← Back to research
·11 min read·company

Faros AI

Faros AI is an enterprise engineering intelligence platform from three ex-Salesforce Einstein founders — 100+ tool integrations, DORA metrics, and an AI-impact module tracking Copilot, Cursor, Claude, and Devin — with $36M disclosed funding and its own April 2026 report finding bugs per developer up 54% at high AI adoption.

https://www.faros.ai/ ↗
engineering-intelligenceenterprisegithub

Key takeaways

  • Founded in 2019 by three Salesforce Einstein alumni (Vitaly Gordon, Matthew Tovbin, Shubha Nabar) with $36M disclosed — $16M seed plus a $20M Series A led by Lobby Capital in June 2023; no Series B has been publicly announced
  • The AI impact module tracks GitHub Copilot, Cursor, Windsurf, Claude, Devin, and Amazon CodeWhisperer — adoption, acceptance rates, and percent AI-generated code by repo — and ties them to delivery and quality outcomes via cause-and-effect analysis across 100+ tool integrations
  • Its own AI Engineering Report 2026 (22,000 developers, 4,000+ teams) found bugs per developer up 54% and incidents-per-PR more than tripling at high AI adoption — and the once-promoted open-source Faros Community Edition has quietly disappeared from GitHub

FAQ

What is Faros AI?

Faros AI is an enterprise engineering intelligence platform that aggregates data from 100+ engineering tools — Git, Jira, CI/CD, incident management, and AI coding assistants — into unified dashboards covering DORA metrics, developer productivity, and AI coding-tool impact.

How much does Faros AI cost?

Pricing is not publicly listed. Faros sells three tiers (Professional, Enterprise, Ultimate) through demos and custom quotes; a competitor analysis claims typical contracts start around $150K/year.

How does Faros AI measure AI coding impact?

It ingests telemetry from GitHub Copilot, Cursor, Windsurf, Claude, Devin, and Amazon CodeWhisperer — adoption, acceptance rates, lines generated, percent AI-generated code by repo — then correlates them with PR cycle time, review duration, bugs, and incidents using baseline-based cause-and-effect analysis.

How is Faros AI different from Swarmia?

Faros is a heavyweight data-aggregation platform aimed at large enterprises with 100+ integrations and custom schemas; Swarmia is a lighter-weight engineering effectiveness tool with transparent per-seat pricing aimed at mid-size teams.

Executive Summary

Faros AI is an enterprise engineering intelligence platform founded in 2019 by Vitaly Gordon (CEO, previously VP of Engineering at Salesforce and a founder of Salesforce Einstein), Matthew Tovbin, and Shubha Nabar — three Salesforce Einstein alumni who left in 2019 to build the data platform they wished their own engineering org had.[1] The product connects to engineering systems — GitHub, Jira, CI/CD pipelines, incident management, and AI coding assistants — and aggregates them into dashboards covering DORA metrics, software delivery, developer experience, and, increasingly, AI coding-tool impact.[2][3]

The company has leaned hard into AI measurement as its growth wedge. Its AI impact module tracks GitHub Copilot, Cursor, Windsurf, Claude, Devin, and Amazon CodeWhisperer adoption against real delivery outcomes, and its April 2026 AI Engineering Report — built on two years of telemetry from 22,000 developers across 4,000+ teams — became an industry talking point by finding that high AI adoption correlated with 54% more bugs per developer alongside 34% more tasks completed.[4][5] Funding is more modest than the category leaders: $36M publicly disclosed ($16M seed; $20M Series A led by Lobby Capital, June 2023), with PitchBook listing $39.8M total raised and no publicly announced Series B as of June 2026.[1][3][6]

AttributeValue
CompanyFaros AI, Inc.
Founded2019, by Vitaly Gordon (CEO), Matthew Tovbin, and Shubha Nabar — ex-Salesforce Einstein[1]
Funding$36M disclosed: $16M seed (SignalFire, Salesforce Ventures, Global Founders Capital) + $20M Series A led by Lobby Capital (June 2023); PitchBook lists $39.8M total[1][3][6]
CustomersAutodesk, Coursera, Discord, Vimeo, Box, GoFundMe, Riskified[3][1][4]
Open SourceFormerly — Faros Community Edition launched 2022, repo now removed from GitHub[1][7]
HeadquartersUnited States (founded out of the Salesforce Einstein team)[1]

Product Overview

Faros positions itself as the single pane of glass for engineering operations: connect every system a software org runs on, normalize the data into one schema, and answer leadership questions — where is delivery slow, what did the AI tooling spend buy, which teams are healthy — with dashboards instead of spreadsheets.[2][1] The AI-impact module is the newest and most-marketed surface: it claims customers see measurable value within one day of a proof-of-concept, with marketing citing outcomes of "up to 10x higher PR velocity, 40% fewer failed outcomes."[4]

