Features
Deep Research
Run long-form research workflows for complex questions and decision support.
Deep Research
Deep Research is built for complex questions where quick answers are not enough.
It spends more time gathering, comparing, and structuring information.
Screenshot
[IMAGE PLACEHOLDER: Deep Research mode enabled in chat with long-form structured output]
When to Use Deep Research
| Use Deep Research | Use normal chat instead |
|---|---|
| Vendor evaluations | Simple factual lookup |
| Market opportunity analysis | Quick daily summaries |
| Policy/trend tracking | Single-step tasks |
| Strategic planning inputs | Fast drafting requests |
How Deep Research Differs from Normal Chat
| Area | Normal chat | Deep research |
|---|---|---|
| Runtime | Short | Longer |
| Exploration depth | Moderate | High |
| Output shape | Short answer | Structured report |
| Best use | Everyday tasks | High-value decisions |
Strong Prompt Framework
Research [topic] for [audience/use case],
cover [scope],
for [time range],
and return [output format].Example prompt
Research customer support AI tools for B2B SaaS companies (100-1000 employees),
covering pricing, strengths, risks, and implementation complexity,
for the last 12 months,
and return a decision table plus recommendation.Output Formats You Can Request
- Executive summary
- Comparison table
- Decision matrix
- Risk register
- 30/60/90-day rollout plan
Example:
Return:
1) one-paragraph executive summary,
2) a comparison table,
3) final recommendation with 3 reasons.Quality Checklist
| Check | Why it matters |
|---|---|
| Scope is explicit | Reduces irrelevant research |
| Time range is explicit | Improves freshness and comparability |
| Output format is explicit | Makes output easier to use |
| Audience is explicit | Improves tone and decision quality |
Cost Guidance
Deep research is higher-cost than normal chat and should be used for important decisions.
Use Deep Research for high-impact questions, not routine daily checks.