Chat with AI

Overview

Solatis Chat with AI is your intelligent conversational assistant powered by advanced language models (GPT-4, Claude). Ask questions, analyze content, and get insights from your documents in natural language.

Getting Started

Starting a Conversation

Method 1: Quick Start

Click "New Chat" button (top navigation)
Or press N keyboard shortcut
Start typing your question

Method 2: From Document

Open any document
Click "Chat" tab or button
Ask questions about that specific document

Method 3: From Dashboard

Recent conversations widget
Click "Continue" on existing conversation
Or "New Conversation" button

Your First Conversation

Example Conversation:

You: What is this document about?

AI: This is a quarterly financial report for Q3 2024. 
It covers:
- Revenue performance ($2.5M, up 15% YoY)
- Major expense categories
- Market expansion results
- Forward-looking projections for Q4

The report was prepared by the Finance team and 
includes detailed breakdowns by region and product line.

Follow-up Questions:

You: What were the biggest challenges mentioned?

AI: The report highlights three main challenges:

1. **Supply Chain Delays** - Increased lead times for 
   key components, affecting product delivery timelines
   
2. **Currency Fluctuations** - Exchange rate volatility 
   in European markets impacting profit margins
   
3. **Increased Competition** - New market entrants in 
   the SMB segment putting pressure on pricing

Source: Section 3.2 "Operational Challenges", pages 8-10

Chat Interface

Layout Components

Chat Window:

┌─────────────────────────────────────────────────────┐
│ Financial Report Q3 Analysis    [Context] [⚙️ ]    │
├─────────────────────────────────────────────────────┤
│                                                      │
│  💭 You: What are the revenue trends?              │
│                                                      │
│  🤖 AI: Revenue shows strong growth trend:          │
│      Q1: $2.1M  Q2: $2.3M  Q3: $2.5M               │
│      Growth rate: +9.5% quarterly average           │
│      📄 [Source: Page 3]                            │
│                                                      │
│  [Suggested questions:]                              │
│  • What drove the Q3 growth?                        │
│  • Compare to last year                             │
│  • Show expense trends                              │
│                                                      │
├─────────────────────────────────────────────────────┤
│ [Type message...] 📎 🎤 ✨        [Send ↵]         │
└─────────────────────────────────────────────────────┘

Context Panel:

Shows documents AI is using
Add/remove documents from context
View token usage
Adjust AI model settings

Message Types:

User Messages (💭): Your questions
AI Responses (🤖): Generated answers
System Messages (ℹ️): Status updates
Suggestions: Recommended follow-ups

Advanced Features

Rich Responses:

Formatted text (bold, italic, lists)
Code blocks with syntax highlighting
Tables and data visualizations
Embedded images and charts
Clickable citations

Interactive Elements:

Click citations to view source
Copy code snippets
Download generated content
Create tasks from responses
Share specific messages

Voice Input:

Click microphone icon
Speak your question
Automatic transcription
Multiple language support

Context Management

Understanding Context

What is Context? Context includes:

Documents you've selected
Previous conversation messages
User preferences and history
Workspace information
Related content

Context Window:

Limited by AI model (up to 128K tokens)
~100,000 words of text
Automatically managed
Prioritizes recent and relevant content

Adding Documents to Context

Method 1: Manual Selection

Click "Add Context" or 📎 button
Search for documents
Select documents to include
Click "Add" (up to 10 documents)

Method 2: Chat from Document

Open document
Click "Chat with this document"
Automatically added to context

Method 3: Drag and Drop

Drag document from library
Drop onto chat window
Confirms addition to context

Context Limits:

Documents: 10 maximum
Total size: ~75MB of text
Automatic chunking for large docs
Priority: Most recently added

Managing Context

View Active Context:

Click "Context" button (top right)
See all documents in scope
Token usage per document
Remove documents not needed

Context Strategies:

Focused Context (Recommended)

Add only relevant documents
Better, more accurate answers
Faster response times
Lower token usage

Broad Context

Multiple documents for comparison
Cross-referencing information
Finding patterns across content
Useful for research

Example Context Setup:

Research Question: "How has our product strategy evolved?"

