Ridecell

📊 Most AI benchmarks do not really test domain specific knowledge. Asset spend work is anything but general, so we built our own: Ridecell SpendBench. It tests leading open-source AI models on the work that matters to asset operators: reading real spend related docs like invoices accurately, and consistently making the right approve-or-reject call. It's how we evaluate the models enabling Ridecell's Intelligence Platform, measured on real asset work and not generic leaderboards. Quick walkthrough below 👇

It's 2:47 AM. Your fleet is still running, and so is Maya. Introducing Maya, Ridecell's Gen-AI agent working on top of our Knowledge Graph specialized for automotive leasing and finance: → catching cost spikes in work orders → finding assets stalled in the repair queue → flagging suppliers who cost more than they should → answering driver questions straight from lease terms Autonomous when you need it. Assistive when you ask. Find out more at https://ridecell.com/maya/

🚚 Every day, fleet operators make hundreds of vehicle assignment decisions. Most still run on spreadsheets and the cost shows up later: as service collisions, avoidable swaps, and margin quietly leaking off the P&L. Fleet visibility only tells you what’s happening but Fleet Intelligence determines what should happen next. Dhruva Dandekar, our Director of Technical Product Management, unpacks the shift in our latest post. 👉 Read it here: https://lnkd.in/eW8wsizd

⏱️ We cut writing release documentation from 8 hours to 15 minutes with AI. Writing good release notes is critical to communicate the value of the awesome innovation the tech teams at Ridecell deliver at breakneck speed. Release notes at Ridecell used to take one product lead about eight hours per release because four products and four different release conventions were all documented by hand in a spreadsheet. Now a pipeline kicks off when a branch is cut or a tag is pushed. It pulls the tickets and linked support cases, drops the noise commits, and has secure AI write the customer-facing summary for each one. It takes about 15 minutes end to end with no misses since we turned it on. But speed should not come at the expense of quality, and we don't want to send AI written slop to our beloved customers. Key learning for us to get the quality high: the prompt mattered much lesser than we thought it would. Weeks of tuning changed the output quality slightly but getting engineers to set fix versions and write one sentence in their words in the ticket itself changed everything. The automation gives each row representing a fix or feature a confidence score. Low ones get flagged for a human, plus security-sensitive changes are filtered out before the model ever sees them. Read the full story here: https://lnkd.in/eNTrdPJm

On behalf of Ridecell, I am happy to be a signatory of this important letter and an early member of the Little Tech Association. Competition makes us better, and open weight models that can be hosted freely in the United States, Europe and other allied countries will benefit the U.S. and the world as a whole. https://lnkd.in/gDi3JSt7 Thank you to Garry Tan of Y-Combinator and Harry Godfrey of Little Tech for advancing open systems, free competition and empowering builders.

On recrute 3 profils tech à Paris (et Bordeaux) — et c'est le meilleur moment pour nous rejoindre. 👇 Chez Ridecell, on construit la plateforme de fleet intelligence n°1 mondiale : on apporte l'IA aux acteurs du leasing et de la flotte — Arval BNP Paribas, KINTO by Toyota, Merchants Fleet — en transformant 100+ systèmes fragmentés en une seule couche de décision. Concrètement : 12M+ d'événements traités / jour, 500K+ véhicules, 14+ pays. Startup YC (W12), petite équipe, vraie vélocité, et l'IA utilisée pour de vrai — pas sur un slide. Les 3 postes ouverts (Paris ou Bordeaux, hybride, CDI) : 📊 Data Engineer (3+ ans) — tu prends la main sur nos systèmes de données et tu les transformes en warehouses propres. Python, SQL, Trino, Dagster, un peu de ML si ça te branche. 🤖 Solutions Engineer (3+ ans) — tu es l'architecte qui transforme les process les plus complexes de nos plus gros clients en workflows automatisés et agents IA. Tu décides où l'IA remplace des tâches manuelles et fastidieuses, tu conçois l'intégration, et tu emmènes le tout jusqu'en prod. Un rôle où tu vois ton impact tourner en vrai, chez des leaders européens de la flotte. 🤝 Solutions Support Engineer (1+ an) — point de contact technique des clients : intégrations, résolution d'incidents, accompagnement pré/post-lancement. Un vrai tremplin. Ce qu'on offre : de l'ownership réel, des comptes stratégiques (pas du mid-market générique), et une équipe où ce que tu repères sur le terrain nourrit directement la roadmap. Français + anglais requis. Autorisation de travail en France (pas de sponsoring visa). Ça te parle, ou tu connais la bonne personne ? Liens vers les offres en commentaire 👇 — ou écris-moi en DM / à francois@ridecell.com.

🙋 Your clients shouldn't have to call you to find out what's up with their assets. With Ridecell Client Intelligence, they get a live self-service view of their own assets: lifecycle status, total cost, EV and ESG reporting, and proactive alerts before something goes off-track. And for you as the FMC or Automotive Finance leader: 30-40% fewer support requests, higher satisfaction, and stronger renewals, with premium insights you can offer as an add-on monetized service. Read more here: 👉 https://lnkd.in/g3FQYfrs

How simplifying our maintenance orchestration stack made us 12x faster 🚀 Splitting a tech system into too many pieces causes massive lag. Think of it like trying to order a coffee, but the barista has to call three different managers in three different cities just to approve your milk choice. The engineering team at Ridecell fixed this by bringing sliced systems together into one unified powerhouse. The results? ⏱️ 12x Faster Operations: A critical job-matching routine went from 60 minutes down to just 5 minutes. 📉 Zero Waiting Around: Field workers get their assignments instantly instead of sitting idle on the clock. 🚫 No More App Freezes: Eliminated peak-hour crashes that used to freeze workers' mobile apps. ⚡ Faster Innovation: Cut deployment times by 77% to ship features much faster. The takeaway? Complexity isn't a badge of honor. True scale comes from choosing the simplest tool for the job. Read the full story here: https://lnkd.in/gtydBufW

🛡️ Data sovereignty is a critical trust factor for enterprise AI. For major leasing and automotive financial services, asset data is treated with the same strict security as a financial ledger. Sending sensitive data to public AI APIs creates unacceptable risk. In the latest blog post from our Head of Compliance Bhumi Shah, CISA, CISM, HiTrust CCSFP, PCI ISA, we detail how Ridecell addresses these security demands through: 🔒 DMZ Customer VPC Deployments: Keeping data firmly within your cloud perimeter. 🌍 IBM Cloud for Financial Services Validation: Meeting banking-grade security standards. 🔀 LLM-Agnostic Design: Bringing the AI to your data to eliminate external exposure. 👉 Read the blog post here: https://lnkd.in/gf77Qsw4

Our VP of Product & Customer Launch, Ashok Sundararajan, breaks down why organizations must move past AI 'productivity theater' and demand real P&L impact. Read his insights on how Ridecell’s deterministic enterprise infrastructure is helping Asset Leasing & Finance leaders turn operational data into measurable savings.