Switch

Switch builds Myles, an agentic AI teammate for fleet and logistics operators. It answers operational questions in plain language, forecasts with stated confidence, and then acts: rebalancing, routing and planning on your city's real data. Covers 3,800+ cities and 350M trips per month.

Switch is the company behind Myles, an agentic AI teammate built for fleet and logistics operators in urban mobility. Rather than bolting a chatbot onto a dashboard, Myles is designed to answer operational questions in plain language and then carry out the resulting actions, either on demand or on a recurring schedule. The product is aimed at ops leads, planners and analysts running shared fleets, micromobility schemes or logistics operations who need fast answers without writing a query or building a report. The core of the product rests on two ideas the vendor calls the where and the when. On the where side, Myles carries the structure of a city: zones, depots, service areas, access windows and infrastructure constraints. On the when side, it holds weekly rhythms, seasonal shifts and weather correlations drawn from operational history. The site cites coverage of more than 3,800 cities and 350 million trips modelled per month, with over 500 live data feeds ingested and a claimed 30 seconds or less from question to answer. Myles grades its own certainty on forecasts and states that confidence alongside the number. In practice the workflow runs in four stages. Operators connect their feeds or start on sample city data, ask a question the way they would ask a sharp colleague, receive both an answer and a proposed next move, then let Myles execute or schedule the work every morning or shift. The stated job families are analytics (idle time, utilisation, shortages, trends), forecasting, operations (rebalancing, routing, pickup priorities) and planning (fleet sizing, breakeven, CO2 and station placement simulated on the operator's own city). Switch positions Myles against general assistants such as ChatGPT, Gemini, Claude and Grok, arguing that a general model has read about mobility but has never been on the road. The pitch is specificity: your city's topology and your fleet's history instead of a generic answer. That makes it a niche tool for mobility and logistics teams rather than a general productivity assistant, and it is still being onboarded through a waitlist.