex member Jeremy Greenberg, who is going to talk to us about a tool his firm has built. It’s called the AI powered audience simulator by the avenue group. Jeremy, welcome to the show.

 

Jeremy Greenberg  00:22

Well, thanks for having me, I’m really excited to be here.

 

Will Bachman  00:25

So Jeremy, I understand this is not yet available to the general public, you’re about to start a beta a closed beta, where people, listeners here, if you get excited about it, you can reach out to Jeremy to sign up. And you’re gonna walk me through the tool, I’m looking at it, the audience is not. So you can kind of give me a verbal description, blow by blow as we go through a test of the tool. But give me kind of the overview of what it does.

 

Jeremy Greenberg  00:53

Yeah, thanks. Well, so in essence, what this audience simulator does, is, it allows you to, to provide a set of custom instructions by saying, this is the type of audience I want to look at is a type of segments, I’m interested in learning more about just like you would do for regular research, for interviews or for surveys. And then once you’ve defined the who, then you can ask them questions of the what? And then so you can ask them different types of questions, qualitative questions, quantitative questions. And then when you input that information, you run it, and within a few minutes, you get results, you get actual results from simulated respondents. And really what this is doing is it’s mirroring the sentiment of collective segments and audiences, just like a chat TPT or other MLMs do on a one on one basis, you see a lot of people using it for understanding what would a celebrity, think of this situation? What does Steve Jobs think about the new iPhone? Or? Or how can I talk to even dead relatives, people are doing that. And so there’s an immense power in these MLMs. But they haven’t been harnessed in this collective way that can mirror the distributions of the underlying population. And that’s where we come in. And we allow you to do that through this tool.

 

Will Bachman  02:17

Right. So we’re basically doing market research, it’s as if we’re running a survey on humans, whether it’s b2b or b2c, and allows us to, in some cases, substitute for a survey, or potentially complement a survey, we could test some runs first, using your tool before we go to the much more expensive trial in the field with human beings, and get an initial sense of what sorts of results we might get. And so you’ll go have like a set of questions. And it’s going to simulate asking it to, you know, hundreds of people where you design the segments, and then you get the results, right, so my understanding that correctly. That’s

 

Jeremy Greenberg  03:00

exactly right. And I would say, you know, I have about 25 years of research experience. And so this is one of these, and I’m not particularly good at coming up with new ideas, or products, I’ll admit that this one really came out of a pretty clear pain point, there’s so many pain points around doing consumer or b2b or expert interview research, where you have to find the people, you have to convince them that you’re not trying to sell them something. So they’ll participate, you have to incentivize them. And then you have to conduct the information and synthesize it. So there’s a lot of pain points, particularly when you’re moving and you’re using panels or other sources, where sometimes the quality of participants is not amazing. And it can be extremely cumbersome. There’s also situations where you want to simulate individuals where they’re really, really hard to get to. So we’ve talked with clients who work in healthcare, and they want to talk with patients who are in stage four breast cancer, well, you don’t exactly set up a focus group for that, right? You want to be very sensitive about that. So to your point on complementing, maybe instead of talking to 20 people, now you can talk to five of them, after you’ve done, you know, some simulations with this. And so what we’re not saying is to use this instead of human research, we’re really saying it to use it for two main reasons. One is to pre test before you’re conducting human research, which will help you prune topics tone, the language, you know, uncover some surprise that you would normally see anyway, when you do a soft launch, except you don’t have to do the software and you can iterate before you get it live. And the second way is you’re unlocked with these infeasible research mechanisms. Things like I want feedback on on secret ideas. I have 10 Really cool product ideas. But I don’t want any consumers to know because I don’t want the word getting out there. So now I’m going to stimulate and I’m going to ask people or I’m thinking I have an m&a opportunity. But I don’t know how the market is going to react. So I can simulate not a stock market per se, but I can simulate power experts, how are employees? How do I think industry, players are going to react when we do this. So there’s, if you think about it, there’s a lot of opportunity here, because really, we’re taking, we’re mirroring the sentiment of collective groups. And if you take it further, which we’re not doing yet, but it’s on our roadmap, you can think about things around using it for jury simulations for complex scenario planning. You can even retroactively create longitudinal studies can you can place collective audiences in different parts of time throughout history. So there’s a lot of really cool stuff here. But we think the coolest and most applicable use is really from research here. Because it’s such a big pain board. And frankly, we’re a research firm. So it made a lot of sense for us to, for that to fit. So I can I can jump you through, we can block that example here as well.

