Turn Your Small Marketing Team into an AI-Powered Growth Engine
August 27th, 2026
This session is for founders, executives, and small marketing teams who want AI to do more than write captions. Bernard breaks down the exact system he used to generate over six million impressions and 300,000 organic clicks last year: how to train Claude and ChatGPT to sound like you, capture ideas before they disappear, and turn one idea into a written post, a video, an image post, a voiceover, and a carousel. Every person who watches this session will walk away with a clear way to connect sales, marketing, and content into one AI-powered feedback loop.
SUMMARY
Table of Contents
-
Small Teams, Big Leverage
-
Capture Ideas in One Place
-
Training vs. Prompting: Teaching AI to Sound Like You
-
What You Feed the Machine: Personality, Music, and the Shows That Move You
-
One Idea, Five Pieces of Content
-
Connecting the Feedback Loop with MCP
-
Getting Past the Cringe
-
What to Do Next
1. Small Teams, Big Leverage
Bernard opens with his thesis: small marketing teams can now beat large ones, not because of budget, but because of how well they use AI, how fast they make decisions, and how quickly they implement without waiting for layers of approval.
"Teams can now beat larger ones, not because they have bigger budgets, because they can use AI better, make decisions faster, and implement without waiting for layers of approval."
He points to the gap most companies never close: valuable insight is trapped inside sales calls, marketing meetings, Zoom recordings, Slack threads, and people's heads. Sales hears customer questions and objections every day, but marketing often gets that information late, or not at all. Connecting sales transcripts, customer questions, company knowledge, and performance data through Claude and MCP closes that gap and creates a feedback loop showing the team what customers actually care about and what to create next.
"Your team is not slow, but your process is."
Bernard shares his own numbers as proof: last year he generated more than six million impressions and over 300,000 organic clicks using this system, posting 2,200 times in a single year. This year, he's on track to double it.
2. Capture Ideas in One Place
Bernard's system starts simple: a notepad. Just capture ideas throughout the week, especially the ones that make you feel something. If an idea moves you, it will probably move your prospect too.
He uses Whisper Flow to voice dictate ideas on his phone, which also cleans up the language automatically. He pairs it with Granola, which records and transcribes meetings, but also captures thinking on the fly, whether that's a strategy conversation with his wife on the school drop-off run or an idea from a client call. Once it's recorded, it lives in the AI database, ready to be used later.
"Intention is in our pockets. You can open up this phone, turn on the camera, go live stream. You can go on video. You can reach hundreds, thousands, if not millions of people for free."
3. Training vs. Prompting: Teaching AI to Sound Like You
Bernard's central argument is that most people and most teams spend all their time on prompting: trying to ask the right question the right way. He believes the real leverage is in training the LLM instead, teaching Claude and ChatGPT your likes, your dislikes, and your language, so it becomes an extension of you rather than a generic tool.
"It's about training the LLM. It's about training the ChatGPT. It's about training Claude on your likes, your dislikes, your language."
The mechanism is MCP. Bernard connects Claude and ChatGPT to Granola, Whisper Flow, Descript, and Vista Social through Settings, then Connectors. Once connected, the AI can look back through his recordings and past posts and write new content in his actual words, not a generic approximation of them.
"I ask it to write things in my voice, what I actually said in my recordings. That way, whatever I write, I've said it."
His instruction to the AI is specific: pull from the recordings, find where he's already said something close to what's needed, and write the post from that; never invent new wording.
4. What You Feed the Machine: Personality, Music, and the Shows That Move You
Bernard goes a layer deeper than most people would think to. He recommends taking personality assessments (Kolbe, Myers-Briggs, DISC, Colors, and Wealth Dynamics) and feeding the results into Claude and ChatGPT, so the AI understands not just what you say, but how you think and operate.
