Pilot Lights

An Editorial Meeting, Automated

Every content operation faces the same question: what should we write next? Traditionally, the answer comes from editorial meetings, brainstorms, SEO keyword research, or someone's gut feeling about what the audience wants.

Pilot Lights answer that question differently. They look at your knowledge base — the actual documents you've uploaded — and identify topics where you have enough source material to write something substantive. Each Pilot Light is a content opportunity: a topic, the source documents that would inform the article, and a confidence score indicating how well-covered the topic is.

Think of it as an editorial meeting where Pilot shows up with a list of story ideas, and every idea comes with its research already done.

How Pilot Lights Work

As documents enter your knowledge base, Pilot continuously maps the topics they cover and the depth of coverage in each area. When a topic cluster reaches sufficient depth — enough source documents with enough relevant material — Pilot surfaces it as a Pilot Light.

Each Pilot Light includes:

• The topic — what the article would be about
• Source documents — which documents from your knowledge base would inform the article
• Recommended channels — which of your configured channels the article is a good fit for
• Discovery type — how the opportunity was identified (new topic cluster, deepened coverage, cross-topic connection)

Pilot Lights update as your knowledge base changes. Upload new documents on a topic, and existing Pilot Lights for that topic may strengthen (more sources available). Upload documents on an entirely new subject, and new Pilot Lights appear.

Editorial Prompting, Not Automated Publishing

Pilot Lights suggest. You decide.

When a Pilot Light appears in the console, you have several options:

• Generate — tell Pilot to write the article. It draws from the identified source documents, applies your voice configuration, and produces a draft for your review.
• Save for later — bookmark the opportunity. It stays in your queue until you're ready.
• Dismiss — not interested. Pilot won't resurface this specific opportunity.

No Pilot Light results in a published article without your involvement (unless you've configured AutoPilot in autonomous mode, which is a separate, deliberate choice). Pilot Lights are the editorial intelligence layer — the system telling you what it can write well. The editorial judgment about what should be written remains yours.

What Makes a Good Pilot Light

Not every topic in your knowledge base becomes a Pilot Light. Pilot evaluates several factors:

Depth of coverage. A topic with one source document that mentions it in passing is thin. A topic with six documents that discuss it substantively is rich. Pilot Lights surface when coverage crosses a threshold where the resulting article would be well-sourced, not speculative.

Source variety. An article drawing from a single document is a summary. An article drawing from multiple documents is a synthesis. Pilot Lights favor topics where multiple source documents contribute different perspectives, data points, or angles.

Distinctiveness. If you've already generated an article on a topic, Pilot won't keep resurfacing it unless new source material adds enough fresh content to warrant another article. Pilot Lights aren't a broken record.

The Feedback Loop

Pilot Lights create a feedback loop between your knowledge base and your content output:

1. You upload documents
2. Pilot identifies content opportunities (Pilot Lights)
3. You generate articles from the strongest opportunities
4. You notice which topics produce the best articles
5. You upload more documents in those topic areas
6. Deeper coverage produces richer Pilot Lights

This loop means your content operation improves over time. The more you feed the knowledge base in areas that produce good content, the better Pilot gets at surfacing opportunities in those areas. It's a compounding return on your knowledge investment.

Pilot Lights vs. Other Approaches

Vs. keyword research: SEO keyword tools tell you what people are searching for. Pilot Lights tell you what you can write well. These overlap sometimes and diverge sometimes. The SEO opportunity might exist, but if your knowledge base doesn't cover the topic, Pilot won't suggest it — and any article you forced it to write would be thin.

Vs. editorial calendars: An editorial calendar is a plan. Pilot Lights are raw material. They inform the calendar rather than replace it. An editor might look at the week's Pilot Lights and decide which three to generate, in what order, for which channels — that's editorial calendar work, informed by the knowledge base.

Vs. content gap analysis: Some content strategy tools analyze your published content and identify gaps. Pilot Lights work from the other direction — they analyze your source material and identify what you could publish. The difference is between "what have we not written about?" and "what do we know enough to write about?"

In Practice

Builder's Benchmark, a demonstration tenant built on roughly a thousand construction research articles, generates dozens of Pilot Lights across topics like structural engineering, energy efficiency, construction safety, and building materials. An editor looking at those Pilot Lights can see at a glance which topics are rich (many strong source documents, high confidence) and which are emerging (fewer sources, lower confidence).

The editorial decision is immediate and informed: here's what the knowledge base supports, here's how strong the sourcing would be, here's which channel it fits. Generate, review, publish.

For how to act on Pilot Lights and generate your first article, see Getting Started. For how to automate generation from Pilot Lights, see AutoPilot.