Entrepreneurs

Meet The AI Writing Campaign Fundraising Emails – Mike Nellis (Quiller)

"If you are looking at this landscape with all the tech changes that are happening now and that are gonna happen in the next year, five years, et cetera, and you are not thinking about how to totally restructure your process and your company, you are already a dinosaur."

Mike Nellis is the founder & CEO of Authentic, a digital-first fundraising and advertising agency with an impressive list of Democratic clients, including President Joe Biden. Mike was an advisor to then Senator Kamala Harris’ presidential campaign in 2020 and he’s a serial entrepreneur. His latest company is Quiller, an AI copilot that helps Democratic campaigns and allied organizations draft and deploy high quality, effective fundraising content.

In our conversation we learn about Mike’s journey launching Quiller, some strategic decisions he made along the way, and his advice for other entrepreneurs building with AI.

Key Points

  1. AI in Political Campaigns: Discussion about the use of AI in political campaigns, specifically in writing fundraising emails. The conversation highlights the restrictions of commercial AI products against political use cases and the emergence of open-source large language models.
  2. Tech Stack for AI Campaigns: Mike Nellis discusses the process of building their tech stack, including iterative trial and error and testing different foundational models. He mentions testing their system against OpenAI.
  3. AI vs. Experienced Digital Strategist: The conversation explores whether AI can replace an experienced digital strategist. Mike shares that they've rolled out their AI tool to 40 clients without any complaints. However, he emphasizes that AI is not near replacing humans in the process but rather aids in efficiency.
  4. Efficiency with AI: The use of AI in writing fundraising emails has significantly reduced the time it takes to create content. The tool helps to start the writing process at 75-80%, and the human's job is to close the gap to 100%.
  5. Future of AI in Campaigns: Mike shares his vision for the future of AI in campaigns, including tools to identify deep fakes and disinformation, analyze voter sentiment, and improve polling efficiency. He also discusses his desire to reduce burnout in the industry and make campaign work more sustainable.

Transcript

Mike Nellis:

If you are looking at this landscape with all the tech changes that are happening now and that are gonna happen in the next year, five years, et cetera, and you are not thinking about how to totally restructure your process and your company, you are already a dinosaur.

Eric Wilson:

I'm Eric Wilson, managing partner of Startup Caucus, the home of campaign tech innovation on the right. Welcome to the Business of Politics Show. On this podcast, you are joining in on a conversation with entrepreneurs, operatives, and experts who make professional politics happen. We are joined today by Mike Nellis, founder and c e o of authentic digital first fundraising and advertising agency with an impressive list of democratic clients, including President Joe Biden. Mike was an advisor to then Senator Kamala Harris's presidential campaign in 2020, and he's a serial entrepreneur, which is what we're talking to him about today. His latest company is Quiller, an AI co-pilot that helps democratic campaigns and allied organizations draft and deploy high quality effective fundraising content. In our conversation, we learn about Mike's journey launching Quiller somes strategic decisions he made along the way and his advice for other entrepreneurs building with ai. Mike, there are a lot of ways that generative ai, which is what kicked off the, the current trend cuz we've been using AI through machine learning in politics for a long time. But there are a lot of different opportunities for generative AI to disrupt campaigns. Why did you start with online fundraising?

Mike Nellis:

I think I started with what I know and like most great businesses, I wanted to start with a problem that I understood and then I thought was worth solving. And for me, as I look out into the political landscape and I got very interested in generative AI and what AI can do, I looked at at, at three core problems in my own business, authentic, which is we started authentic on the online fundraising side to be a firm that can generate high quality content for online fundraising messages that'll actually raise money. And on our side, and I think on both sides, to be honest with you, there's an awful lot of scams and spammy messages and a lot of abuse that I think our, our donors take. And we wanted to build a business where we could run really great, you know, narrative driven fundraising programs.

