← Back to research
·10 min read·company

Milestone

Milestone (mstone.ai) is an enterprise platform for measuring GenAI adoption, productivity, and ROI across engineering teams — $10M seed (November 2025) led by Heavybit and Hanaco, with GitHub, Monday.com, and Kayak as customers and a claimed >90% AI-attribution accuracy.

https://mstone.ai/ ↗
engineering-intelligenceenterprisegithub

Key takeaways

  • Raised a $10M seed in November 2025 led by Heavybit and Hanaco Ventures, with Atlassian Ventures and angels including GitHub co-founder Tom Preston-Werner — and counts GitHub itself, Monday.com, Kayak, and Sapiens among customers
  • Builds a "GenAI data lake" from four read-only pillars — codebases, project management platforms, team structure, and AI codegen tools — and claims over 90% accuracy attributing code to AI
  • Deliberately enterprise-only with custom pricing and a demo-led sales motion; the dashboard targets VPs of Engineering and the C-suite rather than code-level review workflows

FAQ

What is Milestone?

Milestone is an engineering-intelligence platform that measures the adoption, productivity impact, and ROI of generative AI coding tools by correlating AI usage data with real repository and delivery metrics.

How much does Milestone cost?

Pricing is not publicly listed. Milestone sells to enterprises through a book-a-demo motion with custom quotes, and has deliberately declined smaller prospective customers.

How does Milestone measure AI impact?

It ingests read-only data from codebases, project management tools, team structure, and AI tools (Copilot, Cursor, Claude, Qodo, Augment, and others) into a GenAI data lake, then attributes code to AI with a claimed >90% accuracy and correlates usage with delivery speed, code quality, and bug rates.

How is Milestone different from GitClear?

GitClear leads with code-level quality forensics (churn, duplication, copy-paste trends); Milestone leads with executive-level ROI reporting that joins AI usage to delivery and quality metrics across teams.

Executive Summary

Milestone (mstone.ai) is an enterprise platform for measuring what generative AI is actually doing to an engineering organization — adoption, productivity, code quality, and ultimately ROI. It was founded by CEO Liad Elidan and CTO Stephen Barrett, a Trinity College Dublin computer science professor who was Elidan's university lecturer; the Israeli company spans Israel and Ireland and grew out of years of joint research on software engineering metrics.[1] The platform builds what the company calls a "GenAI data lake" from four read-only pillars — codebases, project management platforms, team structure, and the AI codegen tools themselves — and correlates AI usage with feature delivery speed, bug rates, and long-term code maintainability.[1]

In November 2025 Milestone raised a $10M seed led by Heavybit and Hanaco Ventures, with participation from Atlassian Ventures and angels including GitHub co-founder Tom Preston-Werner, former AT&T CEO John Donovan, Accenture senior tech advisor Paul Daugherty, and ex-Datadog president Amit Agrawal.[1][2] Note: this round is a seed, not a Series A as sometimes reported. Customers include GitHub (whose COO Kyle Daigle provides a homepage testimonial), Monday.com, Kayak, Sapiens, Playtika, OverIT, Carbyne, Bluebricks, and Varonis.[3][4][1]

AttributeValue
CompanyMilestone (mstone.ai)
FoundersLiad Elidan (CEO), Prof. Stephen Barrett (CTO, Trinity College Dublin)[1]
Funding$10M seed (November 2025), led by Heavybit and Hanaco Ventures, with Atlassian Ventures[1][2]
CustomersGitHub, Monday.com, Kayak, Sapiens, Playtika, OverIT, Varonis, and others[3][4]
HeadquartersIsrael, with CTO/research roots in Dublin[1]
Open SourceNo — closed source, proprietary SaaS

Product Overview

Milestone's pitch is "avoid shooting in the dark with your GenAI" — replacing seat-count reporting and developer surveys with measurement grounded in repository data.[3] The platform tracks three dimensions: GenAI adoption and usage (who is using which tools, how often), GenAI productivity (velocity and throughput effects), and GenAI quality (stability and bug patterns in AI-impacted code), with the headline claim of over 90% accuracy in attributing code to AI.[3] Customers typically see meaningful data within the first week of connecting integrations, and the company says the most common discovery is uneven adoption — with longer review times partially offsetting coding speed gains.[3]