 

Will Bachman  06:02

Let’s jump into it. So I’m seeing a kind of an entry screen, where it says number one, create audience. And then number two is add traits to create segments and review segment traits. So Marcus, Marcus has sample here. Yeah, so

 

Jeremy Greenberg  06:17

I just want to say to the audience, we’re in the process of updating our front end, you know, so the flow is gonna be a little bit different. But in essence, the first section is the hook. Right? So first, we create our audience, which is your global population that you’re interested in for that particular topic. So in the example we give in our in our video is Americans drink Coca Cola, you could make it as specific or general as you want, but it’s your overall population. The reason you want to set that is because afterwards, we’re going to create sub segments within that, and all the sub the sub segments will have the audience criteria. So if you say American drink Coca Cola, every segment is going to be an American who drink Coca Cola. That makes sense. So it

 

Will Bachman  07:01

gives us some other examples for maybe some b2b type surveys. Sure, so so we have others CFOs, who are currently switching from, you know, Salesforce to some other CRM system, or?

 

Jeremy Greenberg  07:15

Absolutely. So we did a demo for an international company that wanted to understand for its software, its accounting and financial related software, what are the other types of software that that small businesses are using in these different regions that could help them with affiliate and partnership relationships. So what they wanted to understand, they said, basically, decision makers who decide what types of software to use at different small businesses, and they segmented them into different countries, they looked at Australia, New Zealand, UK, in the US in this example, and they and they, and from that, they ask questions around what are the what’s the vendor ecosystem? What kinds of vendors do you use for communication tools? What do you use for your accounting tools would you use for taxes and so forth? And so that can give you pretty quickly, you know, a visual of okay, these are the different software companies that are being used by different sub segments within your audience. And then you can potentially take that and say, Okay, this is a good first kind of potential partners we could work with.

 

Will Bachman  08:22

Amazing, okay. So you create this audience, and that might be that’s like the broader audience. And then maybe it’s, you know, executives responsible for procuring uniforms. All right. Well, you make it sound so sexy. I love it. Well, I mean, that is sexy. I mean, it’s cool stuff, you know, but we,

 

Jeremy Greenberg  08:48

why don’t we put one in let’s make it whatever you want. You want to do that executives,

 

Will Bachman  08:52

executives responsible for the sourcing and procurement of uniform rental. Okay,

 

Jeremy Greenberg  09:00

so I’m typing that into the audience box. Alright. As best I can recall it. For

 

Will Bachman  09:09

uniform rental service. Okay. It’d be perfect.

 

Jeremy Greenberg  09:15

All right. So we’ve got executives responsible for the sourcing and procurement of a uniform rental service services.

 

Will Bachman  09:21

Yeah, services there. We save that.

 

Jeremy Greenberg  09:24

And then do we want to have it just in the US? Do we want it overall?

 

Will Bachman  09:27

In the US? Sure. In

 

Jeremy Greenberg  09:29

the US, okay, so nobody could save. Yeah. So that’s our audience. And now what we can do is we can create segments and the way we create segments and this will be a little bit different in our new iteration, but effectively, you come up with different characteristics to define the each segment so I’d

 

Will Bachman  09:48

like to do it maybe by industry, so I bet there’s probably executives in the food service industry, who, you know, procure uniform service.

 

Jeremy Greenberg  09:59

We don’t think So we’re repeat executives or some we can just say like work in food industry or work in food industry

 

Will Bachman  10:06

perfect.

 

Jeremy Greenberg  10:06

And then what we do is we do as follows, we can add more traits if you want. So for this Okay, so let’s just say, let’s just make it like 20 people. So we can do that pretty quickly. So we’ve just created a segment of people. And you can see our review area here of within the audience. These are those who work in the food industry. And so now we’ll do some other industries.

 

Will Bachman  10:33

Let’s try the chemical industry.

 

Jeremy Greenberg  10:38

All right, so we’ve done we’re adding a segment with one trait, which is work in chemical industry 20 For that, as well.

 

Will Bachman  10:45

And let’s do one more, maybe how about janitorial services, they probably have uniforms, janitorial services they do.