It's easy for us to be able to do that when I have 20 years of experience and most members of my team have significant amount of experience and a culture and a and and standard operating procedures that lend itself to that. But a lot of folks can't do that cuz they don't have the time, they don't have the experience. So I think a tool like Coer can help that by using AI to generate really high quality online fundraising content. I also am seeing a tremendous amount of burnout on our side because we're asking, you know an increasingly small number of people who wanna be involved in the political space to do more work than ever before. Yeah. So building a co-pilot that can do this type of work, I think is going to like help people stay in the business longer, prevent burnout, let people go from, you know, 60 hour weeks to 50 hour weeks to 40 hour weeks, et cetera.

And there's a lot more applications to that. And I also wanna make sure that, you know, anybody who wants to be able to run a high quality fundraising program can do that down ballot because right now, you know, authentic is expensive. Most of the digital firms in the space are expensive. I imagine that's true on the republican side as well. If you're running for state senate, city council, what have you, you can't really afford to pay someone like me or to pay a firm like mine. And so that's the problems that we're trying to solve right

Eric Wilson:

Now with political tech. There's always this build versus buy calculation, right? Where there are things that, do we need to build it for our, just our industry? Kind of that vertical software question or is there an off-the-shelf solution? So where were other tools like chat G P T or Jasper AI is another one that I've played around with. Where were those falling short that you thought there was an opening for for something that was industry specific?

Mike Nellis:

So the first piece of that, particularly on chat G P T, but really on any of these like broad scale foundational AI models or large language model is their terms of service actually prohibit or severely restrict the ability to draft political content, right? The reason for that is cuz they're worried about, you know, disinformation campaigns, deep fakes, the sort of like, you know, worst cases of this type of technology, which maybe we'll get into at some point during this conversation, but it also ends up restricting like legitimate uses of the tool, right? To write fundraising emails, to write press releases and write tweets and stuff like that. So now you can trick chat g p t into writing you a fundraising email for a candidate like, you know, Adam Schiff or Kamala Harris or what have you, but if you try to do it in scale, they're gonna shut you down.

And over time I think these platforms are gonna get more restrictive on this as they find the sort of use cases like in the beginning of chat G B T, you could get it to give you financial advice, right? <Laugh> down on that significantly because there's like a liability and a public relations concern to doing things like that. So I think we're gonna see more of that and that's why it's necessary for there to be like tools that are specific to us. Jasper's sort of the same, my understanding of Jasper is like, while it is how much more narrow focused AI designed to develop marketing content for businesses, it's still built on top of open ai. And if you look at Jasper's standard operating, or you look at their usage agreement about it does prohibit like political content and some of these others, like I know the founder of Go Charlie is another one.

And I really like them for like using like for company marketing but they prohibit political content too. So I think we're gonna have to build our own solutions. I would encourage people to think about these tools almost like they're, they're big 10 companies, their Facebook, their Twitter, et cetera. Like they're gonna restrict what you can and cannot do. And while I can't go build a political Facebook, and I think that would be a tremendous waste of everyone's time, <laugh> I can build some like narrow use AI that has real practical applications that are gonna help people in this election save a lot of time and reduce a lot of burnout and raise a lot of money. Hopefully

Eric Wilson:

That approach to politics as sort of untouchable by tech companies, especially the newer ones that are coming into the market post 2016. Right? The, I think there's a real divide. So you look at any new tech that's coming out, they don't wanna get involved in politics at all. I don't think that's particularly fair or, or helpful because the bad actors are always gonna figure out how to get around that. It's almost like this criminalization of politics and we're, we're seeing it with, you know, whether it's restrictions on targeting that you can use for Facebook, what you can and cannot say on TikTok. It's just a pattern that I'm seeing here and, and it's prominently raised with OpenAI and Chat G P t

Mike Nellis:

I think it's a concern and one of the largest challenges here is that there's just like no regulation on any of this stuff. You know, even let's set generative AI aside for a moment, like on the work that like you and I do, like, there's almost no real enforceable regulations on like online fundraising, for example, right? Right. and that's why the New York Times has done a really good detailed series of articles about how the Trump campaign is just like openly scamming their supporters every single day. And I've had a number of conversations with the reporters who have worked on that and like, that's the thing that happens. There's plenty of people on the on the left who do that too. There's all kinds of like scam super pacs with like BS names that aren't real and they aren't directing the money. When I worked on Bernie Sanders in 2016, I remember that there were like all kinds of scams of people who were like siphoning off money that meant to go to Bernie to their own committees and stuff like that.

