Responsible intelligence

Ethics built for
politics and markets.

VisionHQ sits beside sibling initiatives in the broader ecosystem, but is its own product: enterprise buyers should expect independence, restraint, and documentation—not vanity metrics dressed as certainty.

Principles

Credibility is a feature—not a footnote.

Political clients need defensible neutrality. Market clients need discipline around material non-public information and hype cycles. Our ethics program names the overlap and the differences up front.

Permitted sources only

We work from public channels, licensed feeds, client-supplied datasets, and APIs that explicitly allow the use case. If a source requires consent, that consent sits with you—and we document the chain.

No platform games

We do not bypass authentication, rate limits, bot defenses, or terms of service. If a platform does not permit a workflow, we redesign the workflow instead of sneaking around it.

Methods you can inspect

Sampling, weighting, NLP choices, and alert logic are described in plain language. Black-box scores without provenance are not how we charge for enterprise work.

Retention with an exit

Active engagements retain data for operation and reconciliation. After offboarding, client-held copies are purged on a fixed schedule unless law or a written legal hold says otherwise. You may request accelerated deletion where compatible with audit duties.

What we do

  • ✓Ingest public posts and articles through compliant interfaces where available.
  • ✓Process survey or CRM extracts that you certify you have the right to share.
  • ✓Aggregate and anonymize individual-level inputs before they enter shared models.
  • ✓Log model versions, prompts (where applicable), and data lineage for review.
  • ✓Separate political intelligence from market product logic so incentives stay clear.

What we refuse

  • ×Harvest private messages, closed groups, or non-public inboxes without authorization.
  • ×Operate sock-puppet or disguised accounts to manipulate platforms or people.
  • ×Sell or merge identifiable voter or customer records with unrelated third parties.
  • ×Present probabilistic outputs as guarantees, official results, or personalized voting instruction.
  • ×Offer stealth scraping as a “premium” lane when APIs already forbid the behavior.

Political independence

VisionHQ does not endorse parties, candidates, or ideologies. Analytical choices are applied uniformly; clients do not receive “custom math” that flatters a predetermined storyline. Outputs are simulations and estimates with documented uncertainty—not official election results or voting instructions.

Market intelligence boundaries

Signals surfaced in dashboards and feeds are for research, risk sensing, and narrative monitoring—not personalized investment advice. Teams remain responsible for compliance with securities law, insider-trading policy, and internal trade rules. We help you see chatter early; we do not tell you how to trade.

For how models are constructed, see Methodology. For commercial terms, see Pricing or speak with our team.

Book Technical Briefing