 

Jeremy Greenberg  10:52

I’ve done a bunch of work there, actually. This is going to call just to be consistent janitorial services industry. And so now what we’ve done is we’ve created our three segments. And so for those who are listening, we have a review section section. And it basically says, what are the different traits? What are the different elements that comprise each of these segments, and then we have a sample for each and the percentage base of the total. So we have our who we ever who, right now, what we want to do is we want to ask our questions. And so as of now we download a CSV and we add the questions in the new the new version, you can just type them directly in. But let’s go ahead and do the previous one now. So basically, here, we can ask, you know, any types of questions we want. We would recommend not asking questions about like politics or current events, just given the the time horizon of the model. But we can ask a lot of a lot of questions about about, you know, to this to this group, in

 

Will Bachman  12:00

terms of, you know, some of the questions might be some open ended, like what of either do you currently use or and then might be, what were your key buying criteria for

 

Jeremy Greenberg  12:17

so I like starting broad and going gluten deeper with research, generally, one thing we may want to do is kind of double check our trades, right? So you could ask them something like, what company? Do you work for? What industry or something like that? Where we can say, Okay, does it actually match? Right? Okay. So we could start pretty broad like that. We could just say, what industry? Are you in? You know? And presumably, they’ll answer correctly, but we’ll find out. Now, we know the industry. You know, we’ve got questions. You know, these are all executives responsible for sourcing, procurement of rental services, uniforms.

 

Will Bachman  12:57

And we keep what, what company do you work for? What is your title? We do that exciting?

 

Jeremy Greenberg  13:04

What company do you work for? What is your title? We could ask things like how many years you’ve been in? We could ask that. How many years? Have you been in this industry? All right. Okay, so that’s a little bit of background information, we could obviously ask, you know, any demographic information you want anything about really any questions? But let’s, right now, these are all open ended questions, right? So we can we can also add clothes into questions, we add quantitative questions, I like to do rating questions, because then you get a lot of value out of that. So one thing we could ask is something like, what maybe I would want to know is like, to what degree? Do they have a split wallet around the, you know, of the uniforms? Like, how many how many vendors are you working with might be an interesting question.

 

Will Bachman  13:58

Okay.

 

Jeremy Greenberg  13:59

And then we could then we could actually turn that into a multiple choice question that’s quantitative. So we could say, how many vendors do you have? For whatever they call it? uniform?

 

Will Bachman  14:13

Uniform rental services,

 

Jeremy Greenberg  14:15

uniform rental services, all right. And then we can say, sort of like one? Well, we could test it and say zero. If they won, say two to five and let’s say greater than five. So then that will get us some sort of distribution we could see like, you know, how much is the wallet fragmented, perhaps. Okay. And then something maybe we could say is like, what is your prime who is your primary vendor? Oh, yeah. And then let’s have them rate them. So now we have the name and we can say On a scale of one to five, where one is, you know, very unhappy, let’s say, and five is, oh, to satisfy a little bit better one, on a scale of one to five, where one is very unsatisfied, and five is very satisfied, satisfied, are you with this vendor? And then we can ask why. So we get some open ends.

 

Will Bachman  15:34

And they also want to throw in there. And maybe just one more maybe. What was your key decision criteria when you selected this vendor?

 

Jeremy Greenberg  15:47

I like that. You don’t have just one or multiple? Oh,

 

Will Bachman  15:55

yeah, sure. You could have it multiple, your top three or something? Okay. What are your top three? And you know, in normal real life, if we were just sort of typing this, while we record, we would write, we would do things and you can go and type that we would probably do things like, you know, have you, you know, gone through an RFP process was in the past two years to select a vendor, and we’d screen people out and determine if they had gone through that process, and, and so forth to make sure that it wasn’t just sort of a legacy thing, but they’ve actually looked at this themselves. And then we could, but I think that’s this is probably good enough for just drawbars. Jeremy.

 

Jeremy Greenberg  16:33

So we just added what are your top three criteria for selecting your top vendor, hopefully? Or your primary vendor? Yeah.

 

Will Bachman  16:41

Actually, I mean, one more question I have, I’m just curious results would be like, What is your approximate annual spend on uniform rental services? And be kind of curious to see how the, your model creates those sorts of numbers? Okay, so So, and I’ll just describe for listeners. Yeah, you type that up. What what Jeremy’s doing is just a simple Excel file where he has just a list of these questions in column A. So that’s all that’s what we’ve done. So far. I just typed up this list of questions. Nothing too fancy. I, as you said that the next iteration, will you be able to type this directly into your into your browser based tool? So that’s exactly right, Jeremy here is he’s saving this to his laptop, and then I think he’s going to just upload it to the tool. So it’s a pretty quick kind of save and then upload.

