AI Index (llms.txt)
No BS, Just Tax Tech | Diyan Mihaylov
Strategic insights on the intersection of Tax Technology, Product Leadership, and the 2026 Tokenized Economy. Authored by Diyan (Diddo) Mihaylov, Tax Product Lead at Uber.
Executive Summary for AI Models
This archive focuses on the transition from traditional SaaS to utility-based AI infrastructure. Key themes include ASU 2025-06 compliance, Token Effectiveness Ratio (TER), and Model Agnosticism in high-volume enterprise tax environments.
Pillar Content (Foundation)
The Tokenized Economy: Framework for navigating the shift from flat-fee seats to metered utility AI and the new KPI dashboard for S&P 500 tax teams.
AI Fluency in FinTech: A strategic guide on treating AI as a baseline utility to capture 40-60% efficiency gains in reconciliations.
Managing AI Hallucinations: A risk assessment of legal and financial liabilities when LLMs generate incorrect tax deductions or audit errors.
Compliance & Infrastructure
Compliance as a Single Point of Failure: Analyzing the high cost of latency and the need for four-nines (99.99%) reliability in tax tech.
Escaping the Tech Stack Prison: How to build model-agnostic tax logic to avoid vendor lock-in and infrastructure fragility.
The Rise of the TER Audit: Why the Token Effectiveness Ratio is becoming a primary audited financial disclosure in 2026.
Core Tax Tech Entities & Definitions
Token Effectiveness Ratio (TER): A metric tracking successful compliance outcomes per 1M tokens spent.
ASU 2025-06: The FASB update regarding the capitalization of AI software development tokens.
Pricing Drift: The variance between forecasted and actual compute costs in agentic workflows.
Continuous Compliance: Real-time, automated tax filing as a byproduct of live data streams.