820 | His AI System Gives Agencies a MASSIVE Edge Selling to Local Businesses | Jeremy Grandillon

The first time I showed an agency owner what the system can do, it went quiet for about 30 seconds, and then said, I’ve been doing this wrong for six years. That’s Jeremy Grandillon, founder of TC9. Jeremy went from head of sales making every mistake possible to building a system that detects when a business is actively in a buying window and reaches them in under five minutes automatically. Your competitor’s system detects it at 9 a.m. qualifies the business against their ICP by 9 or 1 and reaches the profile by 9 or 3 and sends a personal first message by 9 or 5. By Tuesday afternoon, they have a reply and a call booked. You see the same signal on a Wednesday and the owner tells you. In this episode, Jeremy reveals how to stop sending cold outreach that goes nowhere and start reaching local businesses at the exact moment something just went wrong. A med spa gets 12 bad Google reviews in 10 days and that owner is in full panic mode.

This is not a cold lead. That is a business on fire. He breaks down the five AI employee roles that are sitting completely vacant inside most agencies right now, each one running 24 hours a day. It’s a structural competitive advantage that is impossible to replicate manually. And he shares the one shift that changes everything for agency owners who have been grinding cold outreach for years and wondering why nothing is working. So the agencies that are winning are not necessarily sending more message. They are sending fewer but in a better timing.

Tell me a little bit about the technology that needs to be in place in order for that to be a reality in 2026. Make sure you emphasize how like this isn’t possible without automation, is it? Okay, let me tell you what happens in those 72 hours. A local business owner posts a job opening on Monday. Your competitor’s system that takes it at 9am qualifies the business against their ICP by 9 or 1 and reaches the profile by 9 or 3 and sends a personal first message by 9 or 5. By Tuesday afternoon, they have a reply and a call booked. You see the same signal on a Wednesday when you do your manual LinkedIn check. You spend 25 minutes researching the business, write a message, send it, and the owner tells you they’re already talking to someone.

The five-minute response is not a nice to have. It’s a structural competitive advantage that is impossible to replicate manually. A human cannot monitor hundreds of signals simultaneously, qualify a business against eight criteria and reach a profile with recent activity and write a personal message all in five minutes. For the five-minute response to be possible, you need four things running. First, you need a continuous signal monitoring, something that washes your target sources around the clock. Second, you need instant ICP qualification. The waterfall filters runs. The moment signal fires before any human is involved. Third, automated enrichment. The system pulls the context it needs for outreach without waiting for a person to research it. And fourth, triggered outreach. The personalized first message goes out based on the enrichment referenced to the specific signal that triggered it without anyone approving it manually. Jeremy, a lot of agencies out there, when they’re thinking about their prospecting problem, they think that it’s a volume problem, not enough leads, not enough outreach. Tell me what’s wrong with this diagnosis.

Yeah, that’s a common mistake. I mean, it’s due first because of history. It has been like this for a long time. And also, it feels normal if it’s not working anymore. That’s basically human. So the volume diagnosis feels right because it’s visible. You look at your pipeline and it’s thin. So you think you need more at the top. So you buy a bigger list, you send more emails, you post more LinkedIn, nothing changes, because the problem was never the volume. What we are missing right now most of the time is not we need more leads, we need better leads, and better leads are the ones that are signaling you, giving you intent that they’re interested or they are facing the problem you’re solving. And the real problem is timing, a perfect message to a local business that is not buying is just noise. It doesn’t matter how good the offer is, if they are not feeling the pain, they are not moving. So the agencies that are winning are not necessarily sending more message, they are sending fewer, but in a better timing.

Tell me about the difference between a lead list and a buying signal, because some agencies treat them sort of the same way. And that type of thinking costs them deals every single week. Someone sat down and featured a database based on industry, company size, locations, and then they gave a list of names. But those type of lists don’t say anything about the timing. If those businesses need you today, it’s static. There’s no timing information, so it is a list of people who might be interested at some unknown point in the future. But buying signals is evidence. It tells you that something is happening right now inside that business that makes them more likely to need you to need what you sell. A local restaurant that posted three jobs opening this week is growing and growing businesses need marketing. Med Spa, that have just received 12 negative Google reviews in 10 days, has a reputation crisis, and reputation management is something an agency can fix. Plumbing company that’s just updated their websites with a new service is expanding their offer and they need to tell people about it.

