Aeonav
  • Pricing
  • Changelog
  • Roadmap
  • Vision
  • Docs
GitHubSign up
IntroductionQuick Start
OverviewProvidersCLI LabRegistration & LoginBrand SetupPrompt WizardDashboard OverviewVisibility TrackingSentimentCitation AnalysisResearch AgentSettingsReportsAgent Readiness Audit
Edit this page·View as Markdown·Report an issue·Join our Discord

Get product updates

New features and releases from Aeonav, straight to your inbox.

Product

  • Features
  • Pricing
  • Off-Site AEO
  • Changelog
  • Roadmap

Resources

  • Blog
  • Documentation
  • Support
  • API Reference
  • Provider Status
  • Issues

Learn

  • What is AEO?
  • What is GEO?
  • AEO Glossary
  • AI Search Guides
  • AI Search Statistics
  • AEO by Industry

Tools

  • Tool Directory
  • Free & Open-Source Tools
  • Compare Tools

Company

  • About
  • Vision
  • Brand Assets
  • Contact
  • llms.txt

© 2026 Aeonav

TermsPrivacyCookiesSubprocessorsAcceptable Use
v0.7.8GitHub repositorysupport@fivecoder.com

Research Agent

Turn a natural-language SEO objective into an evidence-backed report or product brief.

What It Does

The Research Agent accepts a plain-language objective, such as “Find fast-growing US keywords for a web app that helps small teams forecast cash flow, excluding physical products.” Aeonav translates that objective into a bounded research plan, measures search demand, checks Google rankings, and drafts a report from the collected evidence.

Use it when you know the outcome you need but not which specialist SEO tool to start with. It is designed for web apps, software, and service businesses. It is not a general-purpose browser and does not click through arbitrary sites on your behalf.

Start a Run

  1. Open a brand and select SEO Research → Research Agent.
  2. Describe the audience, market, product type, and any exclusions. Be specific about constraints that should change the result.
  3. Choose Report Only, PRD Only, or Report + PRD.
  4. Select Start Research and leave the page open to follow progress.

The run continues in the background. You can return later and open it from Recent Research.

What the Agent Measures

The current release uses two DataForSEO operations:

  • keyword suggestions for up to three seed terms;
  • Google organic SERP rankings for up to three high-demand keywords.

A run is capped at eight paid provider calls. Empty or failed sources are marked as unknown or partial; they are not treated as proof that demand does not exist.

The agent does not crawl arbitrary websites, run backlink or full site-audit checks, publish content, send outreach, or make purchases in this release.

Evidence and Reports

Every measured conclusion links to an evidence record with its provider, operation, query, observation time, excerpt, and structured data. Inferred statements are labeled as inferred and do not cite measured evidence they do not have.

Open Evidence under a completed run to inspect the source data. If a provider call fails but other evidence is available, the run can complete as partial and will show the failures. If no usable evidence is returned, the run fails instead of generating a fabricated report.

Reports and PRDs

Report mode includes a summary, findings, opportunities, a 30-day action plan, risks, and unknowns. PRD modes add problem framing, target users, scope, requirements, acceptance criteria, and open questions.

Reports are tenant-scoped. A member of another organization cannot list, open, or cancel your run, even if they know its identifier.

Troubleshooting

  • The run stops during planning — Check that the configured structured-research provider is available and that its key is valid.
  • No evidence was returned — Check the DataForSEO credentials and account balance, then try a shorter seed term that describes a software category.
  • A run is partial — Open the evidence drawer and review the failures. The report remains limited to evidence that was actually collected.
  • Start Research stays disabled — Enter at least ten characters and wait for the page to finish loading.

Was this page helpful?

On this page

What It DoesStart a RunWhat the Agent MeasuresEvidence and ReportsReports and PRDsTroubleshooting