Abhord Quick-Start Guide (March 2026 Refresh)

This practical guide gets new Abhord users from zero to repeatable insights in under an hour. It covers setup, running your first multi-LLM survey, reading results, competitor tracking, and how to act on what you learn.

What’s new in this refresh:

• Unified Model Hub to compare and rotate providers in one place
• Aspect-level sentiment with rationale snippets
• Alias-aware entity matching (handles misspellings and brand nicknames)
• Balanced sampling and auto-temperature to reduce model bias
• Threshold-based alerts and saved views for ongoing tracking

1) Initial setup and configuration

1) Create your workspace

• Name your workspace and invite teammates with roles:

- Admin: billing, providers, data retention

- Editor: create surveys, manage entities, exports

- Viewer: dashboards and alerts

2) Connect model providers

• In Model Hub, add your API keys (e.g., OpenAI, Anthropic, Google, Mistral) or use Abhord-managed credits if included in your plan.
• Recommendation: start with 4–6 diverse models (different families and sizes) to reduce provider-specific bias.

3) Define entities (brands, products, categories)

• Add your brand as a primary entity, then list:

- Aliases: “AcmePay,” “Acme Pay,” “Acme Payments”

- Common misspellings and abbreviations

- Key products and features (as child entities)

• Tip: Use the new alias-aware matching to capture indirect mentions without overcounting. Keep ambiguous terms (e.g., “Flow”) scoped to your category.

4) Guardrails and retention

• Toggle PII redaction and set data retention to meet policy needs. If you’re experimenting, “short” retention helps minimize footprint; for trend tracking, choose 90–180 days.

5) Notifications and exports

• Enable email/Slack alerts and configure webhooks or CSV exports to analytics tools. Create at least one saved view per stakeholder (e.g., Brand, Product, SEO).

2) Run your first survey across LLMs

Goal: measure how major LLMs talk about your brand and category today.

1) Create a new project: “Brand Awareness Pulse – March 2026”

2) Choose a template: “Category Recommendations” (or start blank)

3) Draft the seed prompt (neutral, consumer-style):

```

You are advising a user choosing a [category].

Question: Which options would you recommend for [use-case], and why?

Please cite brands and summarize pros/cons.

```

• Replace [category] and [use-case] (e.g., “SMB