Atomic Content

Niche Content at Scale

The idea of Content Atomization isn't new, but up until this point has been largely of manual process of taking a single piece of content and from it creating many variations, channels, media, & schedules.

AI gives us unprecedented opportunities to hyper-automate the generation of niche content at scale across creative, video, articles, landing pages, emails, & social posts.

Why Atomic Content:

• Go beyond category level messaging
• Resonate with EVERY customer niche
• Automatically generates all niche variations
• In your brand style and persona
• Atomized for every distribution channel (Email, Linkedin, YouTube, TikTok and more)

Atomic Content Use Cases:

• Thought Leadership of the Day
• Newsletter
• Product or Service promotions
• Sales Outbound
• Customer Success Onboarding
• Digital Influencer Channel
• Conference & Event Announcer / Translator

Experiments:

MM6: Talking Heads

Goal: Test Autonomous Influencer Marketing with a Synthetic New Anchor
Objectives:

1. Generate affiliate $$$ by orchestrating digital influencers
2. Thread multiple inputs and outputs together at a speed humans cannot.
3. ​Sequence multiple A.I. tools (news trends, scripting, character generation, omni-channel posting) together to achieve objectives
4. Consume conversion data to iterate improvements

Results:

• Processes trending topics
• Writes a new script
• Generates a synthetic new anchor to read the script
• Generates a video short with a featured image of the news story and the influencer
• Posts to YouTube
• Post to TikTok

https://www.youtube.com/channel/UCRehYeMNJgp4NykxOWVU7jg

https://www.tiktok.com/@mattbuenoma

‍

MM7: Code 19 Maveric:

Code19 Autonomous AI Racing Team

Goal: Create an AI Driver Persona named Maverick

Objectives:

1. Automate social content and connect with fans
2. Generate sponsorship $$$ via AI Driver impressions
3. In-car interviews and real-time race commentary

Results:

• From Design Sprint to POC in 1 week
• AI Persona watching live F1 stream and commentating in real-time

Deep Dive:‍
This is the 1st test of "Maveric", the autonomous driver personality for the Code19 racing team in the Abu Dhabi Autonomous Racing League (A2RL).

This simulation used the Abu Dhabi Grand Prix (an F1 race) as a proxy for what Maveric would be capable of executing during a live A2RL race.

She is a professional race car driver and broadcast personality. In this test she is tasked with consuming any content being published by F1 related to the live Abu Dhabi Grand Prix in real-time, and watching the live broadcast of the race on ESPN to consume real-time action and commentary about the race.

As new information flows into her "Feed", she internalizes these data into her knowledge set and generates new commentary based not just the new bits of info - but in context with the AD GP as a whole, including her background data on the league, the Yas Marina Circuit (race track), all previous seasons & the current season's schedule / results / standings, the entire history of F1 racing, the bios of all drivers and constructors (car manufacturers), and Code19 sponsors.

In near real-time, new information is consumed and new commentary is generated. This commentary is ran through an audio generation pipeline to speak Maveric's commentary using her own voice and streamed to the online user experience.

Key Takeaways / Next Steps from this 1st test:

- Apply a custom voice for Maveric to provide better tone & inflection
- Adjustments in her instruction set to make her commentary more confident, halting use of starting qualifiers such as "It appears", "It seems", "It sounds like"
- Replace batching of new Feed item insertion into knowledge with streams to get even closer to real-time stream-of-thought.
‍

‍
‍