 

Jeremy Greenberg  17:47

Great, so yes, we’ve got our questions. And then so now we’re going to do is upload the file. Okay. And then we have a couple of minutes where we wait. But you can see I’ve just selected the file with the questions we just did, I click Run. And the way this works is, I’m logged in, so it knows my email automatically. And so it’s going to email me the results. For me to take a look at one of the things we’re going to change in the future, is you’re gonna be able to see everything on the page here, you’re gonna see synthesis, and you’re gonna see the data in. It’s interesting, when we started, I thought that people, this is an interesting learning for me, like I thought people would really want to like dig into the data and like work from the ground up and bottom up. And that’s sort of what I normally do during what I want to do. What I find what we found is people just weren’t like the answer. So what pretty quickly. And so we’re billing in an element where yeah, you can look at the data, you can download the data if you want. But that sort of the leading thing that people really want is like, give me the insight, right? In fact, you know, that’s something I think generally with research that you don’t always see, right? You can you can get all of that you can ask all the questions you want. But if you’re not getting synthesis, if you’re not getting like the so one out of it, then you’re making people work harder. So we’re trying to create a tool that’s easier, right. And so, and frankly, a lot of those elements are not that hard to build in. The hardest part is making this thing work well and churn correctly.

 

Will Bachman  19:24

So while we’re waiting for the download, I’m curious. And that’s so funny, because definitely two personality types, right? Some people Oh, just show me the answer. And someone who’s a former consultant wants to be able to personally touch the data and mess around with like, I can do a pivot table on my own. Thank you very much. So tell me just a little bit how you built this tool. I’m curious, you know, what was involved in building something like this? Yeah,

 

Jeremy Greenberg  19:50

so first, just wanted to know if the inspiration for this really came from some amazing research that’s been done in academia. There’s so much work that’s being done to research LMS because even a company like open AI and Google in these creators, they don’t even know how their models are working right, they don’t fully understand the capability, the, you know, the the options, the experiences that their model could could actually unleash. And so I was really interested in this space, generally speaking, and listening to a podcast called me myself in AI, with in November with a professor from HBS, named Professor Isla Israeli. And I probably listened to it like five times in a row afterwards, and I’ve listened to it actually, let’s do again a couple days ago. And so what that talked about some initial research that they had done on chat GPT 3.5, which is an older version, and they started to test some basic hypotheses around things like, does chatty PD, understand price sensitivity related to income. So if I make more money than someone who makes that someone makes less money, I ought to be on average, less price sensitive, because I have more money. And it turns out, that’s true. They also tested things like the value of a brand does chat UBT understand that human consumers will value Colgate toothpaste over Acme toothpaste? And the answer is yes. So they did a bunch of these things that start to show indications that you could actually do things to simulate, you know, to have the model simulate individuals. The next step is what we did it said, Okay, well, how do you take that into the collective level? Because ChaCha PD can be great, as you know, pretending like it’s will Bachman. But maybe I want to talk to a distribution of 100 types of people like will Bachman and I want understand the long tail, I want to understand the distributions of different responses, such that they actually mirror the reality of the world, our results came in, we can go to in a second. And so so that’s really what we’ve done here is and when we’ve had the fortunate, really grateful for having a, an amazing advisor named will not mine key who is a member of Ethan Molex L M research team at the Wharton School, arguably the top researcher in this area, in the world in the country, and academia, Wall Street Journal, this is a big piece on him, he’s providing a lot of input to the US government has a best selling book out would really recommend everyone follow Ethan Moloch. On LinkedIn, he’s got the latest and greatest research. And I think, working with Linaro, has really Unleashed our ability to learn how to use this tool in ways that are a bit more advanced, a bit more creative than what one would typically do. So from an engineer, we built it. And now we’re in the process of really bolting up, you know, all the features and capabilities it has to offer. So

 

Will Bachman  23:00

from a kind of process or structural perspective, does your tool first go and generate? In our case, we had three segments 20 people each, so does it first kind of go and create the 60 different personas, and then run like interview each one of them and come up and get their answers and store it? Is that is that kind of the process? Or does it create one at a time? Or, you know, what’s the sort of sequence of events that are happening inside its brain of your tool?