Key Capabilities

CapabilityDescription
Data aggregation100+ tool integrations — Git, Jira, Azure DevOps, CI/CD, incident management — normalized into one queryable schema[4][3]
DORA & delivery metricsLead time, deployment frequency, cycle time, and review-time dashboards with baselining[2][4]
AI coding-tool trackingAdoption (daily/weekly/monthly), acceptance rates, lines generated by language and editor, percent AI-generated code by repo, license utilization and power-user identification[4]
AI impact attributionBaseline-based "cause-and-effect analysis" tying AI usage to PR merge rates, cycle times, review duration, test coverage, bugs, and incidents[4]
Developer surveysSatisfaction surveys layered onto telemetry for experience measurement[4]
Benchmarking researchAI Engineering Report built on customer telemetry (22,000 developers, 4,000+ teams)[5][8]

Product Surfaces

SurfaceDescriptionAvailability
Web dashboardsProductivity, DORA, and AI-impact analyticsAll tiers[9]
AI impact moduleCopilot/Cursor/Claude/Devin/Windsurf/CodeWhisperer tracking and ROI analysis[4]GA
API & schema accessProgrammatic access; full schema access on Ultimate tier[9]Enterprise/Ultimate
Custom connectorsCustom and hybrid connectors beyond native integrations[9]Enterprise+

Technical Architecture

Faros is a managed SaaS platform built around an ingestion layer (native SaaS connectors, plus open-sourced connector tooling derived from Airbyte), a normalized engineering data schema, and an analytics layer on top.[1][7] In 2022 the company launched Faros Community Edition, an open-source version of the full product; as of June 2026 the faros-community-edition repository returns a 404 and no community-edition repo remains in the public GitHub org, whose largest remaining public repo is a 34-star CLI utility — the open-source strategy has been quietly wound down.[1][7]

Key Technical Details

AspectDetail
DeploymentManaged SaaS; tiered SaaS/custom/unlimited connector access[9]
AI tools trackedGitHub Copilot, Amazon CodeWhisperer, Cursor, Windsurf, Claude, Devin[4]
Integrations100+ tools across SCM, ticketing, CI/CD, incidents, and AI assistants[4]
Open SourceNo (Community Edition discontinued; remaining public repos are small CLIs and SDKs)[7]

Strengths

  • Breadth of aggregation — 100+ integrations normalized into one schema is the widest data footprint in the category, and the original founding premise: engineering orgs run on more systems than any dashboard covers[4][1]
  • AI measurement is real, not bolted on — tracks six AI coding tools today with acceptance rates, percent AI-generated code by repo, and license utilization, tied to delivery and quality outcomes rather than usage vanity metrics[4]
  • Credible research flywheel — the AI Engineering Report 2026 (22,000 developers, 4,000+ teams, two years of telemetry) produced findings honest enough to cut against AI hype — bugs per developer up 54%, incidents-per-PR more than tripled at high adoption — which earned third-party press coverage and buyer trust[5][8]
  • Pedigreed founding team — Salesforce Einstein founders with ML-at-scale backgrounds, backed by Salesforce Ventures, SignalFire, and Lobby Capital[1][3]
  • Enterprise logos — Autodesk, Coursera, Discord, Vimeo, Box, GoFundMe, and Riskified are named customers[3][1][4]

Cautions

  • The open-source story evaporated — Faros Community Edition was launched in 2022 as a free, self-hosted version of the full product and marketed as a differentiator; the repo has since been removed from GitHub without a public deprecation notice, leaving teams who adopted it stranded and signaling a hard pivot to closed-source enterprise sales[1][7]
  • No public pricing, heavyweight contracts — three tiers, all demo-gated; a competitor analysis claims typical contracts start around $150K/year, and that customers under 1,000 engineers report needing 1–2 FTE data engineers to maintain custom dashboards and pipelines (competitor-sourced, so discount accordingly — but directionally consistent with the platform's complexity)[9][10]
  • Funding stage lags the marketing — $36M disclosed across seed and a June 2023 Series A; no Series B has been publicly announced despite a series-b.faros.ai investor-deck page existing, and the company has not disclosed ARR or customer counts since 2023[3][6]
  • Attribution is correlation with baselines, not experiments — "cause-and-effect analysis" means baselining and cohort comparison across teams, not randomized trials; per-line AI attribution (which specific AI-generated code caused which bug) is beyond what the telemetry supports[4]
  • Its own research undercuts simple ROI claims — the same report Faros markets with found median review time up fivefold and incidents-per-PR up 242.7% at high AI adoption, which complicates the "prove your Copilot ROI" pitch the AI module sells[5]
  • Vendor-published outcome stats — "up to 10x higher PR velocity, 40% fewer failed outcomes" are marketing claims without named-customer methodology behind them[4]

Pricing & Licensing

TierPriceIncludes
ProfessionalCustom (demo-gated)SaaS tool connectors, SSO, productivity benchmarks, AI productivity consultant[9]
EnterpriseCustom (demo-gated)Custom and hybrid connectors, SSO/SAML, API access, advanced RBAC[9]
UltimateCustom (demo-gated)Unlimited connectors, full schema access[9]

