Pilotpilot
Ask a question...
How It Works

Getting Started with Pilot

Getting started with Pilot takes about ten minutes: sign up for the beta, create your tenant, upload documents, configure your voice, and generate your first cited article.

Ten Minutes to Your First Article

Pilot is designed to get you from signup to published content quickly. Here's the path from zero to your first article — the whole thing takes about ten minutes.

Sign Up and Create Your Tenant

Start at pilotwme.com and request beta access. Once approved, you'll create your tenant — a private workspace where your documents, knowledge base, voice settings, and content live. Tenants are fully isolated. Your material is separate from every other Pilot user's material, your voice settings are your own, and your content publishes to your channels only.

You'll need one thing before you start: an API key from your LLM provider (Anthropic or OpenAI). Pilot uses a bring-your-own-key model — you pay the LLM provider directly for the tokens Pilot uses. This keeps costs transparent and puts you in control of which model powers your content.

Enter your API key in Settings under the API Keys tab. Pilot stores it encrypted and uses it only during content generation — never logged, never shared.

Upload Your First Documents

Go to the Library in the Pilot console. Upload a batch of documents — ten is a good starting number, though you can start with as few as three or four.

What to upload first:

  • Your strongest material. Reports, articles, whitepapers, or analyses that represent your organization's best thinking.
  • Focused topics. Documents that cluster around a few related subjects give Pilot enough depth to write substantively. Five documents about the same topic area produce better articles than five documents about five unrelated topics.
  • Any format. PDFs, Word docs, markdown files, HTML, plain text. Upload whatever you have.

Files upload instantly. Processing happens in the background — Pilot reads each document, extracts the content, identifies topics, and maps relationships to other documents in your collection. Depending on the size of your batch, this takes a few seconds to a few minutes.

Watch Your Knowledge Base Take Shape

As documents process, your knowledge base begins to form. Topics emerge from the content — not tags you assign, but clusters of related knowledge that Pilot identifies across your documents.

The Library view in the console shows your uploaded source documents with their status and the topics they cover. You'll see connections between documents that share subject matter, and you'll get a sense of where your knowledge base is deep (many documents on the same topic) versus thin (a single document on a subject).

This is the foundation. Everything Pilot writes draws from this knowledge base and nothing else.

Configure Your Voice

Navigate to Settings and then Voice. Here you'll set the parameters that control how Pilot writes.

Start by adjusting the main sliders:

  • Brevity — tight and concise, or expansive and detailed?
  • Formality — conversational or professional?
  • Confidence — hedged and cautious, or direct and assertive?
  • Citation density — light references or heavy attribution?
  • Tone — warm, neutral, or authoritative?

There are additional parameters for fine-tuning, but these five define the broad editorial posture. The console shows a cost estimate for generating content with your current settings — more expansive, heavily cited content costs more in tokens.

Don't overthink the initial settings. You'll refine them after you see the first few articles. The goal right now is a reasonable starting point. For detailed guidance on each parameter, see Voice Configuration.

Generate Your First Article

There are several ways to trigger content generation, but for your first article, the most direct path is through Pilot Lights.

Check the Pilot Lights page in the console. Based on your uploaded documents, Pilot has already identified content opportunities — topics where your knowledge base has enough material to write something substantive. Each Pilot Light shows the topic, the source documents that would inform the article, and the recommended channel.

Pick one that looks interesting. Click to generate. Pilot reads the relevant source documents, applies your voice configuration, and produces a draft — complete with inline citations linking each claim to its source material.

Review the draft. The citations are clickable — you can verify that each claim traces to a real document you uploaded. Edit if you want. Adjust the wording, restructure sections, remove claims you don't want to make. The draft is yours to shape.

Publish

When you're satisfied with the article, publish it to one or more channels:

  • Web channel — makes the article available via Pilot's API for your website to display
  • CMS webhook — pushes the article as a draft to your WordPress, Contentful, or other CMS
  • Email — formats the article for newsletter distribution
  • Social — creates format-appropriate versions for LinkedIn, X, or other platforms

For your first article, the web channel is the simplest. Publish it, and it's immediately available through Pilot's public API.

What Happens Next

Your first article is just the beginning. From here, the editorial loop takes over:

Upload more documents. Each new document deepens your knowledge base. New topics emerge, and new Pilot Lights surface content opportunities you didn't have before.

Tune your voice. After reviewing a few articles, you'll have a clearer sense of what the sliders should be. Adjust, regenerate, review. The voice settings get closer to your ideal with each iteration.

Set up AutoPilot. Once you trust the output, configure AutoPilot to generate articles on a schedule. Start with supervised mode — drafts queue for your review. Move to autonomous mode when you're confident in the quality.

Add channels. Connect additional publishing surfaces. Push drafts to your CMS, format articles for your newsletter, generate social posts from the same source material.

Monitor costs. The Analytics page in the console shows exactly what each article cost to generate — which model was used, how many tokens it consumed, and the dollar amount. No surprises. See Pricing and BYOK for details.

Going Deeper

For detailed reference documentation, visit docs.pilotwme.com. For the full picture of how Pilot's components fit together, read How Pilot Works. For the ideas behind the system, see What Is a Wisdom Engine?.

Last updated March 3, 2026

Related