 

Jeremy Greenberg  23:32

Yeah, so part of that answer is actually proprietary, because it depending on how you approach it, it can have a different impact on the results. But in essence, what you said is correct, what in essence is happening is it is creating, in our case, right? 60 different simulated respondents eat 20 For each of these three segments. And each of the each of the rows, right, if you think about rows of the respondents is internally consistent. Okay, so if we had 100 questions, and in question three, you asked, Have you ever worked at a consulting firm? And they said, No. And then question 98 says, Have you worked at McKinsey? And they say, yes, that would never happen. That does happen actually, in human research. Because people are going fast, they’re making a mistake, they don’t care, whatever that is. So so that’s sort of one of the elements elements. The other element that’s really critical is to make sure that the subsequent respondents are aware of how the previous respondents responded such that they’re not repeating them, and that they’re creating a distribution for that group of 20 that is representative of that actual segment. And that’s where things get a bit a bit creative in terms of like how we actually do that, but getting from the one To the many in a way that actually mirroring the real populations is where a lot of this, you know, a lot of the value comes in.

 

Will Bachman  25:09

Okay, fantastic. Well, let’s take a look at the results that have come in,

 

Jeremy Greenberg  25:11

take a look at the results. Okay, so I’ve been emailed the results. And you can open up an Excel, we’ll open up here and Google Sheets, because I’m maybe a convert, not really sure. And I’m

 

Will Bachman  25:27

looking at a Google sheet here, it just came in the CSV and you open as a Google sheet. And I’ll just read off, so I see column headings that are traits. And those are the three options and then the segment response, the number within that segment, so one through 20. What what industry, are you in? Food industry was the first one. The what company do you work for? What is your title at that company? How many years have you been in the industry? How many vendors do you have for uniform metal? So all the questions that we have, it has their individual answers, right. And so Okay, all right. So so some of these, some of these are probably we might have needed to specify it a little bit more, right? So we probably needed to teach the model because like, we have a barista at Starbucks, and we probably should have made sure the model understood that it was an executive who is responsible for, you know, sourcing, you know, this particular category, right? Yeah. Okay, so walk us through what we’re seeing here.

 

Jeremy Greenberg  26:29

Yeah. So another thing you can do here is, and this is like, a really good example of, you know, sometimes the questions or even the segments may not be like, as clearly defined. And so you can say, all right, well, I’m gonna go back and change them. Right. Let me make sure it’s a decision maker. And that’s a big thing. With a lot of these procurement type projects, right, you want someone that’s either a decision maker or aware of the decision making process, probably, but, you know, just for fun, you can get a feel for that. So you’ve got here, you know, we’ve got our three segments grouped into 20 respondents each,

 

Will Bachman  27:07

we’re redoing this one. Because we just went quickly, if we had wanted it to be, the person’s role is the vice president of procurement or something like that? Because we got some titles that were the host that the restaurant or delivery driver or a waitress, and we didn’t get the kind of executive responsible for sourcing, uniform rental services. So yep.

 

Jeremy Greenberg  27:32

Yeah. So I think there are some ways we can we can play with it for sure. I think the other thing is, I’ve done I’ve actually done some projects in, I could say, at least for the janitorial services, space, and sometimes the people who actually know a lot about that those types of information are actually the people who aren’t executive. So it’s a bit of a tricky thing. But But But definitely, definitely hear you there. All right. So now we can see, you know, for for it

 

Will Bachman  28:06

is it is giving us I’m sorry, it is giving us the primary vendor. And those do seem like names that I recognized, obviously, Aramark, I see Sint OS, I see. G and K Services universe. I mean, those sound like they could be uniform rental companies. I know synth tosses. So that seems and then it, it has some interesting stuff under the Y section about why they liked them and so forth. So

 

Jeremy Greenberg  28:36

all right. Yeah, I struggled through air marks. foodservice in college. So yeah, they definitely have uniform. So So yeah, so we’ve got that for so you can see the vendors for the food industry. And then you can see below that the vendors for the chemical industry, which looks a bit different, right. It’s interesting, you can see some Na is here. Interesting. And so they don’t, I think they’re indicating that they’re not using these services, right. And then you can see to so that to our question, a foreigner earlier on internal consistency, you can see it all and A’s and go to $0. For those those folks. Yeah.