I think some people got caught doing that too, but you have to be so egregious and so brazen to get caught targeted, prosecuted, you know, fined what have you. It's so hard to do. And like there's so few people that are like trying to do it right. And trying to build, you know, good quality programs. There are a lot of people who are, you know, trying to direct the money to the right places, to the candidates and causes they're working on, but they're doing it in a way that are gross. So like I I get like big tech and open AI and I get anybody's reticence to working with political campaigns. Like it's gotten messier, it's gotten more polarized. I think the extremes are the extreme use cases, especially things like DeepFakes and stuff like that are gonna make people like even more nervous and like in the absence of like governmental regulation, they're gonna be like, I'm just gonna try to weed it out.

Eric Wilson:

Yeah. It's this really interesting, I don't know, crosscut that we're stuck in right now, which is with more analog means of communication, like TV mail, there's kind of a private place for political speech, right? We get a lower candidate rate, for example. But whenever more of the public square is moving into private companies, absent positive regulation from the government, and I mean positive in the, the, the sense of proposing rules, they're just left to their own decisions and most everyone is choosing, Hey, I just don't even wanna mess with that. So it is a, it's a, it's a challenge. I think a lot of people who are outside of the industry don't really appreciate it. Switching back to Quiller now, it's built for and by Democrats and in fact on the website it says, we will never offer these super tools to G O P candidates. Explain that thought process behind that decision. For us,

Mike Nellis:

I think a big part of the reason that we launch QR was to make sure that our data and the work that we're doing is secure and it's gonna be used to elect the candidates that we believe in. And I'll offense the other side of the aisle, but if I'm teaching a robot to write fundraising emails, I don't want it raising money for Ron DeSantis and Donald Trump in the same way that I think there's probably lots of tools on the right that you wouldn't want to be used by Joe Biden and Kamala Harris. And so I also think there's like a, there's a cultural like inside of the industry aspect of this, which is like, there's distrust of tools that service both sides. And that's why you see, you know, like, you know, you have ActBlue and you have Win Red, right? Ironically, I actually have a nonpartisan client that's using both ActBlue and Win Red right now, which I think is bizarre.

But to me it's a really interesting use case. But at the end of the day, like if I'm giving my contribution data to a company, I want it to be a company that aligns with my values because my contribution data is incredibly important, right? And I think if I'm giving information on my polling information, on my tactics information, on my tone of voice, information on how my my online fundraising program is performing, I would wanna make sure that one, that data is secure by my side, but two, making sure it's not being used to tune an algorithm that's used in other ways. That's part of why, like, I think I get concerned about putting like any voter data or any strategic or any polling data inside OpenAI and chat GPTs apparatus, right? I don't know how that data is gonna be used and I don't know if it's eventually used to help a candidate I might not support or it might not be secure in some way. The, the line that I often joke with people about is like, and this sort of comes after Elon Musk taking over Twitter, and I think turning off like 95% of liberals to Twitter is like, do you trust a company that was founded by Elon Musk? Because I don't

Eric Wilson:

<Laugh>. So first a product question and then I wanna go to kind of like a more strategic entrepreneurial question for you, but do you think that, let's say you gave me access to Quiller, do you think I could write effective fundraising copy for Republicans? Or is it so different that there wouldn't be overlap?

Mike Nellis:

I think, I think the way that the tool is built, you could use it for a Republican candidate if we gave you access to, because the way that the tool works is like we've created a database of like top performing high quality fundraising emails. We taught the tool how to write in certain styles, and then there's a layer on top of it of who the candidate is, why they're running, here's some writing examples, here's how they feel about, you know, this issue, that issue, this issue, here's anything else we wanted to know. So it combines those two, and then you have individual inputs as well. So if I went in there and told the tool, Hey, if this is a Republican and the Republican cares about, you know lowering taxes and appointing right-wing Supreme Court justices, then that's what it's gonna write about. And that's partly why we need to make sure that the tool that we're building and investing in is being used towards our values. And I think you're gonna see more bifurcation of tech like that than than you would think.

