Compiled AI: Deterministic Code Generation for LLM-Based Workflow Automation

Geert Trooskens; Aaron Karlsberg; Anmol Sharma; Lamara De Brouwer; Max Van Puyvelde; Matthew Young; John Thickstun; Gil Alterovitz

cs.SEApr 6, 2026

Compiled AI: Deterministic Code Generation for LLM-Based Workflow Automation

Authors: Geert Trooskens, Aaron Karlsberg, Anmol Sharma, Lamara De Brouwer, Max Van Puyvelde, Matthew Young, John Thickstun, Gil Alterovitz,

Organizations: XY.AI Labs, Palo Alto, CA · Stanford University School of Medicine, Stanford, CA · Cornell University, Ithaca, NY · Cornell University, Department of Computer Science, Ithaca, NY ·

Abstract

We study compiled AI, a paradigm in which large language models generate executable code artifacts during a compilation phase, after which workflows execute deterministically without further model invocation. This paradigm has antecedents in prior work on declarative pipeline optimization (DSPy) and hybrid neural-symbolic planning (LLM+P); our contribution is a systems-oriented study of its application to high-stakes enterprise workflows, with particular emphasis on healthcare settings where reliability and auditability are critical. By constraining generation to narrow business-logic functions embedded in validated templates, compiled AI trades runtime flexibility for predictability, auditability, cost efficiency, and reduced security exposure. We introduce (i) a system architecture for constrained LLM-based code generation, (ii) a four-stage generation-and-validation pipeline that converts probabilistic model output into production-ready code artifacts, and (iii) an evaluation framework measuring operational metrics including token amortization, determinism, reliability, security, and cost. We evaluate on two task types: function-calling (BFCL, n=400) and document intelligence (DocILE, n=5,680 invoices). On function-calling, compiled AI achieves 96% task completion with zero execution tokens, breaking even with runtime inference at approximately 17 transactions and reducing token consumption by 57x at 1,000 transactions. On document intelligence, our Code Factory variant matches Direct LLM on key field extraction (KILE: 80.0%) while achieving the highest line item recognition accuracy (LIR: 80.4%). Security evaluation across 135 test cases demonstrates 96.7% accuracy on prompt injection detection and 87.5% on static code safety analysis with zero false positives.

Figures & tables

Panel of Figure 1 — Figure 1: Token consumption comparison across baselines (BFCL, n=400). Compiled AI incurs a one-time 9,600-token generation cost, then executes at zero marginal cost. Break-even vs. Direct LLM at $n^{*}\approx 17$ ; at 1,000 transactions, compiled AI reduces consumption by 57 $\times$ vs. Direct LLM and 84 $\times$ vs. AutoGen.

Panel of Figure 2 — Figure 2: The code foundry architecture. Business intent (YAML) enters; validated Temporal activities emerge. Dashed red arrows show the regeneration loop on validation failure. The LLM runs once at generation time. Generated artifacts may include bounded LLM tool calls for steps requiring semantic understanding at runtime.

Method	GenTokens	Runtime/tx	Total@1K	BreakEven	CompRatio
Direct LLM	0	552	552,000	n/a	n/a
LangChain	0	740	740,000	n/a	0.75 $\times$
AutoGen	0	805	805,000	n/a	0.69 $\times$
Compiled AI	9,600	0	9,600	$\sim$ 17 tx	$\sim$ 57 $\times$

Table 1: Token consumption comparison (BFCL, n=400). Runtime tokens reported for the function-calling task; the Code Factory variant for document intelligence is reported in Table 3 .

Method	Inference	Infra	Total TCO	CostRatio
Direct LLM	$21,500	$500	$22,000	40 $\times$
LangChain	$28,900	$500	$29,400	53 $\times$
AutoGen	$31,400	$500	$31,900	57 $\times$
Compiled AI	$55	$500	$555	1 $\times$

Table 2: Total cost of ownership at 1M transactions/month (Claude Opus 4.5:

15/1M input,

75/1M output tokens).

† Median latency; mean 10,263 ms due to 10% retry outliers (max 183s). First compile: 57s.
Paradigm	Approach	KILE	LIR	Latency	LLM Calls
Compiled AI	Deterministic (Regex)	20.3%	59.7%	0.6 ms	None
Compiled AI	Code Factory	80.0%	80.4%	2,695 ms †	Compiled
Runtime AI	Direct LLM	80.0%	74.5%	6,339 ms	Per-request
	LangChain	80.0%	75.6%	6,207 ms	Per-request
	AutoGen	77.8%	78.9%	13,742 ms	Per-request

Table 3: DocILE results. KILE: Key Information Line-level Extraction (exact-match on key fields, e.g. invoice number, date, total). LIR: Line Item Recognition (accuracy on structured line items). Code Factory makes compiled LLM calls; Deterministic uses pure regex.

Code Generation

Code generation studies methods for turning natural-language, visual, or formal specifications into executable programs and code edits.

See all 313 papers

Auto: The AGI Compiler
Jul 5, 2026Jaber Jaber, Osama JaberContinual Learning for LLM Agents LLM-Based Program Synthesis
RightNow AI
Evaluating LLM Trade-offs for Enterprise Automation: Lessons from Workflow Generation in a Production Enterprise Platform
Aug 4, 2026Xavier Wrenn, Radoslav Raykov, Aleksandar Angelov +3LLM Evaluation Workflow Automation
IBM
Broken by Default: A Formal Verification Study of Security Vulnerabilities in AI-Generated Code
Apr 7, 2026Dominik Blain, Maxime NoiseuxLLM Security LLM Code Generation
Cobalt AI Gatineau, QC, Canada

Jul 5, 2026J/K move · Enter open · S save

Auto: The AGI Compiler

Jaber Jaber, Osama Jaber

RightNow AI

Continual Learning for LLM Agents LLM-Based Program Synthesis LLM Agent Self-Improvement LLM Agent Verification Runtime Enforcement for AI Agents LLM Agent Efficiency

Open paper AskPDF arXiv

Every LLM agent run re-derives its behavior token by token on a frontier model: brilliant, expensive, slow, and unbounded. We present Auto, a compiler that records live agent behavior, measures which parts are secretly deterministic, extracts them into verified programs or distilled specialists, and emits cognition binaries: WebAssembly artifacts whose manifests carry measured guarantees and whose declared capabilities are physically enforced by the sandbox. A tiered runtime executes compiled behavior behind conformally calibrated guards; guard trips deopt to the reference agent, and the captured trace recompiles back down, so nothing is figured out twice. We use "AGI compiler" in one narrow, testable sense: a system that autonomously converts novel experience into permanent, verified, near-free skill while measuring what it does not know. On AUTO-BENCH, a benchmark we introduce and pre-register, 87.1% of 560 recorded frontier-agent spans are witnessed-deterministic (three of the four censused task families measure 100.0%). On a 300-item stream with three scheduled distribution shifts, the closed loop compiles three artifact generations and drives marginal cost from 59 to 2 micro-dollars per item (6.4x end-to-end) at 96.9% parity on witnessed inputs with zero errors. The same stream also quantifies the failure modes: a loose guard silently mislabels 48.9% of compiled answers, and an unfaithful deopt reference causes the verification gate to refuse recompilation. Calibration and reference fidelity, not model capability, decide whether cheap stays correct. Code: https://github.com/RightNow-AI/auto

Abstract

Code Generation

Explore similar work

Auto: The AGI Compiler

Evaluating LLM Trade-offs for Enterprise Automation: Lessons from Workflow Generation in a Production Enterprise Platform

Broken by Default: A Formal Verification Study of Security Vulnerabilities in AI-Generated Code