Abhord Quickstart Guide (Refreshed 2026 Edition)
This practical guide gets new Abhord users from zero to insight in under an hour. It’s the refreshed 2026 edition with updated recommendations for today’s LLM landscape.
What’s new since the last edition
- Broader LLM coverage: survey more commercial and open-source models in one run.
- Smarter clustering: dedupes near-identical answers across models to avoid double-counting mentions.
- Sentiment v2: adds mixed/neutral handling and confidence bands to reduce overreaction to small swings.
- Faster scheduling: recurring surveys and alerting are now first‑class, with cost guardrails.
- Competitor automation: add a list once; Abhord keeps the baseline, alerts on anomalies, and tracks share-of-voice (SoV) over time.
1) Initial setup and configuration
Goal: create a clean foundation so downstream metrics (mentions, sentiment, SoV) are accurate.
- Create your workspace
- Add your brand name, product lines, and canonical entities (company, product, category).
- Enter synonyms and common misspellings (e.g., “Acme Pro,” “AcmePro,” “AP”). Include legacy names if they still appear online.
- Connect models and surfaces
- Choose the LLMs and answer engines you want to survey. Start with at least three to reduce single‑model bias.
- Set budget caps and max parallelism. Tip: cap spend by survey and by day; enable retry-on-failure to mitigate provider hiccups.
- Define intents (recommended)
- Create 5–10 intent categories that mirror your funnel: awareness (best X), comparison (X vs Y), evaluation (reviews, pricing), support (“how to…”), and navigational (brand queries).
- Map each seed query to one intent; this powers segmented SoV and sentiment.
- Governance and access
- Assign roles (admin, editor, viewer). Turn on audit logs if multiple teams will run surveys.
- Configure notifications (email/Slack/webhooks) for critical alerts (e.g., SoV drop >10% week over week).
2) Run your