Eric Wilson:

Yeah, and I agree with you on your decision from a strategic perspective and it's advice that I give to our companies here at Startup Caucus. Like a lot of people, especially from the the business world, they look at political tech and say, why can't this be nonpartisan? It's tech, and you've listed several good reasons, but also there's that we just don't have a good pattern of this being successful. The, the total addressable market is Republican politics and democratic politics. It's not politics. And I think that's an important nuance. And I really, when I talk to potential founders and they're, they say, Hey, you know, they talk to me because they lean, right? But they don't wanna build a partisan company. I'm saying, well, that's just does not work in the political space even though we wish it weren't the case. I think there are tools that you guys have that I wish we could use and vice vice versa. But for all the reasons you list it, it's gotta be separate. And I think also our user bases are different, right? You know, the people who support Republican candidates, the way Republican campaigns are run is different than the way democratic campaigns are run or the, the people who support democratic candidates.

Mike Nellis:

Well, and what, and what motivates these bases is different too. And I know that from working on some nonpartisan stuff that I've done over my life where I've tried to sort of either, you know, play to the middle or try to play to both. And I've, I've run campaigns where I've actually bifurcated the digital program into like, here's what a liberal focused, why here's a conservative focused view. It's on the same topic essentially, but you would sort of change it out and kind of like mess up, change the messaging a little bit. I might go more heavy, you know, freedom u s a design graphic here, and I might go a little bit more like, you know, empathy, compassion oriented, longer form narrative stuff. So mixed results. I don't know that anybody's really figured that out.

Eric Wilson:

So coming from authentic, you, you obviously had a ton of emails and data to train the AI model on. I think that just for our listeners who may not be familiar with this these large language models like Chad, G B T and what you've built with Quiller rely on training data. They're just, this is very reductive, I understand it, so don't send me emails on it, but ai, large language models, chatbots are, are really sophisticated auto completes. And so you, you're providing this data to it. How did you decide what went in and, and what didn't?

Mike Nellis:

I think one way to think about just chat G P T is that it's basically a giant search at John and you can communicate with it. You can kind of treat it like a centi robot, but it's really not. It's a conventional wisdom machine, right? Right. <LAUGH> and G B T is fed on basically the entirety of the internet up until 2021, I think. Whereas Quiller is more of what I would call like a narrow focused AI where I'm narrowing down what I'm trying to train it on. So we sort of, you know, I met with my leadership team and you know, folks that I trusted the industry and basically said, all right, what makes a good fundraising email? What's the construction of a good fundraising email? What do you need? What do you don't need? You know, we would talk about things like a solid theory of change, what makes for a good subject line?

Why do we think this email is good and kind of deconstruct it that way. And then really went to the campaigns and causes that I've worked on over time and try to identify, you know, some of the best performing emails and some of the emails that we think, you know, are gonna, are gonna turn out, some of 'em are gonna be replicatable, right? Because there's gonna be some that are like highly specific to a candidate, right? Like if I started, you know, probably some, probably if I, if I looked at it, the best performing fundraising emails I've ever sent are probably like the emails we sent around Ruth Gater Ginsburg's death in 2020, right? Those are not really applicable fundraising. You must, so we didn't put stuff like that in there. Interest, we kind of looked interesting, one of the skeleton and bones of like, what we need here and then kind of go from there. And at a certain point it was like, you put in a thousand, you put in 2000, et cetera. It's like now it's producing new content. And what we ended up doing was just basically like retreading the model, you know, every couple of days and like being like, here, read all these emails, write some emails, then we gave it edits and then we told it to absorb those edits and we kept going, right? It was just not an iterative process. On and on.

Eric Wilson:

Did you include performance, like open rate, click through rate, amount raised as part of that training? Or was it just we're giving it the, the content, the email copy?

