Our verdict
8th Light is a good hire for an established company that wants one agentic AI workflow built properly on its existing stack, or a platform that has to hold up before agents run on it. Its senior-led, time-boxed model suits careful buyers in regulated industries more than teams that want a large build crew.
The right buyer
- Enterprise teams in regulated industries that want an AI agent running in production on their own stack within a few months.
- Engineering leaders who need platform modernization and AI work from the same senior team.
Reasons to look elsewhere
- Senior US and London consultants put it above nearshore rates.
- A team of about 140 cannot staff very large, multi-year programs.
8th Light, At A High Level
Paul Pagel and Micah Martin founded 8th Light in Chicago in 2006, and it built its name in the software craftsmanship movement, teaching client teams practices such as test-driven development and pair programming. Jenny Farver became CEO in 2022, when Pagel moved to president.
Its 140-plus team members work from Chicago, London, and satellite offices in Los Angeles, New York, and Austin. In June 2026 it launched an Agentic AI Studio that builds one production-ready agent workflow per engagement in eight to twelve weeks, running live on the client’s existing stack. Public client work includes DevSecOps for Pfizer, platform modernization for Cars.com, and a genomic resource for GeneDx.
Scores by criterion
| Criterion | Weight | Score | Basis |
|---|---|---|---|
| Production track record | 20% | 7.5 | Twenty years of custom software delivery, with public enterprise clients. |
| Technical depth | 15% | 7.0 | Platform engineering plus a dedicated agentic AI practice. |
| Industry depth | 10% | 6.5 | Regulated industries such as pharma, health, and genomics. |
| Delivery model and handoff | 15% | 7.5 | Senior leaders stay for the full engagement and teach client teams. |
| Commercial clarity | 10% | 7.0 | Studio engagements are scoped to one workflow in eight to twelve weeks. |
| Platform independence | 10% | 7.5 | Builds on the client’s existing stack with no single cloud tie. |
| Stability and scale | 10% | 7.0 | Founded 2006; about 150 employees on LinkedIn. |
| Independent evidence | 10% | 6.5 | Press coverage and client case studies; few public reviews. |
| Overall | 100% | 7.1 | Weighted average |
Where 8th Light fits
8th Light’s buyers are mid-size and large companies, often in pharma, health, and other regulated fields, that want custom software or an AI agent built to production standards inside their existing systems. It is best for a first serious agent workflow, platform rebuilds that agents will depend on, and teams that want their own developers to learn better practices along the way. On price it sits in the senior onshore tier, above nearshore firms and close to other US craft consultancies.
Its closest substitutes for the same work: Test Double for senior consultants who embed in client teams; Atomic Object for onshore custom software and AI development; Keyhole Software for senior US consultants who modernize core systems with AI; The Gnar Company for legacy modernization and Claude-based AI work.
What clients and leaders say
“The hard part of agentic AI isn’t the model. It’s everything else: security review, legacy integration, and the organizational gravity that quietly kills demos before they ship.”
“They worked inside our repo and our review process, and our own developers came out of it writing better tests.”
“The pilot was tightly scoped to one workflow. Anything beyond that became a second engagement and a second budget.”
Where customer reviews couldn’t be found, our analysts attempted to give a voice to buyer sentiment, drawing on real information from review sites, forums and product documentation. Based on: 8th Light About page, Agentic AI Studio launch release.
