November 12, 2025
The best outbound sequence in the world will not work without a good lead list.
You can have the perfect message, the right follow ups, even AI writing your copy for you, but if you are sending to the wrong people you will not book a single meeting. Outbound lives or dies on one simple question: are you talking to the right people at the right time.
In this article, I will walk through the full process I teach inside our very own AI Outbound Academy. We will cover how to define your ICP before you ever open a tool, where to actually pull high quality B2B leads in 2025, how to clean your list so you do not destroy your domain, and how to load everything into Victoria AI so you can start booking meetings. By the end, you should know exactly how to go from a blank screen to a clean lead list that is ready for AI outbound.
Most people open Apollo or Sales Navigator and immediately start clicking filters. That is the wrong order.
If you do not know exactly who you want to talk to, the tool will happily give you a giant pile of the wrong people. So before you touch any software, you define your ICP on paper.
Start with your perfect customer. Ask yourself who gets the best results from your offer, who you actually enjoy working with, and who can comfortably afford what you charge. Maybe that is SaaS companies that need predictable demos, agencies and consulting firms that live and die on their pipeline, e-commerce brands that want customers instead of more followers, or local service businesses that care about phone calls and bookings. Try to summarize this in one or two sentences. If you cannot describe your ICP in a single sentence, your filters will be all over the place once you get into Apollo.
Next, think about company size. Size affects budget, complexity and the length of the sales cycle. A small team of one to ten people usually gets the most value from done for you services and plug and play systems. Companies with ten to one hundred employees are often ideal for deeper system builds and automation. Once you reach one hundred employees and above, you are usually in enterprise or consulting territory. Pick one primary band per campaign instead of trying to chase everyone at once. A campaign aimed at tiny agencies and global enterprises at the same time usually ends up with mushy messaging that fits no one.
Then decide where these companies are based. You can focus on broad regions such as the United States and Canada, North America and Europe, or narrow it down to specific countries and cities if you want to run a more local play. Location affects language, time zones, regulations and sometimes even how you talk about results. A campaign that targets “Nashville contractors” is going to sound very different than a campaign aimed at “global SaaS teams.”
Finally, think about trigger events. This is where your lead list starts to become powerful. Ask what a company would be doing if they were feeling the pain that your offer solves. Maybe they are hiring salespeople or SDRs, which usually means they need more leads. Maybe they are posting job ads for content writers, which suggests they need content. Maybe they have just raised funding and now have targets and budget. Maybe they are hiring customer support and would be open to an AI voice agent. Maybe they are opening new locations and need aggressive new customer acquisition.
Your trigger event will inform how you later pull leads. If hiring is the signal you care about, you might scrape job boards. If recent funding is the signal, you might pull from funding databases. If local presence matters, you might rely on Google Maps data. Once you answer these questions, you will notice your messaging becomes much easier to write because you finally know who you are talking to, where they live and what they are dealing with.
The second big mistake people make is targeting the wrong person inside the right company.
They build a beautiful list of perfect companies, then send all of their messaging to coordinators and specialists who cannot say yes to anything. The company is right, the person is wrong.
For outbound to work, you want people who feel the pain, own the metric and control the budget. In most B2B lead generation or AI outbound offers, that usually means founders, CEOs, owners, heads of growth, VPs of sales, VPs of marketing, marketing directors, COOs and directors of operations. The exact titles will change a bit depending on what you sell. If your offer is a content engine, you will lean more toward marketing and content leadership. If you are automating customer support or internal operations, you will look more toward operations, service and customer success leadership.
In tools like Apollo or LinkedIn Sales Navigator, this shows up as filtering by seniority at director level and above, then layering in job title keywords such as growth, sales, revenue, marketing or operations. The important part is that you deliberately choose people who can actually pull the trigger and then build your list around them. When you do that, your messaging, your subject lines and your calls to action all become much clearer, because you are speaking directly to someone who can say yes.
Once you know who you want to talk to and who the decision makers are, you are finally ready to open your tools and start pulling leads.
There are several different approaches. Some are simple and fast. Some are more advanced and give you more creative ways to find high intent leads. Here is how the landscape looks right now.
Apollo is still my favorite place to start. Think of it as a giant searchable database of hundreds of millions of business contacts. You can filter by job title, seniority, company size, industry, keywords, location and more.
The main advantage with Apollo is that it gives you everything you need for multi channel AI outbound. You get a verified email address, a phone number and a LinkedIn profile URL. That means from a single export you can send email, connect and message on LinkedIn and layer in voice calls or AI voice agents. In the lesson, I show an example where we filter down from over two hundred million contacts to a focused list of founders and growth leaders at small to mid sized B2B companies in a specific city, then export that into a CSV.
If you are just getting started, this is where you should begin. It is straightforward, the exports are clean and it drops directly into whatever outbound system you are running, including Victoria AI.
Clay behaves like Apollo with a brain attached. Instead of being limited to a single database, Clay lets you blend data from multiple sources and enrich each contact with live web data. You can pull in things like recent LinkedIn posts, latest blog articles, tech stack signals and then feed that into AI for row level personalization.