Mike Nellis:

It was, it was part, it was part of the thought process as to what we chose. But it wasn't it wasn't what we gave the tool. We just told it to take it at face value. Like, this is what we're asking you to draft, we're asking you to draft things that are like this. And a lot of it is like, as sophisticated as Quiller has become, it's really only a tool that's like 12 or 14 weeks old. And I think the fact that we've gone from, you know, concept to prototype to MVP, to kind of having like an alpha and a beta that's out there now with like 40, we have 40 campaigns that are using it right now. Some with like a great deal of success. It's come so far, but it's still very much just like an mvp. You know, our idea was let's prove that this can do one thing really, really well. And then when we prove that it can do that one thing really, really well, we can prove that we have a market and interest for it. We can prove that campaigns are gonna wanna do this and then we can start investing in other stuff. And now I feel like we've gotten to that point, so we're building like additional features and additional opportunities on the backend while continuing to iterate and improve on the model we have. Now

Eric Wilson:

You're listening to the Business of Politics Show. I'm speaking with Mike Nellis, c e o and co-founder of Authentic Campaigns about his latest product Quiller ai. So Mike, this is obviously brand new territory. What, what advice would you share with our listeners who might be looking to incorporate AI into their own products?

Mike Nellis:

My top level advice would be stay away from like the shiny object syndrome of these things. Like, it's, it's queued that like the, the G o P put out, you know, that like AI generated video. And I think there's a place for like what I would call stunt tech or something like that, which a good example of that would be on on May 4th. I work with the shift campaign and the, and the campaign put out this AI generated image of Adam Schiff as a Jedi. You know, that's, that's cute stuff. It's gonna get reporters' attention, but it's not gonna make a difference. I think my advice to anybody who's working on this, whether it be on a campaign in an organization at a business, it's what is an actual problem that I have that I understand that can be solved with this. It's not a shiny object. It might not, it might not be sexy, right? I don't, I would honestly argue like writing online fundraising emails is like the least sexy part of my <laugh>.

Eric Wilson:

It's miserable,

Mike Nellis:

Unfortunately. Something I'm really good at though. But it was like something that it seemed like this technology could do. It seemed like something I understood and it seemed like something that was achievable when we conceived this idea. We can, we conceived this idea six months ago before CHATT G B T was even launched because the technology to do what chatt G B T does already existed. And so we were already working on it before it dropped. So that's my advice. Find a problem we're solving and go take the next right step.

Eric Wilson:

There's an important part here, which you raised earlier, which is that a lot of the, the sort of commercial AI products have restrictions against political use cases, but we're starting to see some open source large language models. W do you have any advice on like how did you source out your tech stack for this to share whatever you can or feel comfortable doing about

Mike Nellis:

Quiller? I mean, we took a, I mean, I'm gonna sort of not get into the specifics of like what technology we're using and stuff like that, but I will say that we did like a, like an iterative trial and error type process here where it was, we, we built, you know, a really ugly looking prototype in the beginning and basically just had it work off different foundational models that are out there and say, okay, let's look at model A versus model B versus model C. We actually tested it versus open ai. Cuz open AI allowed us to do that. I don't think they would've allowed us to do it at scale like we're trying to do now, but like we wanted to see like, hey, how are these, how are these other algorithms performing? And like, they performed pretty well. Like I didn't feel like we were missing a ton in terms of what we were doing, but to be honest with you, with the amount of data we're giving it, we're not asking it to do a particularly like complicated thing, right? I think it becomes more complicated when it's like, you know, give me advice or like, you know, like the more detailed it becomes, the harder it gets. And so that's where some of the other layering we've done on top of it is. But in terms of the foundational models, they're, they're out there, they're accessible.

Eric Wilson:

All right. Is productizing a best practice enough? Can, can you really replace an experienced digital strategist with artificial intelligence? I mean, have you done like a touring test on this yet?

Mike Nellis:

I haven't done a touring test, but I have, I mean we've rolled this out on 40 clients and I haven't heard a single client complaint. That is how I would put it. So nobody,

Eric Wilson:

Do you think that, I guess do you think that clients could pick it out?