Key Capabilities

CapabilityDescription
GenAI impact measurementQuantifies how AI-assisted coding affects the development lifecycle and team performance[4]
AI attributionAttributes code to AI tools with claimed >90% accuracy[3]
Productivity analyticsTime allocation, trend identification, PR cycle time, and code stability monitoring[4][5]
Engineering investmentResource allocation across repositories; maintenance vs. new development[5]
Multi-level insightsMetrics roll up from individual contributors to teams, groups, and projects, with role-specific dashboards[4]
Predictive planningBottleneck identification and delivery forecasting[4]

Technical Architecture

Milestone is a closed-source platform offered as cloud-hosted SaaS or on-premise, connecting to existing tools through read-only integrations — it ingests events and metadata rather than modifying anything in the development pipeline.[5] The four-pillar data model (code, project management, team structure, AI tools) feeds the GenAI data lake that powers attribution and correlation.[1] The company has partnerships with GitHub, Augment Code, Qodo, Continue, and Atlassian.[1]

Key Technical Details

AspectDetail
DeploymentCloud-hosted SaaS or on-premise; read-only integrations[5]
AI tool integrationsGitHub Copilot, Cursor, Claude, Amazon Bedrock, Qodo, Augment[3][4]
Dev-stack integrationsGitHub, GitLab, Bitbucket, Azure DevOps; Jira, Asana, Linear, Fibery; HiBob, Workday, CSV for org structure[3][4]
ComplianceSOC 2 Type 2; public Trust Center[3][4]
Open SourceNo

Strengths

  • Credible category-specific backing — $10M seed from Heavybit (developer-tools specialist) and Hanaco, plus Atlassian Ventures and angels who built GitHub, Datadog, and Accenture's tech practice[1][2]
  • GitHub as both partner and customer — GitHub's COO publicly endorses the platform for visibility into "how AI-driven engineering practices scale in real environments," an unusual validator for a seed-stage measurement vendor[3]
  • Four-pillar data model — joining code, project management, org structure, and AI-tool telemetry is broader than git-only analysis, enabling team-level ROI questions git data alone cannot answer[1]
  • Read-only, low-risk deployment — no agents in the pipeline and no write access; on-premise available for codebase-sensitive enterprises[5]
  • Academic grounding — the CTO is a computer science professor whose research on engineering metrics predates the GenAI wave[1]
  • Honest-sounding findings — the company itself reports that customers commonly discover longer review times offsetting AI coding speed gains, rather than only positive ROI stories[3]

Cautions

  • Seed-stage, not Series A — the November 2025 round is a $10M seed; total disclosed funding is $10M, and revenue is undisclosed[1]
  • Codebase access is the asking price — investors themselves initially questioned a model requiring access to customer codebases; security-conservative enterprises will scrutinize this even with read-only scope and SOC 2 Type 2[1][3]
  • Incentive alignment question — Elidan notes no customer has ever used Milestone and concluded GenAI doesn't help; a tool sold to justify GenAI spend has a structural incentive to find positive ROI[1]
  • The category's premise is contested — critics like Ed Zitron argue AI ROI measurement is built on sand: a Bain survey of 951 executives found only 4% achieved AI savings above 30%, and faster code generation can net out negative once review burden is counted[6]
  • No public pricing or self-serve — demo-led enterprise sales only, and the company deliberately turns away smaller prospects[3][1]
  • Less code-level depth than forensic tools — the product is an executive dashboard over correlations, not a code-quality microscope; teams wanting churn/duplication forensics need a GitClear-style tool

Pricing & Licensing

TierPriceIncludes
EnterpriseCustom (not publicly listed)Full platform — adoption, productivity, and quality analytics; cloud or on-premise deployment; demo-led sales with an ROI calculator[3][5]

Licensing model: Proprietary closed-source SaaS (or on-premise license), annual enterprise contracts.[5]

Hidden costs: Integration and org-structure setup (HR data via HiBob, Workday, or CSV), plus the internal process work of acting on findings — the dashboard surfaces problems; fixing uneven adoption is on the buyer.[4]

