PostHog

Análise de Dados

Neotask torna o PostHog uma plataforma de analytics continuamente autogerenciada — consultando eventos, gerenciando feature flags e entregando insights de produto por meio de conversa.

O que você pode fazer

With 118 actions — the largest single integration in Neotask catalog — PostHog through OpenClaw gives you complete programmatic control over your entire analytics infrastructure.

Feature Flags (5 actions)

Create, update, and delete feature flags. List all flags with their targeting rules. Manage gradual rollouts por meio de conversa.

Experiments (6 actions)

Create A/B tests, get results, check for statistical significance, update parameters, and run the full experiment lifecycle por meio de conversa.

Error Tracking (6 actions)

List and investigate errors, update issue status, and monitor error trends — connect this to automated alerting so you know about errors before users complain.

Insights & Dashboards (10 actions)

Create, query, update, and manage insights and dashboards. Run HogQL queries for custom analysis. Generate insights from natural language questions.

Surveys (7 actions)

Create, deploy, and analyze surveys. Get global stats and per-survey results automatically.

Data Warehouse (9 actions)

Manage saved queries, materialize views, and run query history — treat PostHog as a full data platform.

Cohorts, Persons, Alerts, Notebooks, CDP Functions, and more — over 100 additional actions for complete analytics automation.

Cada ação é executada de forma autônoma ou requer sua aprovação — você decide.

Tente perguntar

  • "Create a feature flag called 'new-checkout-v2' targeting 10% of users in the United States"
  • "What's the current status of our homepage redesign A/B test — which variant is winning?"
  • "List all errors that have spiked in the last 24 hours and group them by severity"
  • "Create a cohort of users who completed onboarding in the last 30 days but haven't used the core feature"
  • "Build a dashboard showing our top 5 product KPIs with weekly trend charts"
  • "Run this HogQL query: total purchases grouped by country for the last 90 days"
  • "Deploy a survey to users who just cancelled asking why they're leaving"
  • "What insights exist in the Growth team's dashboards that haven't been viewed in 30 days?"

    • Dicas profissionais

  • Schedule automated experiment checks every 48 hours — when statistical significance is reached, get a Slack notification automatically.
  • Connect PostHog alerts to your incident management workflow: when error rates spike, agents auto-create incidents in Linear and notify on-call engineers.
  • Use CDP functions to create automated user scoring — agents can update person properties based on behavior patterns without writing code.
  • Enable controles de aprovação on feature flag creation and experiment launch — these actions have production impact and warrant human review.
  • Equipes multiagente can run full product analyses: one queries PostHog data, one checks Amplitude for business impact, and one summarizes findings for the weekly product review.

    • Works Well With