The 10% AI Insurance Policy: Why Your "Moonshot" Budget Is Now an Agility Fund

Winnie Chan
6 days ago
3 min read

G’day and welcome back to part 2 of my series on the 70:20:10 Performance Marketing framework.

In the first instalment, we tackled the 70%—the bedrock of your predictable revenue—and the 20%, where you optimise the winners. Today, we dive into the 10%: the traditional "Moonshot" budget.

In B2B SaaS, this is often the space for high-risk gambles, like viral TikTok experiments or flashy, unproven ad formats. Last year, I took a more pragmatic path. I treated that 10% as an AI Insurance Policy.

The Shift: From Awareness to Authority Retrieval

I’m going to call it: Social media isn’t for humans anymore. It’s for training the machines.

While that statement might trigger panic or dismissal, the reality is baked into how Perplexity, SearchGPT, and Google’s AI Overviews actually function. AI-first search engines rely on Retrieval-Augmented Generation (RAG).

RAG gives the Large Language Model (LLM) a "research phase." When a user asks a complex B2B question, the AI scans the live web for fresh, high-authority sources before it synthesises an answer. To be cited as that "Source of Truth," you need more than keywords; you need Traffic Velocity.

The Problem: Micromanaging Chaos at a Global Scale

Last year, content became the centre of the universe for my performance team. We were churning out technical blogs and case studies to build domain authority. However, managing this at a global scale presented a massive hurdle:

20+ Local Markets: Each with unique language nuances and customer behaviours.
Fragmented Schedules: A mess of disconnected organic calendars.
The Control Gap: Micromanaging this from a global seat in Melbourne was a recipe for burnout.

The content was high-quality, but it lacked the immediate "pulse" required for AI search crawlers to recognise it as trending or vital.

The Solution: The "Pocket Money" Pivot

So, I took my 10% "Moonshot" budget and turned it into an Agility Fund.

Instead of betting it on one massive global campaign, I did something radical: I bypassed the global red tape and gave local teams "pocket money" via local credit cards.

The brief was embarrassingly simple:

1. Find a high-value piece of technical content on our main blog.

2. Boost the living daylights out of the social posts directing traffic back to that URL.

The Technical Logic: Manufacturing Information Gain

I know what my fellow performance nerds are thinking:

"Is this just glorified PPC, Winnie?"

Well, listen. This goes beyond traditional PPC. This is about buying Retrieval Signals.

By forcing a sudden, diverse spike in traffic from various social sources back to a specific URL, you are manufacturing the relevance signals that tell AI crawlers: "This is the canonical piece of truth the industry is interacting with right now."

The Checklist: What Makes Content AI-Citable?

The "pocket money" didn’t go to fluffy "thought leadership" about the future of AI. That’s generic noise. It only went to content that demonstrated high Information Gain.

Information Gain is the measure of unique, original insight a piece of content offers that doesn’t exist elsewhere on the web. It’s exactly what RAG systems crave.

In 2026, the LLM Signal Checklist should prioritise:

Unique Data/Points of View: Perspectives that can't be scraped from generic competitors.
High Click-Through from Citations: Proof that your source provides the "next step" for the user.
Recent Velocity: Evidence that the information is current and being discussed now.

The Silos Are Dead

Here is my final reality check, performance lead to performance lead: The silos are officially gone.

There is no more "ask the SEO team" for organic visibility or "ask the paid team" for immediate conversion. If you aren't using your paid media spend to influence what the AI retrieves, you are just burning cash.

We built more domain authority with a few decentralised credit cards and local autonomy than a centralised SEO model ever could.

The 10% isn’t for moonshots. It’s for feeding the machine what it needs to trust you.

About the Author

Winnie Chan is a Performance Marketing Lead and Strategist based in Melbourne, Australia. With over a decade of experience across Agency and In-House roles, she specialises in bridging the gap between creative brand building and commercial targets. She is currently on an "Involuntary Sabbatical" (ask her about it) and writing about the future of marketing.