Skip to main content
Context and knowledge is the repository of information you provide to Arya to contextualise its AI responses to your specific business.

​
What it can contain

Company info

Description, mission, market, main products/services.

Internal terminology

Glossary of industry- or company-specific terms.

Processes and policies

How you handle leads, customer success, escalations.

Templates and tone of voice

Email examples, preferred tone, recurring phrases.

Customer FAQs

Frequent questions + official answers.

Product documentation

Specs, pricing, key features.

​
Why it matters

Without context, Arya responds generically. With rich context:
  • Responses use your terminology
  • Suggestions aligned with your processes
  • Generated emails reflect your tone
  • Correct pricing and product info in quotes

​
How to feed it

Upload PDF, DOCX, TXT, Markdown. Arya indexes them (without using for global training). Max 100MB per doc, 50 total docs (1000 on Enterprise).

​
Context segmentation

Not all context is relevant for every question. Organise by categories:
Used when AryaChat responds on deals, quotes, pricing, objections.
Used for product technical questions, specs, use cases.
Internal policies, HR processes, benefits, welfare.
Admin processes, fiscal deadlines, expense policies.
Arya automatically picks relevant context based on the question.

​
Context quality

More specific and current is better. Avoid:
  • Outdated documents (old pricing, abandoned processes)
  • Contradictory information
  • Corporate manual generality
Prefer:
  • Concrete examples
  • Real customer dialogues
  • Actually-used templates

​
Periodic review

Arya notifies if:
  • A document has > 6 months without edits
  • There are contradictions between documents
  • Glossary terms no longer appear in recent records

​
Privacy

Context documents are visible only within your workspace. Not shared with other Arya customers, nor used for global AI model training.

​
Effective context examples

Document with: ICP, objection handling, pricing flexibility, success examples, industry use cases.
Template responses to common questions, escalation policy, tone of voice for difficult situations.
Feature matrix, technical limits, public roadmap, supported integrations.
Values, communication style, things NOT to say, ethics approach.