Relevant Documents:
✅ Product Roadmap 2024
✅ Q1 Strategy Meeting Notes
✅ Q2 Strategy Meeting Notes
✅ Q3 Product Review
✅ Customer Feedback Summary

Remove:
❌ Financial Reports (not relevant)
❌ HR Policies (not relevant)
❌ Marketing Campaigns (tangential)

Effective Prompting

Question Types

1. Factual Questions

What is the total revenue in Q3?
Who attended the product meeting?
When is the project deadline?

2. Analytical Questions

What are the main trends in customer feedback?
How do our costs compare to last quarter?
What factors contributed to the delay?

3. Comparative Questions

How does Q3 compare to Q2?
What's different in Version 2.0 vs 1.0?
Compare our approach to competitor X

4. Synthesis Questions

Summarize the key decisions from all strategy meetings
What are the common themes in customer complaints?
Identify patterns in our quarterly results

Prompting Best Practices

Be Specific:

❌ Tell me about the report
✅ What are the top 3 revenue drivers mentioned in the Q3 financial report?

❌ What did they say about customers?
✅ What specific customer pain points were mentioned in the product feedback document?

Provide Context:

❌ What should we do?
✅ Based on the risk assessment document, what are the top 3 recommended mitigation strategies?

❌ Compare them
✅ Compare the marketing ROI between Q2 and Q3 based on the campaign performance reports

Set Output Format:

Create a bullet list of action items from the meeting
Generate a table comparing product features
Write a 3-paragraph executive summary
List all mentioned deadlines in chronological order

Define Constraints:

In 3 sentences or less, explain...
Using only information from the Q3 report...
Without technical jargon, describe...
Focus only on financial aspects...

Advanced Prompting Techniques

Chain of Thought:

First, identify all budget items in the Q4 report.
Then, compare each to Q3 spending.
Finally, highlight any items that increased by more than 20%.

Role-Based:

As a financial analyst, review this report and identify risks
From a product manager's perspective, evaluate these features
Acting as a legal advisor, assess this contract for risks

Few-Shot Examples:

Extract action items in this format:
- Task: [description]
  Owner: [name]
  Due: [date]
  Priority: [high/med/low]

Example from previous meeting:
- Task: Update pricing page
  Owner: Sarah
  Due: Nov 15
  Priority: High

Now extract from this meeting...

Iterative Refinement:

1. You: Summarize this document
   AI: [provides summary]
   
2. You: Focus more on financial aspects
   AI: [revised summary]
   
3. You: Add specific numbers and metrics
   AI: [final version with data]

Multi-Document Analysis

Cross-Document Queries

Pattern Recognition:

What common themes appear across all quarterly reports from 2024?

AI analyzes:

Q1 2024 Report
Q2 2024 Report
Q3 2024 Report
Q4 2024 Report

Result: Identifies recurring topics, trends, issues

Trend Analysis:

How has customer satisfaction changed from Q1 to Q4 based on all survey documents?

Returns:

Quantitative changes
Qualitative shifts
Notable patterns
Anomalies or outliers

Contradiction Detection:

Are there any conflicting recommendations between the strategy document and implementation plan?

Finds:

Misalignments
Inconsistencies
Conflicts of approach
Discrepancies in data

Gap Analysis:

What information is covered in Document A but missing from Document B?

