Skip to main content
This guide shows how to get the most out of Pria with precise prompts, complete context via IP Vault, verifiable outputs with Agent Details, and repeatable workflows using Assistants. It aligns with Praxis AI’s operating model and links to relevant documentation where applicable.

​
At a glance

1- Be specific

Craft precise, bounded prompts with clear deliverables, tone, and constraints.

2- Use IP Vault for websites

Scrape full sites to give Pria clean, complete context instead of noisy partial HTML.

3- Verify with Agent Details

See which agent ran, sources used, success status, and returned data.

4- Manage conversations

Create/choose the right conversation to preserve context and retrieve history.

5- Set perspective

Tell Pria which role to adopt (e.g., “AI engineer”) for audience-aligned output.

6- Create Assistants

Automate multi-step workflows with clear, ordered instructions and error handling.

7- Apply CRISPE

Design robust Assistants with a structured capability and context template.

8- Iterate & validate

Use Agent Details to identify gaps; refine prompts or sources accordingly.

9- Stay organized

Name conversations and Assistants consistently for easy retrieval.

10- Respect governance

Follow data, privacy, and safety rules; prefer IP Vault and RAG over ad-hoc copy-paste.

​
1 Be Specific: The power of detailed requests and context

Specificity directly improves precision and reduces rework. Define objective, audience, format, constraints, and examples.
1

State the objective and audience

What should be produced and for whom (e.g., teachers vs. executives).
2

Constrain format and scope

Word count, headings, components, style, and what to exclude.
3

Provide source context

Attach RAG/IP Vault content or quote key excerpts.
4

Define success criteria

Checks, tests, or acceptance conditions.
Example prompt
Objective: Draft a 500–600 word announcement for small businesses about Q1 SEO trends.
Audience & tone: Conversational yet authoritative; non-technical.
Format: 4 sections (Overview, Trend 1, Trend 2, Actions); include 3 actionable bullets.
Constraints: Cite any stats inline; avoid jargon like "synergy"; no images.
Sources: Use my IP Vault document for "AcmeSEO Blog Q1 2025" and my RAG file "Search Trends Notes".
Success: Clear 90-day action plan and linked glossary terms.

​
2 Website research via IP Vault: Comprehensive scraping

Why IP Vault?
  • Full-context capture: Pulls complete site content for deep analysis, not just partial HTML.
  • Clean data: Filters out extraneous JS/CSS and delivers structured text.
  • Repeatable: Documents are reusable across conversations and Assistants.
1

Create a site snapshot file

Add the site to IP Vault and complete the scrape.
2

Name and tag consistently

E.g., VendorX Docs (2025-10-02) | pricing, api, auth.
3

Reference documents explicitly

Tell Pria which file to use by name.
4

Limit scope as needed

Ask for specific sections to avoid noise.
5

Refresh on updates

Re-scrape periodically when content changes.
Use IP Vault: "Compare pricing tiers from 'VendorX Docs (2025-10-02)' and map features to our needs.
Avoid direct URL reads; use my document only."

​
3 Agent Details: Unveil process and verification

Agent Details shows which agent executed, whether it succeeded, which sources were used (RAG, IP Vault, browser), and what data was returned.
  • Agent identity and version
  • Success/failure status and error messages
  • Source breakdown (RAG sections, IP Vault documents, web fetches)
  • Returned artifacts (tables, JSON, files)
  • If sources are missing: attach or reference the right RAG/IP Vault items
  • If a step failed: retry with narrower scope or corrected parameters
  • If inference is high: ask Pria to separate “retrieved facts” vs “model inferences”

​
4 Manage conversations: Organize context and history

1

Start a new conversation per topic

Prevents cross-topic contamination.
2

Pick the correct conversation before asking

Ensures relevant history is available.
3

Name conversations descriptively

E.g., “Mintlify Writer — Pria Best Practices”.
4

Leverage ‘Clear selected’ to broaden history

Allows Pria to draw across prior threads when appropriate.
If something “disappears,” check a different conversation thread. Context is scoped per thread by design.

​
5 Use perspective: Define Pria’s role explicitly

“As a social media strategist, create a 7-day content plan prioritizing short-form video
for small retailers. Include hooks, CTAs, and platform-specific notes.”

“As an AI engineer, outline a Node service to summarize PDFs using our IP Vault documents.
Include architecture, error handling, and timeouts.”
State domain, deliverable, constraints, and evaluation criteria in the first message to lock in the perspective.

​
6 Create Assistants: Automate and optimize workflows

Assistant core pattern Assistants turn reliable processes into repeatable, multi-step runs with consistent quality. 
You are the "XXX" assistant and your role is to create YYY by following the steps below:

1- Collect input from user through conversation
2- Analyze the data
3- Generate your response
4- End of job

Error Handling:
A- Not enough or invalid input data: Continue or restart at step 1
B- Off topic: Redirect off topic back to the focus of this assistant
Keep steps atomic, verifiable, and source-aware (RAG/IP Vault names). Add logging expectations and acceptance checks.

​
7 Error handling patterns for Assistants

Validate inputs

Check required fields; ask targeted follow-ups.

Constrain sources

Prefer specific RAG/IP Vault documents over free web.

Plan-then-execute

Have the Assistant outline its plan before executing.

Degrade gracefully

Return partial results plus a “next steps” list.

​
8 Advanced design with CRISPE

Why CRISPE? A structured methodology to define Capabilities, Role, Inputs, Steps, Policies, and Evaluation for robust, auditable Assistants.
1

Capabilities & constraints

What the Assistant can and cannot do.
2

Role & audience

Who it is, who it serves.
3

Inputs & sources

Required fields; named RAG/IP Vault assets.
4

Steps & checkpoints

Atomic tasks with intermediate validations.
5

Policies & safety

Privacy, redaction, and compliance boundaries.
6

Evaluation & QA

Acceptance tests and quality thresholds.
Learn more: Building Assistants with CRISPE

​
9 Iterate and validate with a tight loop

“Run with sources A/B. In Agent Details, show: (1) retrieved passages,
(2) any inference steps, (3) unresolved items. If missing, request the exact inputs needed.”
Treat outputs as drafts until Agent Details confirms the right sources were used and success criteria are met.

​
10 Governance, privacy, and safety

Use IP Vault documents and RAG files instead of raw copy-paste or unvetted links.
Follow your organization’s privacy and safety policies; restrict sensitive content and verify recipients for any exports.
Maintain human review for outputs impacting decisions, grades, compliance, or policy.
For institutional commitments and protections, see your Praxis AI DPA and internal governance policies.

​
Putting it all together

In practice:
1

Start a correctly named conversation
2

set a clear perspective
3

select precise files from IP Vault
4

state objective, constraints, and success criteria
5

run
6

inspect Agent Details
7

refine inputs or steps
8

repeat until acceptance tests pass
9

templatize into an Assistant with CRISPE
10

add fallback for edge className.
If results feel generic, raise specificity (objective, structure, constraints) and narrow sources (IP Vault/RAG). If results feel incomplete, check Agent Details for missing or failing steps and adjust accordingly.