Abhord Quickstart Product Guide (Refreshed for January 2026)
This practical guide helps new Abhord users go from zero to running cross-LLM surveys, interpreting results, tracking competitors, and acting on insights. It reflects updates and recommendations since the last edition, including improved normalization across models, cost/latency controls, smarter deduplication, and stronger alerting.
1) Initial setup and configuration
Before you start
- Workspace and roles: Create a workspace and assign roles (Admin, Editor, Viewer). Use least-privilege by default.
- Data connections: Optionally connect your site map or key docs for grounding (RAG). Enable PII redaction for uploads.
- Entity canonicals: Define your brand, products, and canonical domains, plus common misspellings and abbreviations.
Configure your first project
1) Create a Project: Name it clearly (e.g., “Q1 Brand Perception – US”). Add an objective and target audience (geos/languages).
2) Add Entities: Brand, product lines, and a starter competitor set. Include domain(s) and social handles for each.
3) Select Model Panel: Choose a balanced panel spanning at least 3–4 model families to reduce single-model bias. Set budget and latency guardrails.
4) Standards and Baselines: Turn on cross-model normalization and deduplication. Set a 28-day baseline window for trend comparisons.
5) Notifications and Exports: Connect Slack/email for alerts. Enable scheduled exports (CSV or Sheets) and API access for pipelines.
What’s new since last edition
- Cross-model normalization: Smarter per-model sentiment scaling and variance controls for more comparable scores.
- Cost and latency guardrails: Per-run caps with auto-retries and graceful degradation when a model spikes or rate-limits.
- Advanced deduplication: Better clustering of near-duplicate mentions across models and runs.
- Templates: Prebuilt survey templates (Brand Perception, Category Landscape, Purchase Drivers) to start faster.
2) Running your first survey across LLMs
1) Pick a template: Start with Brand Perception