Licensing model: Proprietary closed-source SaaS, annual enterprise contracts quoted in US dollars via order form; the formerly open-source Community Edition is no longer available.[9][7]

Hidden costs: Connector rollout and data-quality work; a competitor analysis claims sub-1,000-engineer customers carry 1–2 FTE of internal data-engineering maintenance ($200–400K fully loaded) on top of contracts it pegs as starting around $150K/year.[10]

Competitive Positioning

Direct Competitors

CompetitorDifferentiation
SwarmiaSwarmia is lighter-weight with transparent per-developer pricing for mid-size teams; Faros is a heavyweight aggregation platform with custom schemas sold top-down to large enterprises
LinearBLinearB pairs metrics with workflow automation (PR routing, goals); Faros goes deeper on data breadth and AI-impact attribution across 100+ sources[4]
CortexCortex approaches from the internal-developer-portal side (service catalogs, scorecards); Faros approaches from analytics — they overlap on engineering visibility but solve different primary jobs
JellyfishJellyfish anchors on engineering-to-business alignment (allocation, R&D capitalization); Faros anchors on telemetry breadth and AI measurement, and Jellyfish actively markets against it[10]
DX / GetDXDX leads with developer-experience research instruments; Faros leads with system-of-record telemetry, layering surveys on top[4]

When to Choose Faros AI Over Alternatives

  • Choose Faros when: you run 1,000+ engineers across many tools and need one normalized schema for DORA, delivery, and AI-impact reporting — and have the budget and data-engineering capacity to operate it[10]
  • Choose Swarmia or LinearB when: you want self-serve setup and transparent pricing for a mid-size org without an enterprise sales cycle
  • Choose Cortex when: your primary problem is service ownership and production standards, not analytics
  • Choose a specialist (e.g., GitClear) when: you need per-commit AI code attribution rather than platform-level adoption-to-outcome correlation

Ideal Customer Profile

Best fit:

  • Large enterprises (1,000+ engineers) with sprawling toolchains and a platform/DevEx team to own the deployment[10]
  • Engineering leadership that must defend AI tooling spend to a CFO with adoption-to-outcome data across multiple AI assistants[4]
  • Organizations standardizing DORA and delivery reporting across many business units[2]

Poor fit:

  • Teams under a few hundred engineers — contract size and maintenance overhead are disproportionate[10]
  • Buyers requiring open-source or self-hosted deployment — that option existed and was withdrawn[7]
  • Teams wanting per-line AI code attribution or commit-level provenance rather than cohort-level impact analysis

Viability Assessment

FactorAssessment
Financial HealthModerate — $36M disclosed, last public raise June 2023; no Series B announced and no ARR disclosure since the Series A[3][6]
Market PositionCredible enterprise player — named logos (Autodesk, Discord, Vimeo, Box) and category-defining research, but squeezed between Jellyfish/DX above and Swarmia/LinearB below[3][10]
Innovation PaceStrong on AI measurement — six AI tools tracked, AI Engineering Report 2026 published April 2026[4][5]
Community/EcosystemWeak — open-source Community Edition withdrawn; public GitHub footprint reduced to small utilities[7]
Long-term OutlookCautiously positive — AI-impact measurement is a growing budget line, but the company needs a new funding event or disclosed growth to confirm trajectory

Faros has the strongest data-breadth story in engineering intelligence and was earlier than most to rigorous AI-impact measurement — its own telemetry research is the category's most-cited evidence that AI coding gains come with quality costs.[5] The open questions are commercial: a three-year-old Series A, an unannounced Series B, a withdrawn open-source edition, and competitor-documented six-figure contract economics all suggest a company betting everything on the large-enterprise motion.[6][10]

Bottom Line

Faros AI is the data-platform heavyweight of AI engineering intelligence: 100+ integrations, DORA and delivery analytics, and the most complete multi-tool AI measurement module shipping today — tracking Copilot, Cursor, Windsurf, Claude, Devin, and CodeWhisperer against real delivery outcomes.[4] Its honest, large-sample research (bugs up 54%, incidents-per-PR tripled at high AI adoption) is both its best credibility asset and an implicit warning label for the category it sells into.[5] The trade-offs: opaque six-figure-class pricing, real operational overhead, a quietly abandoned open-source edition, and a funding story that has not been publicly updated since June 2023.[10][7][3]

Recommended for: Large enterprises (1,000+ engineers) that need defensible, multi-tool AI-impact reporting and unified delivery metrics, with platform-team capacity to run it.

Not recommended for: Mid-size teams, open-source/self-host requirements, or buyers needing per-line AI code attribution.

Outlook: Watch for a Series B announcement or ARR disclosure — the AI-measurement wedge is well-timed, but the company's public financial signals are two-plus years old, and faster-moving, cheaper rivals are marketing aggressively against its contract economics.[6][10]

Research by Ry Walker Research • methodology

Sources

Share on XShare on LinkedIn