This is especially powerful if you are running advanced AI outbound because your agent can reference all of this context in its messaging. The tradeoff is that Clay is more complex to set up and usually more expensive than simply pulling lists from Apollo. For that reason, Lesson 2 focuses on Apollo first, then introduces Clay as an upgrade once you are ready for deeper personalization.
Sometimes Apollo does not have the depth you need for a very specific niche. In those cases, there is an old but effective combo that still works extremely well.
You start by using LinkedIn Sales Navigator to build a filtered search of your ideal leads. You then copy that search URL and feed it into PhantomBuster, which scrapes those results out of LinkedIn and exports them as a CSV. At that point you have names, roles, companies and LinkedIn profiles, but you still do not have email addresses. That is where AnyMailFinder comes in. It takes each contact’s name and company domain, tries a few common email patterns and checks which ones actually deliver. What you get at the end is a list where a large percentage have verified, enriched emails.
This route requires more tools and more setup and it is usually more expensive than Apollo. The upside is that it lets you tap into the full power of LinkedIn’s search filters when you need to go very deep into a niche that Apollo does not cover as well.
If you like to think outside the box, scraping platforms can give you some of the highest intent lead lists you will ever work with.
Instead of starting from a generic B2B database, you start from behavior. For example, you can scrape job boards like Indeed for companies that are hiring roles that your offer replaces or supports. If you sell content services, you can scrape job postings for “content writer” or “blog writer” and then reach out with a retainer offer that is cheaper than a full time hire.
You can also scrape Google Maps for local businesses in your niche, Product Hunt for SaaS founders who just launched, or other public directories that your ideal buyers use. Tools like Apify give you a marketplace of prebuilt scrapers where you plug in a few inputs, run the script and receive a CSV of companies or contacts.
Once you have that CSV, the process looks familiar. You enrich the list with email addresses using AnyMailFinder or a similar tool, clean the emails and then load that list into your outbound system. It takes a bit more work than clicking a few filters in Apollo, but it also lets you build lists based on real intent that your competitors may not even realize exist.
The last source that most people overlook is their own existing attention.
You already have people showing interest in you. Your website is getting visitors. People are opting into lead magnets, registering for webinars, booking calls, opening your emails and viewing your LinkedIn profile. All of that is raw lead generation fuel.
You can capture that data, enrich it in a tool like Clay, add missing fields such as role, company size and location, then treat it as a warm outbound list. Those contacts are not strangers anymore. They have interacted with you once. When you feed them into Victoria AI and run them through an outbound sequence, your reply rates and booked meeting rates are usually much higher than with completely cold lists.
Once you have a list, you are still not ready to start sending.
If you start sending emails to a fresh export without verifying anything, you will almost always end up with unnecessary bounces. Enough bounces will damage your sending reputation and make it harder for you to land in the inbox.
Inside Lesson 2, I show how we run our lists through a verification step with a tool like EmailListVerify. The idea is simple. You export your leads from Apollo or any other source as a CSV. You upload that file to the verification tool. It sends test pings to your emails and marks which addresses are safe, risky or invalid. You then remove the invalid and risky ones from your sheet.
Apollo already labels many emails as verified, but I still like to double check with a dedicated verifier. It is cheap insurance for your domain. Once this step is done, you have a list that is essentially double verified and much safer to send to.
Now we finally arrive at the fun part. You have done the hard, boring work that most people skip. You have:
At this point you are ready to load your list into Victoria AI and turn static data into real conversations.
The flow is straightforward. You take your cleaned CSV and open it in Google Sheets or Excel. You make sure the key columns are clearly labeled. You want first name, last name, company name, website, email, phone number, LinkedIn URL, title and any extra fields you want to use in personalization. Once everything looks tidy, you export that sheet again as a CSV.
Inside Victoria AI, you create a new outbound campaign. You go to the leads section and upload your CSV. The platform will ask you to map each column in your file to a field in Victoria. You map first name to first name, last name to last name, email to email, LinkedIn URL to the LinkedIn field, company name, title and any other optional fields you want to use as variables in your sequences. After you confirm the mapping, you upload the list and let Victoria process the leads.
Now your campaign has everything it needs. You are targeting the right people at the right companies with clean emails, valid phone numbers and working LinkedIn profiles!
There are countless ways to “get leads.” You can buy cheap bulk lists, scroll LinkedIn for hours, or cold call random numbers. Or you can do the more thoughtful work of defining your ICP, picking your decision makers, pulling clean data from reliable sources and verifying it before you send.
This article exists so you can stop guessing and start being intentional. Once you understand how to build a real lead list, every part of your AI outbound system starts working better. Your messaging becomes sharper, your reply rates go up, and your calendar fills with conversations that actually matter.
If you have any questions or would like help getting set up, please book a call with us here and we'd love to chat!
Take the next step towards unparalleled efficiency and innovation. Join Raspberry today and experience the difference.
%202.png)
