The evidence room is no longer just a physical place. It is a sprawling digital dump of phone logs, financial records, surveillance footage, and witness statements, a haystack where the needle is a conviction or an exoneration. Longeye, a San Francisco startup founded in 2025, is selling a new kind of filing cabinet for that data, one powered by a private AI that investigators can query in plain English. The company’s early traction suggests a real appetite for the tool, with more than 50 justice agencies signed on in its first year of operation [Business Wire, October 2026]. The question for procurement officers is whether a startup can navigate the long sales cycles and stringent technical requirements of the public sector, and whether its AI can maintain the chain of custody that every case demands.
The wedge into the evidence locker
Longeye describes its product as an “AI-powered investigative workspace,” a single interface for analyzing the heterogeneous data types that define modern cases. The wedge is speed and organization, not prediction. Investigators can ask a chatbot questions to extract evidence, construct timelines, and identify discrepancies, with the platform claiming to analyze evidence up to 100 times faster than manual review [Longeye Mid-Market Account Executive posting]. Crucially, every AI-generated finding remains traceable to the original source document, a non-negotiable requirement for courtroom admissibility [Forbes Councils profile]. The platform processes text, images, audio, video, and billions of rows of structured data like financial and phone records [Citybiz, October 2026]. For enterprise buyers in police and sheriff’s departments, the pitch is about closing cases faster by cutting through data overload.
A bet on private, compliant AI
The technical architecture is as important as the interface. Longeye runs its AI models through a private, fully closed system on Amazon Bedrock, with a firm policy that customer data is never used to train its or its providers’ models [Longeye Public Defenders page]. This self-hosted, secure-by-design approach is a direct response to the unique sensitivities of criminal justice data. It is also a strategic moat. Competing with general-purpose AI tools means offering something they cannot: a system built for the evidentiary standards and privacy mandates of the public sector from the ground up. The company has extended this model to public defenders, offering the platform free of charge to approximately 30 defender offices, a move that builds goodwill and expands its footprint within the justice ecosystem [LinkedIn post].
The team and the $25 million war chest
The founders bring a blend of operational scale and mission-driven focus. CEO Guillaume Delépine is a former Skydio leader who left a joint Harvard Business School and Kennedy School program to pursue public safety [Forbes, October 2019]. CTO Danielle Man is an MIT-trained engineer who previously led the 40-person Core Platform team at Apollo GraphQL, scaling the engineering org from 20 to over 200 employees [Longeye About]. That experience in building and managing complex software systems is a relevant credential for a product that must handle billions of data points reliably.
Their vision has attracted $25 million in total funding, including a $20 million Series A led by N47 in October 2026 [Business Wire, October 2026]. The earlier $5 million seed round was led by Andreessen Horowitz’s American Dynamism Fund, which backs companies solving national priorities [Longeye About]. The capital is likely earmarked for scaling the sales motion and further hardening the product for larger, more demanding agency deployments.
2025 Seed | 5 | M USD
2026 Series A | 20 | M USD
Traction and the path to scale
Longeye’s reported metrics paint a picture of rapid early adoption. As of October 2026, its platform was in use at more than 50 justice agencies, which had collectively worked on over 2,300 investigations [Business Wire, October 2026] [Citybiz, October 2026]. The aggregate data processed is staggering: over 15 million pages of documents, 10 million images, 32,000 hours of audio and video, and more than 1 billion rows of financial and telephone records [Citybiz, October 2026]. These numbers are a powerful proof of concept, demonstrating that agencies are willing to feed real, sensitive casework into the system. The current hiring focus on enterprise and mid-market account executives suggests a deliberate push to move upmarket from early adopters to larger, budget-holding departments [Rippling].
Where the wheels could come off
For all its momentum, Longeye faces hurdles intrinsic to its chosen market. Selling to government agencies involves procurement cycles measured in quarters or years, not months. Budgets are often inflexible and subject to political winds. The most credible near-term risk is that growth stalls as the company exhausts its pool of early-adopter agencies and hits the slower pace of mainstream public-sector sales. Furthermore, while the AI claims are compelling, the platform’s ultimate effectiveness will be judged on its ability to withstand legal scrutiny. A single high-profile case where evidence processed by Longeye is challenged on procedural grounds could create a significant headwind.
The company’s most plausible answer is already being built: a relentless focus on security, auditability, and compliance as product features, not afterthoughts. By architecting for the highest standards of evidence handling from the start, Longeye aims to turn those stringent requirements into a competitive barrier.
The next twelve months
The coming year will be about proving the enterprise sales motion. Watch for announcements of contracts with state-level or large metropolitan agencies, which would signal an ability to navigate complex procurement. The other key milestone will be the expansion of the platform’s analytical capabilities, likely moving beyond review and organization toward more proactive investigative leads, all while maintaining the ironclad provenance that is the product’s foundation.
The ideal customer profile here is clear: a mid-sized to large law enforcement agency or prosecutor’s office drowning in digital evidence but constrained by manpower and time. They have a dedicated IT and compliance team, a budget for software that improves operational efficiency, and a pressing need to reduce case backlogs.
The realistic competitive set is fragmented. On one side are legacy evidence management systems, which offer organization but little intelligence. On the other are general-purpose AI and data analytics platforms from large tech vendors, which may lack the specific workflows and legal compliance built into Longeye’s product. The startup’s bet is that by owning the intersection of specialized justice workflows and secure, explainable AI, it can define a new category before others decide to enter it.
Sources
- [Business Wire, October 2026] Longeye Raises $20M Series A as More Than 50 Justice Agencies Adopt Its AI Investigative Platform | https://www.businesswire.com/news/home/20261021359350/en/Longeye-Raises-$20M-Series-A-as-More-Than-50-Justice-Agencies-Adopt-Its-AI-Investigative-Platform
- [Citybiz, October 2026] Longeye Raises $20 Million Series A to Expand AI Investigative Platform | https://www.citybiz.co/article/922569/longeye-raises-20-million-series-a-to-expand-ai-investigative-platform/
- [Longeye About] Longeye About | https://www.longeye.com/about
- [Longeye Mid-Market Account Executive posting] Longeye Mid-Market Account Executive posting | https://ats.rippling.com/longeye/jobs/af12bcf3-2df9-496a-b2d7-40c3700a0fae
- [Forbes Councils profile] Forbes Councils profile | https://councils.forbes.com/profile/Guillaume-Delepine-Co-founder-CEO-Longeye/261180e0-1a94-4267-8f02-7f26752b16f9
- [Longeye Public Defenders page] Longeye Public Defenders page | https://www.longeye.com/public-defenders
- [LinkedIn post] LinkedIn post | https://www.linkedin.com/posts/freightsourcing_xeneta-existing-on-the-market-for-14-years-activity-7492819411880288256-2eaZ
- [Forbes, October 2019] Leaving Business School Isn’t Easy, But Your Passion May Lay Outside Of It | https://www.forbes.com/sites/frederickdaso/2019/10/27/leaving-business-school-isnt-easy-but-your-passion-may-lay-outside-of-it/
- [Rippling] Longeye Careers | https://www.longeye.com/careers