Competitive Positioning

Direct Competitors

CompetitorDifferentiation
GitClearGitClear leads with code-level quality forensics — churn, duplication, copy-paste trends — from git history; Milestone leads with executive ROI reporting that joins AI-tool telemetry to delivery metrics
Exceeds AIExceeds goes deeper on per-commit AI detection and code-level analysis; Milestone is broader and more C-suite-oriented, with org-structure and project-management data in the model
Faros AIFaros is a full engineering-operations platform (DORA, allocation, surveys) that added AI measurement; Milestone is GenAI-impact-first and narrower by design
DX / JellyfishEstablished engineering-intelligence platforms adding AI-measurement modules to broader productivity suites; Milestone is a focused newcomer with AI attribution as the core product

When to Choose Milestone Over Alternatives

  • Choose Milestone when: the buyer is a VP of Engineering or C-suite needing board-ready GenAI ROI reporting joined across code, delivery, and org structure — especially with on-premise requirements[5]
  • Choose GitClear when: you want code-quality forensics and AI-impact research grounded in commit-level diff analysis
  • Choose Exceeds AI when: per-commit AI detection depth matters more than executive dashboards
  • Choose Faros AI when: you want one platform for all engineering operations, with AI measurement as one module

Ideal Customer Profile

Best fit:

  • Enterprises with hundreds-plus engineers and meaningful GenAI license spend to justify (Copilot, Cursor, Claude)[3]
  • VPs of Engineering / CTOs who must report GenAI ROI upward to the CFO or board
  • Organizations on GitHub/GitLab/Bitbucket with Jira-class project management already in place[4]
  • Regulated or codebase-sensitive companies that need on-premise deployment[5]

Poor fit:

  • Small teams and startups — Milestone deliberately declines smaller customers[1]
  • Teams wanting code-level quality forensics rather than executive analytics
  • Buyers requiring public pricing or self-serve onboarding
  • Organizations unwilling to grant read-only codebase access to a third party[1]

Viability Assessment

FactorAssessment
Financial HealthEarly but well-backed — $10M seed (November 2025) from Heavybit, Hanaco, and Atlassian Ventures; revenue undisclosed[1][2]
Market PositionStrong logos for seed stage (GitHub, Monday.com, Kayak); crowded field as DX, Jellyfish, and Faros add AI modules[3][1]
Innovation PaceFocused — staying on engineering GenAI measurement, explicitly declining to expand into marketing or other functions[1]
Community/EcosystemPartnership-driven — GitHub, Atlassian, Qodo, Augment Code, Continue; no open-source community[1]
Long-term OutlookPromising niche with two risks: incumbents bundling AI measurement, and the ROI-measurement premise itself being contested[6]

Milestone has the right investors, the right logos, and a defensible four-pillar data model for the moment when every CFO starts asking what the GenAI line item is buying. The open questions are whether a seed-stage pure-play can outrun broader engineering-intelligence platforms adding AI modules, and whether enterprises keep paying for ROI measurement if the broader AI-ROI narrative sours.[6]

Bottom Line

Milestone is the executive-dashboard entrant in AI engineering intelligence: a $10M-seeded, enterprise-only platform that joins read-only code, project, org, and AI-tool data to tell VPs of Engineering whether GenAI spend is working — with GitHub itself as a marquee customer.[1][3] It trades code-level depth for breadth and C-suite legibility, and its honest acknowledgment that review times often offset coding gains lends it more credibility than most vendor dashboards.[3]

Recommended for: Enterprise engineering leaders who must quantify and report GenAI ROI across teams, especially those needing on-premise deployment and a vendor-neutral view across Copilot, Cursor, and Claude.

Not recommended for: Small teams, code-forensics buyers, or organizations that cannot grant third-party codebase access.

Outlook: Cautiously positive. The seed round, category-specialist investors, and GitHub/Atlassian partnerships position Milestone well for the budget-scrutiny phase of enterprise GenAI — but it must convert focus into depth before the engineering-intelligence incumbents close the gap.[1][2]

Research by Ry Walker Research • methodology

Sources

Share on XShare on LinkedIn