Mike Nellis:

I I don't. I really don't. Yeah, but the thing, the thing about it is like, I do not believe that like any of this technology is anywhere near replacing humans in the process. I view this as like when, right now, pre chat G B t pre generative ai, pre coller, if I needed to write a fundraising email, I had to sit down, open up a blank Google Doc and spend an hour writing that fundraising email, right? Yeah. And pick your poison for me, it probably takes 20 minutes for my junior staff, it might take 90 minutes, but the average is about an hour right now. I know what I need generally, broadly, and I can go in and instead of starting in zero, I start at 75, 80% and then my job is to close the gap between 80% to a hundred and get that ready for the client cuz there's nuances and things that I might know that the tool doesn't know, the tool doesn't recognize, et cetera.

So we're closing the gap between how long it takes to do it without generative AI and how long it takes to do it with it, right? If it goes from an hour to run a fundraising email to 10 minutes, you know, all told after edits and client approval, like that's a huge acceleration of our ability to get content out the door, right? If we teach this tool to code the fundraising emails, which we're working on those CRM integrations now, then you go from a process that takes 30 to 40 minutes to do and is highly repetitive, something that takes five minutes to do and it could let us work on more accounts, it could let us run more efficient programs, it could let us spend less time focused on repetitive tasks and more time doing other things. Creative problem solving, strategic thinking doing more, more things with like reactivation campaigns. I'm already seeing that play out on some of our, on some of our accounts right now. So, but I don't think we're gonna replace people with this, nor do I think we should. The other point that I'll make here is, and I've said this to like my staff cuz I think if I was,

Eric Wilson:

They're getting nervous

Mike Nellis:

<Laugh>, right? If I, if I was anybody in this field, I mean, I, I mean frankly I went to build Quiller and it sort of accelerated building quiller after Chachi BT came out because Cheche bt scares the hell outta me. It scares the hell outta me. Like I'm a creative type that spent the last 20 years making a life off of writing fundraising emails or writing digital ad copy it's stuff. And like now it, a robot can do that in a matter of seconds and it's <laugh>, you know, it's not as good as mine yet, but like, it, it's going to be so, like, I'd rather build the tech myself, but I can't speak for the G o P side. But on the temp side, we don't have enough people right now to do the work that we wanna do, right? Which is why we're burning through people. So I'm not gonna replace a single member of my team because I built this tool. I'm gonna unleash my team to go do bigger, better, more creative, more human-centered things than writing fundraising emails and coding fundraising emails.

Eric Wilson:

So Mike, I, I personally have the opinion that there's, there's too much moral panic around ai. You mentioned some of the, the worst case examples of, you know, deep fakes disinformation, that sort of thing. But there is an ethics question that I, I was raised by a previous guest on this show that I also hear in my conversations. And so as professionals, how do you think we should approach sending deliverables to clients that we're generated by ai?

Mike Nellis:

I believe that you should disclose that to your clients. Yeah, I do. And then we, we have disclosed to our clients that we are using this tool. I don't think you need to stop every single time and be like, Hey, this was generated by ai, this wasn't what, say like, what we did was we went to our clients and we said, Hey, we built this tool, we're gonna start using it. Here's, you know, some frequently asked questions that we imagine you might have. If you have other questions, reach out to me, but just know every single thing is gonna be reviewed by a human. Because I think as we all know, Chad, G B t, Bard, all these other large language models and AI tools, they hallucinate quiller is no different than that. So like the idea that like, as great as Quiller is and it is a really great and special tool, it still makes mistakes.

And so like, I can't set it loose without a human that knows what they're doing. And we're not gonna do that. So I think, you know, you've gotta disclose that to your client. Whether or not you need to disclose that to voters is sort of another topic of conversation. And the, the f e C kind of came in and decided not to require that to happen last week. I think is probably gonna be the guiding principle for most folks. I, I sort of oscillate wildly between how I feel about that, to be honest with you. I

Eric Wilson:

Wanna know what Quiller hallucinates about. That's, that's probably some, some weird like daily coast comment section <laugh> talk?