Then he adds another layer: music. He keeps a Spotify playlist of songs that move him and drags that into his AI training the same way Spotify itself uses those likes to learn his taste. He does the same with the movies and shows that shaped his sense of design and story, citing Shogun and Samurai Rabbit as direct influences on his current visual brand.
"In order to stand out in a feed, you have to stand out. That's why these designs are very curated."
His point: everyone has access to the same AI tools. The only way your output looks different from everyone else's is if it's actually built from you.
"The only way it will stand out is from you. We are one out of billions and will never be created again the same."
5. One Idea, Five Pieces of Content
This is the mechanism behind the 2,200 posts. Bernard starts with one written post. He records himself reading it on camera using Descript, which turns it into a video. He takes a screenshot from that video and lays text over it, which becomes an image post. He strips the talking-head footage and layers B-roll over the audio, which becomes a voiceover. Finally, he drags the transcript into Claude or ChatGPT and asks for a five-to-seven-slide carousel.
One idea. Four or five formats. All from a single sitting. He then reposts that same set of assets the following quarter, and the quarter after that, since most of the audience hasn't seen it the first time around.
"For those interested in making one piece of content into multiple pieces of items, that's exactly how I posted 2,200 times last year. I'm not gonna sit here and create 2,000 pieces of content. I'm gonna create one content and turn it into multiple pieces."
6. Connecting the Feedback Loop with MCP
Bernard walks through the mechanics simply: go into Settings in Claude or ChatGPT, hit Connectors, and connect Granola, Whisper Flow, Descript, and Vista Social. Once it's connected, you can ask the AI to look back through a year of posts and explain why the top performers worked, then generate new posts based on that same framework, pulled from your own recordings.
"Please take a look at my past year's post and give me the framework of why the top post did well... make sure that you look through my Fathom recordings, my Granola recordings, and my Whisperflow, and use my language and create a new post based on the framework of the past performing post."
Because Vista Social connects to LinkedIn, Instagram, YouTube, and Facebook, the whole loop becomes self-reinforcing: post, measure what worked, feed that back into the AI, and let the next batch of content improve automatically.
7. Getting Past the Cringe
Bernard addresses the fear directly, especially for people who feel like posting isn't natural to them. He admits he felt the same way for years, which is part of why he barely posted at all before 2024. His reframe is simple: make the pain of regret bigger than the pain of the cringe.
"What if we made the pain of regret larger than the pain of the cringe? It will only be cringe for a little bit, and then it won't be cringe when it's working."
He points to his own first videos as proof that no one starts polished. Nobody liked the first ten. His mom liked the sixteenth. His brother liked the seventeenth, after being asked to. The twenty-eighth took off. The only way through the early, awkward stretch is reps.
"People don't remember what you said. They remember how you made them feel." — Maya Angelou
He closes the content portion of the talk with a challenge to create more than you consume.
"You need to fight for creativity like it's a fight for the knife in the mud. Because when you grab that blade and you wipe it off, you will cut through all the noise."
What to Do Next
Bernard is hosting an in-person, hands-on workshop on September 4th in Irvine, California, at CoLab Space, from 9:00 AM to 3:00 PM. Seats are limited to a small conference room, and the price is under $300, since the goal is meeting people in person rather than running it as a profit driver.
In the workshop, attendees will set up their AI tools, clarify their positioning, and turn their expertise into content with the help of a professional videographer on-site to shoot and edit a one-minute pitch video live. The day also covers building a content creation system, a brand and design kit using Claude Design, and connecting MCP and Vista Social to turn captured ideas into multiple pieces of content.
For those not able to attend in person, Bernard posts consistently on LinkedIn and runs regular LinkedIn Live sessions built around the same core idea:
"Your presence is your strategy. Find your voice, be memorable, and share it at scale."
The best next step is to follow Bernard on LinkedIn, drop a comment on the replay with what you're going to walk away and do differently, and connect with others who were in the room.
WATCH NEXT
Stop Trying to Find Your Voice.
Create It. Share it at Scale.