 

Will Bachman  29:27

And for the chemical, we’re seeing companies like Granger, which that would seems like it makes sense for Granger for the kind of services to manufacturing type firms. Cool. So and then it gives what, so you’d probably iterate and play with this. And it’s not necessarily perfect and probably take some iteration, but at least it gets you a sense of, you know, who might the vendors be the types of answers that you might get For people, and allows you to iterate the actual survey that you’re going to put into the field and say, Oh, if I ask it this way, I get a short stubby little answer. That’s not the greatest. But I asked this different question, get some more insightful answers. So it lets you kind of test run your survey, and a quick way. Yeah, exactly. It’s

 

Jeremy Greenberg  30:19

like, you know, are they answering in a way that I expected them? Okay. In this case, it looks like we’re off a little bit on the executive. So let’s go back. And like, you know, let’s tweak that. Maybe we’re, maybe we’re missing vendors, right. And we want to actually probe on particular vendors. Another thing I’m seeing here is the annual spend seems pretty low here. Right? So I want to push a bit on that. And I guess this is a suspend that for you is this for your entire company, we do say for your organization. But sometimes these types of questions need to be clarified for humans as well. I mean, I see case all the time, we’ll see, you know, very large numbers for question like that. And like ridiculously small numbers. The other thing I want to say, well is, you know, we’re constantly in the back end, improving the model. And there are some questions that were definitely better than other questions. And we’re part of the beta is to really understand what are the types of questions that are most effective, and if we can feel, you know, really, really comfortable about, you know, different types of questions being as accurate as possible, then we really want to orient people to use those use cases, as opposed to having a tool that can be anything for everybody. So feel pretty good about the ability to, to segment different people and replicate, you know, what they, you know, their perspectives, I think, on the question type category, and like, exactly, there’s like, so many different question types that we’re, we’re analyzing, some have insanely high RF squares, when you when you compare it to human research, I’m talking, you know, over point eight or point nine when you compare series data, and then some of them are lower. So, you know, with the beta firm coming up, we want to really work collaboratively with a number of different organizations across different areas, including consulting firms, research firms, researchers at different types of companies, whether it’s healthcare, or b2b, or higher ed, we’ve talked with, there’s a lot of different areas here. And so there’s more tire tire kicking we want to do to get people comfortable. The other thing I’ll note is, I don’t see a future where this is not going to happen. And what I mean by that is the sort of panel industry in the way that their their structure and the frustration around them. And the difficulty of getting surveys and finding people is so pronounced. And the power of these models is so strong, and you know, accelerating getting better and better. I don’t see a world where if you fast forward six months, a year, whatever the right timeframe is, this type of work, whether it’s, you know, our company, or other companies or combination, someone’s going to be doing this, and I think it’s going to take a pretty big, you know, it’s gonna be pretty impactful for doing research. What’s been interesting in the demos is, and my guess is you did a demo like this too, is often people come with like their pet, sort of like, what’s been on their mind idea if that makes sense. And a lot of times, when we talk with executives, they they have like these like, gut topics, they just want like a gut feel, right? It’s like, I just want to get a general sense of people think this is a silly idea, or it’s a good idea or order of magnitude, which of these five segments do I think is going to be most interested and then see could run that really no risk at all, like no sharing of proprietary PII, there’s nothing like that going on here. And like you said, you get a directional indication. And from that, if you feel comfortable, then you can prune your research, or you could skip your research, depending on where you are, I think what everyone, including when we do the beta program, we really want people to get comfortable comparing it with their own data, because everyone’s gonna have a different, you know, opinion in terms of how their segmentation schematic is set up their sampling, they’re going to know customers differently. And so we need to get that comparison, you know, pretty early in the process. And we think that that’s probably gonna be part of the general, you know, workflow for people that are onboarding anyway. You’ve got to get really comfortable where this is going to work and where it’s not going to work as well, you know, within your organization. Great.

 

Will Bachman  34:50

So, for listeners who are interested in you know, signing up to be part of your beta or Are you getting on the waitlist? Where would they go? Yeah,

 

Jeremy Greenberg  35:04

so please email us at info at av group.com. It’s info at AV e like Avenue group.com. And, and we’ll get back to you, we do have a good amount of interest, which is great. And you know, we will we will have a waiting list. We will keep the waiting list engaged as best we can. And really, we’d love to hear, you know, ideas people have on like how they’re going to use it. And, you know, we’re looking to get this commercially viable within the next few months, you know, as a fast follow up to the beta firms. Yes, please reach out and vote average.com If you’re interested in being part of our program. Fantastic.

 

Will Bachman  35:46

And Jeremy, I know that you have a kind of a demo video that I watched and we will include a link to that in the show notes for people who want to see it with their own eyeballs. Thank you so much for joining today. It’s very cool what you’re building. Thanks. Well,

 

Jeremy Greenberg  36:01

thanks for having me. Appreciate it.