Comparative Tables:

Create a table comparing product features across all spec documents

Result:
| Feature        | Product A | Product B | Product C |
|---------------|-----------|-----------|-----------|
| Price         | $99       | $149      | $199      |
| Storage       | 100GB     | 500GB     | 1TB       |
| Users         | 5         | 25        | Unlimited |

Citations and Sources

How Citations Work

Automatic Citation: Every AI response includes sources:

AI: Revenue increased 15% in Q3, driven primarily 
by enterprise sales growth. 📄 [Q3 Report, p.3]

Customer acquisition costs decreased by 8% compared 
to Q2. 📄 [Marketing Analysis, p.12]

Citation Components:

Document name
Page or section
Confidence level (high/medium/low)
Direct link to source

Verifying Information

Click Citations:

Click [Source] link
Opens document at exact location
Highlighted relevant passage
Compare AI response to original

Confidence Indicators:

✅ High Confidence (95%+): Strong evidence
⚠️ Medium Confidence (70-95%): Probable
❓ Low Confidence (<70%): Uncertain

When to Verify:

Critical business decisions
Legal or compliance matters
Financial data and numbers
Dates and deadlines
Quotes and attributions

Manual Verification:

You: What page discusses pricing strategy?

AI: Pricing strategy is discussed on pages 15-18 
of the Product Strategy document. 📄 [Source]

[Click source to verify]

Conversation Management

Organizing Conversations

Naming Conversations:

Auto-generated from first question
Edit to descriptive title
Use consistent naming: "[Project] - [Topic]"
Examples:
- "Q3 Analysis - Revenue Trends"
- "Product Launch - Marketing Strategy"
- "Legal Review - Contract Terms"

Starring Important Conversations:

Click star icon
Quick access from sidebar
Highlights critical insights
Easy retrieval later

Folder Organization:

Create conversation folders
Group by project, topic, or date
Drag conversations to folders
Nested folder support

Tagging:

Add tags to conversations
Multiple tags per conversation
Filter by tag
Shared team tags

Conversation History

Viewing History:

All conversations sidebar
Search by content or title
Filter by date range
Filter by documents used

Exporting Conversations:

Open conversation
Click "..." menu → "Export"
Choose format:
- Markdown (.md)
- PDF
- Text (.txt)
- JSON (with metadata)
Download file

Sharing Conversations:

Click "Share" button
Options:
- Copy link (view-only)
- Invite team members (collaboration)
- Export and share file
Set permissions:
- View only
- Can comment
- Can continue conversation

Archiving and Deletion

Archiving:

Hide from active list
Still searchable
Easy to restore
Preserves all content

How to Archive:

Open conversation
Click "..." menu → "Archive"
Or: Swipe left (mobile)

Restoring:

Go to "Archived" folder
Find conversation
Click "Restore"

Deletion:

Permanent removal
30-day recovery period
Warning before deletion
Cannot recover after 30 days

Special Features

Response Regeneration

When to Regenerate:

Response not quite right
Want different angle
Need more detail
Prefer different format

How to Regenerate:

Hover over AI response
Click "Regenerate" button
Optionally adjust prompt first
Get new response

Variations:

Different phrasing
Alternative perspectives
More/less detail
Different examples

Creating Tasks from Chat

Convert Responses to Tasks:

Find action item in AI response
Click "Create Task" button
Auto-fills task details:
- Title from action item
- Description from context
- Source links included
Assign to person
Set due date
Create in connected tool (Jira/Asana)

Example:

AI: Based on the meeting, here are the action items:
1. Update pricing page by Nov 15 [Create Task →]
2. Schedule customer interviews [Create Task →]
3. Review contract terms [Create Task →]

Advanced Settings

AI Model Selection

Available Models:

GPT-4 (Default)

Most capable
Best for complex analysis
Slower response time
Higher token cost

GPT-4o-mini

Faster responses
Good for simple queries
Lower cost
Suitable for drafts

Claude 3 Opus

Excellent for long documents
Strong analytical capabilities
Good with technical content
Alternative perspective

Changing Models:

Click ⚙️ (settings) in chat
Select "AI Model"
Choose from dropdown
Applies to new messages

Temperature and Creativity

Temperature Control:

Low (0.3): Factual, consistent, focused
Medium (0.7): Balanced (default)
High (1.0): Creative, varied, exploratory

When to Adjust:

Low: Financial analysis, legal review
Medium: General questions, summaries
High: Brainstorming, creative writing

Token Management

What are Tokens?

