Welcome to Week 6 of our NegoAI series.

In this week's video, I demonstrate how to build custom AI agents that analyze LinkedIn profiles for personality assessment and negotiation strategy development - potentially replacing paid tools with tailored solutions.

Understanding AI Agents

AI agents are specialized systems designed to perform specific tasks autonomously within defined parameters. Unlike general LLMs, agents focus on particular functions and can work together in sequences to solve complex problems.

Research shows that effective negotiation preparation involves understanding counterparts' behavioral tendencies and communication preferences. Traditional approaches rely on subjective assessment, but AI agents can now systematically analyze digital footprints to generate evidence-based behavioral profiles.

Key components of functional AI agents include:

• Defined scope: Narrow focus on specific tasks

• Independent operation: Ability to process inputs without continuous guidance

• Structured outputs: Consistent, formatted information delivery

• Temperature control: Adjustable creativity vs. determinism balance

Building a Three-Agent System

In the video, I demonstrate a custom workflow built using Cassydi's agent framework that analyzes LinkedIn profiles to generate negotiation-relevant psychological profiles:

1. Data Preprocessing Agent (Temperature: 0.1)

   • Ingests LinkedIn profile data through a scraper

   • Cleans and structures information

   • Applies natural language processing techniques

   • Functions deterministically to ensure consistent data preparation

2. Psychometric Analysis Agent (Temperature: 0.15)

   • Processes prepared data from Agent 1

   • Applies statistical models to identify behavioral patterns

   • Generates multiple profile types (DiSC, OCEAN, Thomas-Kilmann)

   • Maintains strict determinism for consistent analysis

3. Strategy & Reporting Agent (Temperature: 0.25)

   • Transforms psychological profiles into actionable strategies

   • Generates communication recommendations

   • Structures information for practical application

   • Uses slightly higher temperature for nuanced strategic insights

Temperature in Cassidy AI can be set between '0 (deterministic) and 1 (highly creative)

Practical Applications and Results

When testing the system on my own profile, it identified me as primarily:

• DiSC Profile: Dominant and Conscientious

• Thomas-Kilmann Conflict Style: Collaborating with Competitive tendencies

The output included engagement strategies, communication preferences, and negotiation approaches tailored to this profile - comparable to insights from commercial tools like Humantic AI.

This type of preparation enables:

• Data-driven pre-negotiation intelligence

• Systematic adaptation of communication approaches

• Prediction of potential friction points

• Development of personalized influence strategies

Technical Implementation Notes

For those interested in building similar systems:

• Temperature settings: Lower temperatures (0-0.3) produce more consistent, deterministic outputs

• Agent sequencing: Define each agent's specific role and information flow between agents before implementation - carefully map which outputs from one agent become inputs for the next

• Instruction design: Structure system instructions for each agent using this framework:

  1. Role: Define the agent's specific function and expertise

  2. Context: Provide background information and purpose

  3. Instructions: Detail specific tasks and processes to follow

  4. Criteria: Establish standards for successful completion

  5. Examples: Include sample inputs and desired outputs

  6. Continuous Development: Guidelines for improvement

• Validation and evals: Test against known profiles to verify accuracy and compare results against human-created technical profiles - this provides benchmarks for system performance

Looking Ahead

While agent-based analysis provides valuable insights, remember that these are starting points rather than definitive assessments. Real-world interactions must take precedence over AI-generated profiles.

This implementation demonstrates how accessible agent technology has become. Tasks that previously required significant development resources can now be accomplished through thoughtful prompt engineering and workflow design.

This Week’s Exercise

Take 20 minutes to analyze a counterpart in your next negotiation:

1. Identify observable behavioral patterns in their communication

2. Note differences between their stated positions and possible underlying interests

3. Consider how their communication style might reflect conflict management preferences

4. Design an engagement approach that aligns with their behavioral tendencies

Preview of Next week

Next week, we'll explore how to design an AI agent that can systematically analyze negotiation scenarios from multiple perspectives, helping negotiators expand their thinking beyond conventional approaches.