How does Intelligent Research Work?

Till now, you understand what the research agent is. But the important part is the accuracy of analysis and the process.

For example, in equity research, we use 5 agents to analyze each stock. But how does this process work?

Multi-Agent Analysis Structure

Financial analyst, qualitative analyst, and technical analyst work separately on each stock to gather data and perform analysis.

The sources of data include:

• Web data
• Financial statements
• News
• Charts

However, these agents do not rely on limited knowledge. They are trained using:

• Instruction tuning
• Few-shot learning
• Prompt engineering

But even that is not enough.

GPT Research Integration

We use a specific model called GPT Research for this process.

Each question asked by an agent is treated as a task for GPT Research. This means:

• Each agent collaborates with GPT Research
• Tasks are executed deeply and comprehensively
• Outputs are more reliable and structured

Role of the Head Analyst

After this process, the separated data is sent to the Head Analyst.

This agent is responsible for:

• Combining results
• Validating analysis
• Producing the final output

The final outputs include:

• Buy
• Strong Buy
• Hold
• Sell
• Strong Sell

Context-Aware Decision Making

With successful training, this system can generate outputs based on the nature of the asset.

For example:

A megatrend company without revenue may not be attractive based on financial data alone. However, it may have strong growth potential when considering:

• Product strength
• Market trends
• Technology

In such cases, the agent prioritizes qualitative data over purely financial metrics.

Valuation Layer

After the main analysis, the Head Analyst passes the result to the Valuation Agent.

This agent is responsible for evaluating:

• Company value
• Stock price positioning

For example:

NVIDIA (NVDA) is a great company with significant growth potential, but it may be overvalued, meaning it is not always the right time to buy.

Final Output for Portfolio Use

At this stage, a complete and detailed decision is available.

This analysis is then used by:

• Investment managers
• Portfolio managers

to make informed portfolio decisions.

How many research agents does Intelligent have?

We have developed multiple research agents across different markets:

• Equity Research → more than 5 agents
• Commodity Research → more than 3 agents
• Forex Research → more than 3 agents
• Bond Research → 3 agents
• Economic Research → 3 agents
• Indices Research → 3 agents
• Crypto Research → 4 agents

We are continuously improving:

• Efficiency
• Accuracy
• Coverage

and adding new agents for:

• Industry analysis
• Foresight
• Scenario building

Where are the researches used?

These research outputs are used in three main parts of the Intelligent investment platform:

1. Robo Advisor

To design hyper-personalized portfolios, the AI Portfolio Manager leverages this extensive analysis.

2. Asset Analysis

Within the platform:

• Pro users can access full analysis on each asset page
• Basic users can access partial insights

3. Human Analyst Layer

This analysis is also used by human experts to:

• Design unique portfolios for wealth management clients
• Provide deeper market insights

Conclusion

In the new era of investment, AI significantly reduces the time required for analysis.

Even without deep financial knowledge, users can access top-tier insights about:

• Assets
• Markets
• Future trends

That’s what we do at Intelligent—empowering wealth with AI.