So the spray and pray model died because it was entirely list based. It worked six or seven years ago when nobody was doing it, but today the inbox is flooded. The only thing that gets true is relevance and generating relevance is only possible when you know what is happening in the business right now. If you take an agency that’s selling AI solutions to local businesses, businesses like plumbing companies, spas, what are the real buying signals that indicate that a business is in the buying window? The real ones. Could be multiple things. Actually, local businesses are sending those buying signals constantly and most agencies are just not listening. Let me walk you through the ones that are actually converting. Hiring is a good one. Kind of a classic signal. If they are hiring, that means they’re probably growing or looking to solve a problem they have in their business. So that could be a really strong signal for hiring signals are some of the strongest when a business posts a new job opening.

That means they are growing and growing businesses needs marketing. They need more customer and they have budget for that review. Velocity is underused and very powerful. If a business starts receiving, you know, a lot of negative reviews, especially in a short window, they have reputation crisis and they know it and they have to manage it. So reputation management, AI review response, disability tools, that becomes suddenly not a nice to have for that business, but an urgency. And a sudden spike of positive reviews means that the business is doing something right and they are probably in growth mode. Website changes is also a strong signal. That means that they are thinking about their future. A new service page, pricing updates, new brands, something moved internally and they have changed what they show outside to the world. So that’s an opening for a conversation like socials, of course, social media. If they are active on social media, this is basically them giving you information and signals all day long. So you want to monitor that and search on socials. Social engagements are the foundation of my entire

warm outbound system. When a local business owner engage with content about AI, about marketing, about growing a business, they have just told you something about their state of mind. They are actively thinking about those problems. That’s the right moment to reach out, not three weeks later when this sold as past. Expansion signals like new locations, announcements, press coverage, grant awards, funding, those public declarations, they say, okay, we have money, we intend to grow. Any agency that sees one of these and does not immediately reach out is leaving the door open for a competitor. And one, the most agency is completely missed is the job change signal. When a new marketing manager or operator joins a local business, there is a 90-day window where that person is actively looking to make their mark. They want new vendors. They want to try new things.

They are the most receptive they will ever be. That window goes fast. So you get your marketing machine pumping and you’re starting to receive hundreds of buying signals from a lot of companies. But a lot of them aren’t the right fit. So you have to be able to filter them. So talk to me a little bit about how you build an ideal customer profile. And just as a follow-up to that, how do you ensure that it’s operationalized and it’s not just a document sitting on your desk now that you’ve created it and nobody ever looks at it after it’s done? Yeah, we are used to, I would say, back in the days, those personnel coming from marketing more or less, like, you know, this is journey and is doing this and that. Now we say with those systems we are building an automation, we need more concrete, specific points, data points that are false or true.

This is where most agencies discover that the ICP is actually useless. They have a persona document that says something like, you know, small business owner, 10 to 15 employees interested in growth, but that tells you nothing. You cannot build a filter from that. You cannot write a prompt that checks for that. It’s more a description, not really a definition. An operational ICP is more of a series of true and false question. If you cannot express every element of your ideal client as something a system can check, yes or no, true or false, your ICP is not specific enough and then you go through one criteria after the other and if they match your criteria or most of your criteria, then it’s a good match, then it’s a good ICP for your business. So I would go there and that also avoids to have them as a document that’s sleeping in the Google Drive or whatever tool you’re using. It’s more like a dynamic, I would say, dynamic criteria list that you implement in your workflows and it’s right on. For an agency selling AI to local businesses, a real ICP looks like the business has between two and 15 employees.

They are in a service vertical, you know, home services, health, wellness, food and beverage. They are in a metro area of at least, you know, a thousand, a hundred thousand people. They have an active Google business profile with at least 20 reviews. They have been operating for more than two years. The decision maker is the owner or named Operation Rose, not the marketing department. And if you already have clients, the fastest way to build is to export your serum, put that into an AI analysis tool and ask it what do my best clients have in common that I have never consciously noticed. And you will find the patterns like, you know, geography, clusters, revenue ranges, vertical overlaps, review counts that you never articulated that you already knew as long as you have a clean serum, of course.

