Skip to content

Named enterprise rollouts

This page collects publicly-reported deployments of AI coding assistants at named companies. Inclusion criteria are deliberately strict: each entry has a verifiable URL pointing to a primary write-up by the company itself, a vendor case study (clearly labeled as such), or coverage by a major outlet.

Read these as marketing-adjacent

Almost all public reporting on enterprise AI coding rollouts comes from companies that consider the rollout a success and from vendors selling the tool. The "what didn't work" reporting in this space is sparse. Treat headline numbers from these write-ups as upper-bound, best-case claims, not generalizable expectations.

How to read each entry

For each rollout below, four facts are flagged:

  • Scale — engineers in scope, where reported.
  • Tool(s) — what is actually deployed.
  • Reported outcomes — what the source claims, with the source's own framing flagged.
  • Source type — [primary] (the company's own engineering write-up or first-party announcement), [vendor] (a vendor case study using a named customer), or [secondary] (press coverage of the rollout).

GitHub — internal Copilot use and external rollout reporting

Scale. GitHub itself is the originating customer of Copilot; its broader claims about productivity uplift across customers are reported on the GitHub Blog and in commissioned research1.

Tool. GitHub Copilot.

Reported outcomes. GitHub's own research blog reports developer- satisfaction and task-time benefits in line with Peng et al.'s 2023 RCT, and reports broader rollout metrics across customers. The methodology overlaps with the RCT [vendor].

Caveat. GitHub publishes both the tool and the research. Cite the underlying peer-reviewed Peng et al. paper (see productivity studies) where possible; the blog content is best read as marketing for the same finding.

Source. GitHub Research blog, "Research: quantifying GitHub Copilot's impact on developer productivity and happiness"1.

Goldman Sachs Engineering — public engineering writing on AI

Scale. Goldman Sachs Engineering publishes on its developer.gs.com site about engineering practice; the firm has publicly stated developer AI rollout intentions in CEO and CIO interviews covered in business press3.

Tool. Multiple, with internal platform integration. Specific vendor is not the focus of the public statements.

Reported outcomes. Public statements emphasize internal platform work and developer enablement, framed as productivity uplift; numeric claims should be checked against the underlying interview, not headlines.

Caveat. Bank engineering blogs are recruiting and brand surfaces. Treat as [primary] for what they did and [secondary] for how well it worked.

Source. Goldman Sachs Developer site2; press coverage of executive AI commentary3.

Stripe — engineering blog on AI-assisted internal work

Scale. Stripe Engineering publishes on AI-assisted internal tools and developer productivity work via its engineering blog and conference talks.

Tool. Internal tooling; commercial assistants used alongside.

Reported outcomes. Stripe's public material centers on internal LLM-powered developer tooling — code search, documentation generation, on-call assistance — rather than on raw "engineers shipping faster" claims [primary].

Caveat. Stripe is a sophisticated engineering shop; the patterns described generalize less easily than the framing suggests. Read for architecture ideas, not for expected ROI.

Source. Stripe engineering blog4.

Shopify — engineering blog on AI-assisted developer workflows

Scale. Shopify Engineering publishes on platform changes, including AI-assisted developer workflow integrations and internal tooling.

Tool. Multiple, with significant internal platform investment.

Reported outcomes. Public posts emphasize workflow integration and platform leverage (e.g., AI integrated into internal CLIs, code review, documentation). Numeric claims appear less often than qualitative outcomes [primary].

Caveat. As with Stripe, Shopify has unusually mature platform investment; the integration patterns will not transfer cheaply to smaller engineering organizations.

Source. Shopify engineering blog5.

What's missing from public enterprise rollouts

A short, honest critique of the public corpus:

  • Survivor bias. Companies that paused or rolled back rarely publish post-mortems. The public set is a non-random sample.
  • Vendor-aligned framing. Where vendors co-author or amplify customer stories, the language tends to favor adoption (engineers using the tool) over outcomes (defect rate, lead-time, change-failure rate).
  • Numeric claims travel without method. "Developer productivity is up X%" rarely comes with the underlying instrument (lines of code? pull requests? self-report?), making cross-company comparison meaningless.
  • Quality and security outcomes underreported. Public rollouts emphasize speed; the security incidents page exists because the same companies generally do not publish the corresponding incident data.
  • Maintenance horizon missing. None of these public reports cover the maintenance cost of code generated under early rollouts. The measurement window in the public corpus is months, not years.
  • What happened in regulated industries. Banks, insurers, and healthcare providers issue press releases; few publish the actual governance frameworks they adopted. The most-cited write-ups are from product-led tech firms, which is the opposite of where most governance pressure lives.

How to use this page in a leadership conversation

These rollouts are useful for vocabulary ("here is what a serious program publishes"), not for targets ("we should also see X% productivity gain"). When a vendor or analyst quotes a headline percentage from one of these companies, ask: what was measured, on what population, by whom, and for how long? Most of the time, the answer is in a sentence the headline did not include.

  1. GitHub, "Research: quantifying GitHub Copilot's impact on developer productivity and happiness", Sep 2022. Vendor-published. ↩↩

  2. Goldman Sachs, Goldman Sachs Developer, accessed 2026. ↩

  3. Reuters, "Goldman Sachs to roll out generative AI tool to bankers", Jun 2024. ↩↩

  4. Stripe, Engineering blog, accessed 2026. ↩

  5. Shopify, Engineering blog, accessed 2026. ↩