When a shipment to a local taco shop risked arriving late, a human-sounding voice called the store’s manager to notify of the delay and reschedule the delivery, answer the manager’s questions, and text a live tracking link. It even suggested a recipe for the evening’s special taco.
The voice belonged to artificial intelligence (AI), not a person. And the taco shop manager? That was Kiren Sekar, Samsara’s chief product officer, roleplaying during a live demonstration at the 2025 Samsara Beyond keynote presentation, which kicked off its annual three-day user conference in San Diego, Calif.
The exchange was powered by HappyRobot — a company that Samsara has recently invested in and integrated its AI voice tech into Samsara’s platform — was just one example of how AI can redefine fleet operations.
During the keynote, the company has showcased what it calls its largest set of product announcements to date — unveiling proprietary tools that stretch across every layer of fleet operations, from maintenance and routing to safety and customer communications.
The company framed this shift as part of a broader technological transformation. During his keynote remarks, co-founder and CEO Sanjit Biswas outlined what he described as three historic waves of innovation: the age of computation, driven by the personal computer; the age of communication, shaped by the smartphone; and now, the accelerating “age of intelligence,” powered by AI.
“It took 25 years for computers, eight years for smartphones, and just two years for AI,” Biswas said of adoption curves of each wave, adding that AI reached a billion users in just two years, since ChatGPT’s launch in 2023. “It happens fast because we’re ready. And at this point, we’re entering an entirely new era, the age of intelligence…It’s so incredible and so useful, especially in our world of physical operations.”
AI-enhanced inspections, end-to-end maintenance
The evolution begins with the driver walkaround. Samsara has reworked its Driver Vehicle Inspection Report (DVIR) process, which will be available later this year, to prevent shortcuts and improve accuracy using location validation, image recognition, and voice-to-text features.
“Drivers are collectively completing a quarter of a billion DVIRs in Samsara every year,” said Kiren Sekar, chief product officer at Samsara, during the presentation. “It turns out 60% of all DOT violations are maintenance-related.”
Drivers using this feature won’t be able to submit old or offsite photos. Instead, a location-based prompt ensures they are near the vehicle, and the camera can identify specific truck components — accepting or rejecting images accordingly. Voice memos are automatically transcribed, and AI helps classify defect severity.
Inspection results flow directly into Samsara’s new native Asset Maintenance platform, that is already available in early access. Fleet managers can now view prioritized fault codes, schedule repairs, and track work orders — all within the same dashboard.
AI interprets fault codes and explains their impact, urgency, and regulatory implications. In one demo, an engine fault involving the aftertreatment system was flagged as high-priority due to performance and compliance risks. The AI assistant suggested repair steps personalized to the vehicle’s service history.
“At this point, traditionally, we would move outside of the Samsara platform into a third-party maintenance application…but we wanted to help solve this problem,” Sekar said.
Samsara now offers complete end-to-end maintenance management, allowing users to create and assign work orders, schedule repairs, and analyze cost data — all within the platform. Maintenance leaders can filter costs by labor, parts, fuel, geography, or vehicle model to identify trends and optimization opportunities.
The platform also supports third-party maintenance invoice processing, Sekar said. Fleets can upload a PDF of a repair bill and Samsara’s AI will analyze the line items, categorize them by service type and cost, and auto-fill the work history.
“For a complex repair, your teams are spending half an hour plus each time,” Sekar said. “[AI] is going to actually understand and reason about onto the line items and figure out what are parts, what are labor expenses, what are our service tasks that these translate to. It does that all almost instantly, and now we can see a complete service history filled out in just a couple seconds with zero manual data.”
Smarter routing and commercial navigation
Fleets using legacy dispatch systems will now have a new option: Samsara’s native AI-powered Route Panning tool. Users can import orders via spreadsheet or API and generate delivery routes based on available equipment, driver shifts, order types, required certifications, and even refrigeration needs.