Units of text (≈4 characters)
Both input and output counted
Affects cost and limits

Token Usage:

Context: 10,000 tokens
Question: 50 tokens
Response: 500 tokens
Total: 10,550 tokens

Optimizing Token Use:

Remove unnecessary documents from context
Ask focused questions
Use summaries for large documents
Clear context between unrelated topics

Use Cases

Research & Analysis

Literature Review:

Summarize key findings from all research papers, 
organized by methodology

Result: Grouped summaries by:
- Quantitative studies
- Qualitative studies
- Mixed methods
- Meta-analyses

Competitive Analysis:

Compare our product features to competitors mentioned 
across all market research documents

Result: Feature comparison matrix with:
- Our strengths
- Competitor advantages
- Market gaps
- Opportunities

Business Operations

Meeting Follow-ups:

From today's meeting transcript, extract:
1. Decisions made
2. Action items with owners
3. Open questions
4. Next steps

Result: Structured summary ready to share

Report Generation:

Create an executive summary of the Q3 results 
including key metrics, challenges, and recommendations

Result: 2-page summary suitable for leadership

Customer Insights

Feedback Analysis:

What are the top 5 customer pain points mentioned 
across all support tickets this month?

Result: Ranked list with:
- Frequency of mention
- Impact assessment
- Example quotes
- Suggested solutions

Sentiment Tracking:

How has customer sentiment changed over the past 
3 quarters based on survey data?

Result: Trend analysis with:
- Sentiment scores
- Key drivers
- Notable changes
- Recommendations

Troubleshooting

AI Seems Confused:

✅ Start new conversation (clear context)
✅ Be more specific in questions
✅ Verify documents are processed
✅ Check document quality
✅ Rephrase question

Irrelevant Responses:

✅ Remove unrelated documents from context
✅ Add specific document names to question
✅ Use explicit constraints
✅ Verify question clarity

Missing Citations:

✅ Ask explicitly: "Please include sources"
✅ Request quotes: "Quote specific passages"
✅ Verify information in original document
✅ Report if consistently missing

Slow Responses:

Reduce context (fewer/smaller documents)
Use faster AI model (GPT-4o-mini)
Split complex questions into parts
Check internet connection

Best Practices

Do:

✅ Start with simple questions, then refine
✅ Provide clear context
✅ Verify important information
✅ Save useful conversations
✅ Use specific document names

Don't:

❌ Ask multiple unrelated questions at once
❌ Include too many documents
❌ Accept AI responses blindly
❌ Use for sensitive data without verification
❌ Forget to check sources

Next Steps

Chat with PDF - Analyze documents in depth
Meeting Transcriber - Question meeting content
Advanced Prompting Guide - Master AI interactions
AI Agents - Automate with AI

Need help? Check Troubleshooting or contact support.

Last Updated: October 11, 2025

Chat with AI ​

Overview ​

Getting Started ​

Starting a Conversation ​

Your First Conversation ​

Chat Interface ​

Layout Components ​

Advanced Features ​

Context Management ​

Understanding Context ​

Adding Documents to Context ​

Managing Context ​

Effective Prompting ​

Question Types ​

Prompting Best Practices ​

Advanced Prompting Techniques ​

Multi-Document Analysis ​

Cross-Document Queries ​

Citations and Sources ​

How Citations Work ​

Verifying Information ​

Conversation Management ​

Organizing Conversations ​

Conversation History ​

Archiving and Deletion ​

Special Features ​

Suggested Questions ​

Response Regeneration ​

Creating Tasks from Chat ​

Advanced Settings ​

AI Model Selection ​

Temperature and Creativity ​

Token Management ​

Use Cases ​

Research & Analysis ​

Business Operations ​

Customer Insights ​

Troubleshooting ​

Best Practices ​

Next Steps ​