Automated Influencer & Content Creator Outreach

Why Video Content Creators?

If you're anything like me, you find marketing difficult and annoying. A large part of it is not wanting to annoy people who don't care about the thing you're marketing. Everybody online today is constantly bombarded with offers. Ads are expensive, they often don't work, and most people hate them.

Given that, the most ideal form of marketing would happen when only the people who actually care about the thing, see the offer. That way you don't have to trouble anybody who is not going to be interested, and only the people who are actually interested see your pitch.

There is one form of marketing that fits this mold, and that is video content creators who make videos or stream on a particular topic. Imagine for a minute you're super into a specific type of video game. Maybe it's 90s first person shooters. Probably what you'd do is seek out content creators who play games in that genre. You actually want to see those games because you like them and want to play more of them.

Now imagine you are somebody making 90s FPS games. It makes complete sense to get your game to streamers who stream those types of games, so that they can show them to their audience who love those types of games. When that happens it's a win-win-win and a perfect discovery loop.

The same is true of product content creators of all kinds. People who vlog about elite cookware, or camping gear, or knitting. People watch those content creators because they want to know more about those topics, including the best products and services available to help them do the thing.

The problem for a producer then becomes, how do you reach those creators so that they can reach the people who want to see your product and offer? You can reach out to popular streamers one by one, but that in itself is time consuming and these people are busy, with more than enough products in their review queue. Often they charge money and for popular streamers that can get expensive fast.

There's a hack though. What many people don't realize is there's a hierarchy among content creators. You have a few people at the top with hundreds of thousands or millions of views, and the pyramid gets wider as you go down, with more and more people with less and less views on their videos.

People further up the hierarchy are themselves enthusiasts and they will watch videos from people both above and below them on the content creator hierarchy. So topics and products will migrate among creators in both directions. You can use that hierarchy to get your product out there by sending it to less popular creators further down the hierarchy who are easier to reach, and their reviews will be seen by more popular creators, who will then make videos on the same topic.

So now you need to get your product in front of many smaller content creators at scale. This is exactly the problem I solved in this project using agentic AI tied together with some custom Python scripts. We used an agent for both the discovery phase and the outreach phase.

The Creator Outreach Pipeline

In the discovery phase, the goal is to uncover the set of content creators who are most likely to be interested in your product. We used a set of agentic Python scripts in three steps:

1. Initial Dataset - We started by manually compiling a list of similar products in the same category, and then built a simple agent to search across YouTube for each of those products to find content creators talking about them.
2. Metrics & Data Extraction - Next the agent used the free and open YouTube RSS feeds for each creator to analyze their played titles and extract style tags from their plain text descriptions.
3. Algorithmic Scoring - The agent also built a score for each streamer using fuzzy matching against our set of similar products, basically answering the question "how often do they create content about products like ours?"

After this initial discovery phase we were then ready to move on to the outreach phase. The script ranked every streamer by how often they streamed about similar products, their view counts, and last recording time and then came up with a relevancy score. After that we had a separate agent draft an email tailored to the particular content creator designed to showcase the aspects of the product that most applied to them.

Critically, this step is 100% human-in-the-loop. Rather than blasting out spammy AI slop emails, the agent instead connects to an IMAP mailbox and drops in a carefully curated draft for human review. At the top of each draft the agent injects a clean summary of important details and context about the creator, curated from their description and videos.

This summary lists their bio, when they last uploaded, their five most recent video titles, and an LLM-suggested personalization draft. This eliminates tab switching and digging through YouTube accounts to find those details. The human then opens the draft in their regular email client, spends 10 seconds reviewing the dossier, edits it, deletes the summary, and then writes a quick authentic greeting tailored to that specific content creator. The goal is not to deceive anybody but to find people who would legitimately be interested in the product and present it in a way that makes it easy for them to see why.

Creator Marketing at Scale

In the end this system was able to send out hundreds of highly personalized, slop-free emails that were carefully checked and edited by a human being. We also included a coupon for free access to the product as a thank you for content creator time and attention. The bottom line is actual content creators made videos about the product which received thousands of views on YouTube.

If you're interested in similar automated outreach and marketing workflows, please feel free to get in touch.