How to Write for Humans and AI: The New Content Strategy for B2B Marketing

For years, content strategy has been built around keywords and search rankings, but with the rise of large language models, it's no longer just about ranking in search — it's about being surfaced in AI-generated answers, and that changes quite a lot about how we need to think about content.

How LLMs consume content

Google is built to crawl pages, match keywords, and rank content based on relevance and authority, pointing users to a list of links. Large language models work quite differently, extracting, summarising, and recombining information to directly answer a specific question, and that changes what content needs to exist.

Instead of "best restaurant Sydney," someone might now ask: "What are the best family-friendly restaurants in Sydney for a family of four, with a gluten allergy and picky eaters?" Ranking for broad terms no longer serves your purpose — you need content that speaks directly to specific needs, written in natural language that reflects how people actually ask questions, and structured in a way that makes it easy for an LLM to extract and use. And businesses are already seeing the impact according to a recent BBC report. They found that companies restructuring their content for AI discovery are seeing traffic grow significantly, with those visitors considerably more likely to convert than traditional search visitors.

Does this mean all content now needs to sound like AI?

This is where I think the biggest risk lies. When AI writes all your content, it prioritises clarity and structure which makes it easier for LLMs, but strips out the storytelling, insight, and persuasion that human readers actually need. You end up optimising for discovery while losing the people you're trying to convert.

In order to convert interest into action, you still need narrative to engage people because no matter how well AI surfaces your content, that part remains entirely human.

This means that content now needs to serve two audiences simultaneously:

• AI systems, which prioritise clarity and structure

• Human readers, who need context, insight, and persuasion

The layered content approach

The most effective content teams are shifting toward a layered approach, where different formats serve different purposes:

Answer layer - Direct, concise responses to key questions

Narrative layer - Context, storytelling, and commercial relevance

Authority layer - Technical depth and expert insight

Each layer serves a different reader at a different moment in the decision-making process. A blog post might carry the answer and narrative, a case study the real-world application, and a white paper the technical depth. Having that breadth across formats also helps with AI discovery, as a consistent body of work across multiple contexts signals authority in a way that a single piece rarely can.

This also has a practical advantage for content planning. The same core idea, broken into the right layers and distributed across formats, gives you more coverage with less repetition and makes it easier to maintain a consistent presence across both search and AI discovery.

What this looks like in scientific content

For science companies, this shift is actually quite natural as scientific information already tends to be modular by nature. The key is making sure content doesn't become overstuffed with data, but is instead structured so each layer builds meaningfully on the last.

Take this as an example, it’s the kind of question a life science company might use to guide researchers toward the right product for their application:

When should you use qPCR vs NGS in oncology diagnostics?

1. Answer layer: Direct, concise response to the question

qPCR is best suited for rapid detection of known variants, while NGS enables multiplexed analysis across multiple genomic targets, making it more appropriate for comprehensive tumour profiling.

2. Narrative layer: Context, storytelling, and commercial relevance

For assay developers, this decision is rarely purely technical — it influences workflow design, cost per sample, and time-to-result, particularly in high-throughput or resource-constrained environments.

3. Authority layer: Technical depth and expert insight

This is where deeper data, mechanistic detail, and referenced expertise build credibility with specialist readers — and signal to AI systems that the content carries genuine authority in the domain.

Each layer earns its place by serving a different reader at a different moment, and together they build the kind of content that performs across both discovery and decision-making.

The shift in how we measure success

Content strategy is at an inflection point. The brands that adapt will be those that understand that search has changed and that AI interfaces are becoming the first place people turn for answers. The way to win is not to choose between writing for machines or writing for people, but to do both well. Page views and keyword rankings still matter, but the fuller measure of success is showing up earlier in the decision-making process. The goal is to build authority that grows with every piece you publish.