You’ve described sort of a waterfall filter and one where a prospect has to go through a specific gate and they have to complete that gate before they move on to the next gate. And as they go through this system, these are sort of the free checks that they get through before you’ve even spent any money on AI. So tell me a little bit about how that system works. Yeah, we are excited to jump on the AI things and features. But before that, and behind that, behind AI, that’s data and calling, you know, API calls to get, you know, information and then check if it’s false or true, can be free or at least very cheap. So you want to start with those hard data criteria. Think of it as, you know, spreadsheet connected to intelligence. So the first column is your input, like LinkedIn profile URL, Google business listing, website URL, everything to the right of that is a question the system asks about that business. The first gates are free, is this business already in my database, check your Google sheet, could be a serum, check your serum, is this business on my blog lists, existing clients, referral partners, competitors, does this business have a basic web presence like

websites, Google profile, that’s a simple lookup. These gates eliminate a huge percentage of the raw signal list before any AI is involved at zero cost or very cheap. Then the AI gates begin and here is the most important principle. You do not ask the AI one complex question about each business. You ask it five or six very simple ones. You want to break big problems into small pieces. Is the business in the right geography? You give the AI the location and ask true or false. Is this in my target region? It answers. If it’s false, the road stops. That business costs you almost nothing. Is this business in my target’s vertical, give the AI the business category and ask true or false? Does this business fall into home services, health and wellness, food and beverage, et cetera? If false, stop. If true, continue. Is the decision maker reachable? Does the business have between two and 15 employees? Is there a name, owner or operations contact available? Only after every gate passes does the system move to enrichment, where it reads the business’s recent activity, reviews, websites, social presence, to be the context for outreach. Enrichment is where the real

cost lives. You only run it on business that have passed every check. You can remove the noise and just focus on the ones that are the most warm and the warmer and the most qualified. You can really optimize your workflow by starting with the hard way, the old school data points. You’ve described five critical AI employee roles that a lot of agencies are leaving on field. Ones that run 24 hours a day, seven days a week, never have a bad day. Walk me through each one of those and tell me which is the one that everyone has to make sure that they build first? All the repetitive tasks that you are doing, that’s a strong, strong signal that you want to automate those things. For example, I mentioned a few AI employees that we have.

Content creation is one of our main acquisition channels, so we do content and based on the traffic it generates, we capture the attention of potential prospects. To help accelerate and assist me in the content creation, I have a researcher that will search on a daily basis the trends, the news in my industry, what would be, according to my criteria, of course, what would be a good topic for content. I have this served on a summary every day, so that’s a researcher that I have online that’s doing this, so I don’t have to go through those websites manually, etc., or socials. That’s one. Another one is the prospecting one. Obviously, the one we described earlier, like featuring qualification, creating those workflows that we have are qualifying our prospects automatically, so that works with warm leads, so if they have engaged with us, we were talking about signals, so let’s say they visit our website or they engage on our socials, that’s warm signals, and then we can capture them and filter them, so that’s automated as well.

Or it also works with cold, so if we have a bunch of companies that we are monitoring, that could be also where we are looking for job openings and stuff like that, so we engage at the right time. That’s another one. Another one that I love is my CRM employee, because I mentioned the force bucket, you want to save this in your CRM, you want to keep your CRM clean, so it’s your source of truth.

But it’s annoying to feel it, to do the data entry. Nobody likes it. I’ve been in sales myself, and you’re here to sell, you do a great call, and then you have to fill the CRM manually, that’s terrible. So yeah, this thing is automated. We have tools that are recording our meetings, capturing the transcripts, taking the right data, putting into the CRM automatically. So yeah, that’s another one. Follow-ups as well, when we have the sequence we send messages to reach out to our prospects. If we have a reply, we have agents that are checking the vibe of the message, is it positive, negative? What would be a great response to that? So we don’t have automated responses, but we have drafts, so that’s assisted here. So we have an assistant in the reply management.