Mike Nellis:

No, it'll just, it like anything, it'll sometimes like hallucinate like a fake part of the bio or something like that. Oh, okay. It's just minor, minor stuff. It's really, I mean, I haven't seen

Eric Wilson:

Not, not as interesting as I was hoping <laugh> No, I tend to agree. I think there's a little bit of concern with like letting the curtain slip with clients, but the way I view it is, it's, it's getting you leverage. It's, it's like a tool, you know, I, you have someone come to the, to the house and build something for you say, look, I don't, I don't care if you use a handsaw or an electric saw, right? I just, I want the, the job done. And, and I think with, with clients in this space, it's, it's similar almost to the point where if you're not using all of the tools at your disposal, I would be disappointed in you.

Mike Nellis:

I think if, if you are looking at this landscape with all the tech changes that are happening now and that are gonna happen in the next year, five years, et cetera, and you are not thinking about how to totally restructure your process and your company, you are already a dinosaur. Like you are basically every organization in baseball that did not rebuild their entire apparatus around Moneyball in the nineties. It like, it's the same core principle. And I am a huge Moneyball fan and I've been joking a lot that I'm living my Moneyball moment, but it does kind of feel like that, like, this is either gonna eat me alive or it's gonna make me stronger. Like that's purely how I feel about it and it's purely because futile, you're still living hell outta me. So, sorry, go ahead.

Eric Wilson:

I said resistance is futile. Yeah,

Mike Nellis:

Exactly. We're, we're all part of the board now.

Eric Wilson:

So Mike, you you hint at this with the coding emails, but what's the next AI tool that you wanna see built for campaigns?

Mike Nellis:

Well, we're, aquir is gonna code emails. I mean, we're pretty close to announcing that. Just sort of putting the final touches on some of those CRM integrations and navigating the other tech companies. I, it's hard to say what I wanna see the new, the next AI tool to be. I mean, to be honest with you, I'm kind of excited for like somebody to release a tool I hadn't thought of yet because that's the stuff that gets like, where I'm like, wow, I never considered a use case like that before. I think that's gonna be so cool. I do think like somebody is Saudi somewhere is gonna be able to develop a tool that kind of identify, you know, a little bit more when there are deep fakes and when there's sort of disinformation happening. Like I think you're gonna find some AI tools that can kind of help with that.

I think that'll be interesting. I think anything that can analyze voter sentiment and polling information in a more interesting way. I think you've been on a lot of campaigns, Eric, you know, the science of polling is hardly a science <laugh>. So no offense to my friends in the polling community whom I love, but I think it's, I'm just, I wanna be more efficient, I wanna be more efficient, I wanna be more effective. I think that the races we're working on are really important. Whatever side of the aisle, I hope you actually care about what you're doing. And I, I know you do Eric, so I think <laugh> finding opportunities to run these campaigns in a more efficient way, and I wanna burn out fewer people. Like I feel like we go through way too much burnout on our side. I wanna, I wanna solve that problem. I want people to be able to go to work every day, feel like they're trying to make a difference, feel like they're trying to make the world a better place and I don't know, clock out at five and have a family, you know, wage and I think that's achievable in our lifetime.

Eric Wilson:

Yeah, no, I'm gonna be the old campaign curmudgeon where, you know, you have to move, move somewhere, work unsustainable hours and live off of volunteer pizza,

Mike Nellis:

<Laugh>.

Eric Wilson:

But you're doing more now because you've got ai.

Mike Nellis:

Exactly. You still work in 80. Yeah, I just don't, I, what I want to do and is really important to me is I don't wanna build a tool that like just fills the cup, right? Where it's like, oh, you're supposed work 80 hours, but now you're twice as productive. Like, that doesn't, I wanna change the culture of it because the culture of this industry is, is, is broken.

Eric Wilson:

Well, my thanks to Mike Nellis for a great conversation. You can learn more about him in our show notes and if you're a Democrat, you can go sign up for quiller Republican listeners should check it out too. Look around the website, he's built something really cool. So if this episode made you a little bit smarter or gave you something to think about, you know, all we ask is that you share it with a friend or colleague and it helps introduce the show to more people, makes you look smarter. So it's a win-win all around. Remember to subscribe to the Business of Politics, show wherever you get your podcasts. That way you'll never miss an episode. And if you'd like, you can sign up for email updates at business of politics podcast.com. With that, I'll say thanks for listening. See you next time.

Continue Listening