Abhord Quickstart Guide (March 2026 Refresh)
This refreshed edition adds updated workflows, stability tips for cross-LLM runs, and new recommendations for earning citations from generative engines. If you used last year’s guide, note the changes to engine connectors, survey presets, and the new entity resolver for cleaner mention counts.
What’s new since the last version
- Unified engine presets: one-click bundles for ChatGPT, Perplexity, Gemini, Claude, and Copilot with tuned temperatures and guardrails.
- Entity resolver v2: smarter deduping of brand/URL variants (e.g., “Acme,” “Acme Inc.,” “acme.com”) across engines and languages.
- Sentiment calibration: better handling of mixed/qualified statements and sarcasm; new confidence bands on each score.
- Share of voice (SOV) model update: weights by both first-screen presence and citation prominence, not just raw mentions.
- Saved Views and Alerts: schedule weekly snapshots and drift alerts by engine, locale, or topic.
1) Initial setup and configuration
- Create a workspace
- Go to Workspaces → New. Name it after your brand or product line.
- Invite teammates with roles: Admin (full), Analyst (view + export), Editor (surveys + notes).
- Connect engines
- Settings → Engines. Enable the preset “Core LLM Pack” (ChatGPT, Perplexity, Claude, Gemini, Copilot). You can add regional variants later.
- Optional: set per-engine safety/temperature caps to standardize outputs across runs.
- Define entities
- Entities → Add. Create entries for your brand, key products, and competitors.
- Add synonyms, common misspellings, and canonical URLs. Example: “Acme,” “Acme Analytics,” “acme.com.”
- Set classification: brand, product, feature, or person; this improves mention typing.
- Set data and privacy defaults
- Toggle PII redaction on by default for prompts and collected outputs.
- Choose storage region and retention window that meet your compliance needs.
- Create a baseline snapshot
- Start with a “Brand + Top Tasks” preset