Prior art
ABSORB is not the first readiness framework for AI in software delivery, and its central observation is not new. Each entry below says how it was read. Last checked 2 November 2026.
DORA, 2025
The State of AI-assisted Software Development report already makes ABSORB's core argument: "AI's primary role is as an amplifier, magnifying an organization's existing strengths and weaknesses." DORA's capability guides also cover visibility of work in the value stream and streamlining change approval.
DORA is a far larger research program with survey data that ABSORB does not have. ABSORB cites it as the source of the amplifier finding and claims only what it adds:
- A whole-path speedup check from stage timing, splitting hands-on time from waiting. DORA's value-stream guide records lead and process time per block but gives no speedup model.
- Autonomy tiers for agent-authored change, tied to an evidence floor. DORA's change-approval guide implies risk-based scrutiny but defines no tiers.
- A small, falsifiable instrument: 24 statements, a written scoring rule, an offline assessment and pre-registered hypotheses.
Read: the report's landing page, the AI Capabilities Model page and the two capability guides. The full report PDF was not read.
Other readiness frameworks
| Work | What it is | How it differs |
|---|---|---|
| Codurance, Data and AI Readiness Assessment | A paid 3 to 4 week consultancy engagement on readiness to deploy AI agents into business processes | About deploying agents, not the software delivery path; no review queues or delivery stages |
| Augment Code, AI readiness assessment framework | A vendor guide compiling several existing models, scored 1 to 5 | No stage timing, no autonomy levels tied to evidence, no tool. Its "any dimension below 3.0 blocks" rule plays the same role as ABSORB's weakest-dimension cap |
Not yet read: an academic paper, Agentic AI Readiness: A Process-Oriented Assessment Framework (University of Hawaiʻi ScholarSpace), and Codurance's separate AI Engineering Assessment.
Foundations
Amdahl's law, the theory of constraints, Little's law, queueing theory, Brooks's law, Goodhart's law and Conway's law are on the concept page. The name comes from absorptive capacity (Cohen and Levinthal, 1990). ABSORB does not claim to measure absorptive capacity in the original sense.
Related work by the same author
ABSORB stands on its own and draws its theory from outside sources. Two earlier records address neighboring questions:
- Governance-Driven Design (doi:10.5281/zenodo.20938778) describes formalizing constraints before agents run. ABSORB's D3 and D4 ask whether an organization has done that kind of work.
- Capability-as-Infrastructure (doi:10.5281/zenodo.23438908) states hypotheses with a validation protocol, the pattern ABSORB's validation plan follows.
The full comparison is in PRIOR_ART.md.