---
description: >-
  The research program behind Rosetta: field evidence first, theory second, falsifiable claims last — with provenance on every published number.
---

# Research & evidence

**Rosetta is not a product looking for a rationale. It is the
implementation of a public research program.** The ideas on this site —
[the delivery gap](../why/delivery-gap.md), [the digital
twin](../concepts/digital-twin.md), [the Spec
Manifest](../concepts/spec-manifest.md), [governed intent
compilation](../concepts/governed-intent-compilation.md) — were not
invented to decorate a platform. They were
developed in working papers, grounded in field studies of real
enterprise engagements, and stated in a form deliberately chosen so they
can be tested — and, where they are wrong, falsified. This section lays
out that evidence base so you can judge it for yourself.

## The lineage

The program comes out of Keenan Vision's research practice, whose
analysis of the Salesforce DevOps and agentic-delivery ecosystem has
been published for years at
[salesforcedevops.net](https://salesforcedevops.net). That reporting
practice grew into a working-paper series: papers on governed intent
compilation, the Spec Manifest, and delivery maturity in the
agentic enterprise, written with the collaborators each paper names on
its own cover — Ian Gotts and Mehmet Orun on the governed-intent-compilation
and Spec Manifest papers. Alongside the papers, the economics of virtual
employees is developed in a research partnership with the UC Berkeley
Haas School of Business.

## The method

The method runs in one direction: field evidence first, theory second,
falsifiable claims last. Structured engagement studies — coded finding
by coded finding — establish what is actually happening in enterprise
delivery; the working papers build theory on that base; and the theory
is then committed to predictions specific enough to fail. Behind every
published claim sits a curated, scope-gated evidence corpus, with each
claim carrying its provenance — which
source, of what kind, with what limits. That is the same provenance
standard [Rosetta itself enforces](../concepts/governance.md) on
delivery knowledge: the research program eats its own cooking. When a
number appears on these pages, it arrives with its instrument attached —
a bounded field sample is labeled as one, a survey association is never
dressed up as a causal result, and an analyst forecast is called a
forecast.

## What is in this section

- **[Field evidence](field-evidence.md)** — the engagement studies
  behind the numbers: the Data 360 Strategy Sprint, the Architecture as
  Strategy interviews, and the wider industry signal they sit inside.
- **[The working papers](papers.md)** — the three papers that carry the
  theory: what each argues, what each explicitly does not claim, and
  where each stands on the road to publication.
- **[Virtual Employee Economics](ve-economics.md)** — the Three Laws,
  the economics of an AI workforce, and why the Spec Manifest
  matters more as the human/machine labor boundary moves.

!!! note "About the papers"
    The working papers themselves are pre-publication drafts. These
    pages carry their ideas in essay form; the papers publish when they
    are submission-ready.

---

*Arrived here first? Start with [the overview](../index.md), then read
[the delivery gap](../why/delivery-gap.md) — the field finding the whole
program is built on.*