Better Call Saul - Your AI Corporate Advisory Team

A full-stack application that coordinates three specialist AI agents (Tax CPA, Legal Attorney, Business Strategist) to provide unified business formation advice. Built with React, Python, Google ADK.

Comment

Inspiration

Starting a business is overwhelming. Entrepreneurs face a maze of decisions: LLC or C-Corp? Delaware or home state? How to handle taxes? What about future funding? The problem is that tax CPAs, corporate attorneys, and business strategists traditionally work in silos—each expert understands their domain but lacks insight into how their advice impacts the others. A tax strategy that saves money today might kill your Series A funding tomorrow. A legal structure that protects assets might create a tax nightmare.

We built Better Call Saul to solve this coordination problem—stitching together three traditionally incompatible professional domains (tax accounting, corporate law, and business strategy) into one unified AI advisory platform that provides holistic, conflict-free guidance.

What it does

Better Call Saul is a multi-agent AI advisory platform that guides entrepreneurs through business formation and corporate strategy. The system consists of four intelligent agents:

  • Tax CPA Agent - Analyzes LLC/S-Corp/C-Corp taxation, pass-through vs double taxation, QBI deductions, state tax implications, and payroll strategies
  • Corporate Attorney Agent - Evaluates liability protection, ownership structures, operating agreements, compliance requirements, state registration, and funding implications
  • Business Strategist Agent - Advises on growth trajectory (bootstrap vs VC), industry regulations, operational complexity, state selection, employee equity/ESOPs, and exit strategies
  • Coordinator Agent - Synthesizes insights from all three specialists, identifies conflicts between recommendations, and presents a unified strategic plan with clear trade-offs and next steps

The key innovation: Unlike real-world consulting where these professionals rarely collaborate, our agents work together on every query. The Tax CPA's recommendations are cross-checked against the Attorney's legal concerns and the Strategist's growth plans, ensuring entrepreneurs get integrated, conflict-free advice that considers all angles.

Users simply describe their business situation through a chat interface, and the agent squad collaborates to deliver comprehensive formation strategies covering entity selection, compliance roadmaps, tax optimization, funding preparation, and growth planning.

Why Frankenstein Category?

Better Call Saul exemplifies the Frankenstein category by stitching together a chimera of seemingly incompatible professional domains into one powerful application:

  1. Tax accounting systems (numerical calculations, IRS compliance, state-specific regulations)
  2. Legal frameworks (corporate law, liability protection, contract structures)
  3. Business strategy methodologies (growth models, funding dynamics, market analysis)
  4. Multi-agent AI coordination (Google ADK, Gemini 2.5 Flash, agent orchestration)

These four domains rarely interact in traditional consulting—they use different languages, frameworks, and decision criteria. Better Call Saul uses AI to bridge these incompatible worlds, creating something unexpectedly powerful: a unified advisory system that ensures recommendations from one domain don't contradict or undermine advice from another.

How We Built It With Kiro

This entire project was developed using Kiro's AI-powered development features, transforming what would have been weeks of manual coding into a streamlined build process.

Kiro Spec-Driven Development

We used Kiro's Spec Mode to architect the entire application before writing a single line of code:

Requirements Phase: We described Better Call Saul's vision to Kiro, which generated a comprehensive requirements.md document with 15+ user stories covering agent coordination, API design, frontend interactions, and error handling.

Design Phase: Kiro created a detailed design.md that specified our tech stack (React + Vite frontend, Flask backend, Google ADK agents), API endpoints, data flow between agents, and deployment architecture for Cloud Run.

Task Breakdown: Kiro generated a tasks.md file with 30+ actionable implementation tasks, organizing the build into logical phases: backend setup, agent implementation, frontend development, integration, and deployment.

This spec-driven approach prevented costly rework—by catching architectural issues during planning rather than mid-implementation, we saved an estimated 40% development time.

Most Impressive Kiro Contributions

Multi-Agent Orchestration: The most complex part was implementing the agent coordination pattern where the Coordinator delegates to three specialists and synthesizes their responses. We described the flow conceptually to Kiro, and it generated the Google ADK sub-agent pattern with proper error handling and response aggregation—code that would have taken hours to research and implement manually.

Cloud Run Deployment Configuration: Kiro generated Dockerfiles, nginx configs, and gcloud deployment commands that worked on the first try, saving extensive debugging time.

Session Management: Implementing stateless conversation history in a serverless environment was tricky. Kiro designed a session-based approach that maintains context across requests while staying within Cloud Run's free tier.

Backend Development with Kiro

  • Built multi-agent system using Google Agent Development Kit (ADK) with Python—Kiro generated the agent class structures and coordination logic
  • Deployed agents as a RESTful API service on Google Cloud Run with auto-scaling—Kiro created deployment configs and health check endpoints
  • Powered by Gemini 2.5 Flash model for fast, cost-efficient agent reasoning—Kiro configured API integration and token optimization
  • Implemented proper agent coordination with sub-agents pattern—Kiro architected the delegation and synthesis flow

Frontend Development with Kiro

  • Created responsive chat UI with React and Vite—Kiro scaffolded components and state management
  • Integrated react-markdown for formatted responses—Kiro added the library and styling
  • Implemented real-time loading states and conversation history—Kiro built the UX flow

What We Learned

  • Spec mode is essential for complex projects: Trying to build a multi-agent system with just vibe coding would have resulted in architectural inconsistencies. Specs provided a clear blueprint that all subsequent code generation followed.

  • Kiro excels at integration tasks: Connecting Google ADK, Flask, React, and Cloud Run involved complex compatibility requirements. Kiro handled these integrations seamlessly, referencing up-to-date documentation we would have spent hours searching for.

  • The real pain point of conflicting advice from independent consultants

  • How tax, legal, and business strategy decisions impact each other

  • The importance of integrated thinking in corporate formation

What's Next for Better Call Saul

  • Document generation - Create operating agreements, shareholder agreements, and IRS forms directly
  • Cost calculator tool for filing fees, compliance costs, and tax projections
  • Comparison mode - Evaluate 2-3 entity structures side-by-side
  • Fine-tune agents with real legal/tax database tools for more accurate citations
  • Connect to real professionals - Bridge users to CPAs/attorneys for implementation
  • Collaborative mode for co-founders to discuss recommendations together
  • Industry-specific templates for SaaS, consulting, e-commerce, etc.

Better Call Saul proves that Kiro-powered development can tackle complex architectural challenges, enabling solo developers to build sophisticated multi-agent systems that would traditionally require a full engineering team. By stitching together incompatible professional domains into one unified platform, we're making expert business advice accessible, integrated, and actionable for entrepreneurs worldwide.

Built With

Share this project:

Updates