The system recognizes that reefer units, for example, cost more per hour and factors that into route optimization, along with traffic, stop time, and historical delivery behavior.
“We just want to make all of these 75 deliveries with the lowest possible cost and no overtime hours,” Sekar said during a demo. The AI-generated routes used only six out of eight available vehicles, completing all stops in under 350 miles with zero overtime.
Drivers then receive their routes directly in the Samsara Driver App — now with built-in turn-by-turn commercial navigation. Just like the route planning tool, it will be available this summer.
The system includes truck-specific data such as height, weight, and hazardous material restrictions, plus voice instructions and live traffic updates. This eliminates the need for drivers to use third-party tools like Google Maps or Waze, which often send trucks down restricted or unsuitable roads, as well as the need to switch between the screens.
Sekar added that the system works for routes dispatched through Samsara, as well as any for routes the driver chooses to enter manually.
Real-time weather intelligence
Weather remains a persistent disruptor in logistics, and Samsara is introducing new tools to help fleets plan and react to conditions on the ground.
The new Weather Intelligence dashboard overlays radar data and National Weather Service (NWS) alerts on top of asset maps. Users can see which drivers and vehicles are affected by events like fog, flooding, extreme heat, or more. Dashcam feeds from Samsara-connected trucks can be pulled up to view conditions in real time — eliminating the need for check-in calls like “Is it raining yet? How bad is it?”
But the most ambitious feature may be StreetSense, an opt-in network that uses anonymized dashcam footage from across Samsara’s customer base to build a live stream of weather conditions across North America. Images are stripped of personally identifiable information and updated every 20 to 40 seconds.
The system also uses AI to translate complex weather alerts into short, spoken alert messages. These can be broadcast to drivers in real-time through their Samsara dashcams — alerting them to conditions ahead without needing to stop and check an app. “NWS has issued a flood watch for the West Texas mountain and plains region starting around noon. The Beyond crowd wants you safe,” the AI-generated voice said during a demo.
Voice AI meets customer service
One of the more surprising demos came midway through the keynote, when Sekar staged a live call between an AI voice assistant and himself, roleplaying as that taco shop manager. The AI rescheduled a delivery, texted a tracking link, and responded various requests.
Had I not known this was AI in advance, I would not be able to tell – the voice used contractions, filler words like “umm” and “yep, real quick…”, and subtle hesitations — the kinds of imperfect, unscripted patterns that I once relied on to distinguish humans from AI. The crowd audibly reacted, laughing and murmuring as the voice handled the call, even more so after it suggested a taco recipe. When asked directly if it was an AI, it admitted so.
This integration comes from Samsara’s investment in voice AI company HappyRobot, which has been deployed in customer service, payment collection, and driver recruiting workflows. “Our customers are using HappyRobot today to automate over a million interactions like this one every single week,” said co-founder and CEO of HappyRobot Pablo Palafox while on stage.
Driver safety redefined
At the same time, Samsara is expanding its safety suite with AI MultiCam — a 360-degree external camera system with pedestrian detection and real-time alerts, which will be available in summer. The cameras are designed for vehicles like school buses, refuse trucks, or box trucks operating in tight spaces.
In one demonstration, the system identified a cyclist in the truck’s blind spot and issued an alert to prevent a potential collision. MultiCam integrates with Samsara’s existing inward and road-facing cameras and is compatible with third-party camera systems.
AI also now contextualizes driver behavior. One dashcam clip showed a driver braking hard to avoid a striking a pedestrian. The system flagged it as “defensive driving” and rewarded the driver instead of penalizing them.
Mobile phone use — even when partially concealed — is detected by the system and automatically triggers targeted micro-training videos. Drivers receive coaching in ‘snackable’ formats rather than mandatory long videos.
A new AI-powered dashboard highlights drivers most at risk of causing the next incident, using over 30 variables like tenure, coaching history, time of day, and event severity. It also flags top performers and enables immediate recognition through a built-in rewards and communications platform.
Worker safety beyond the cab
Rounding out the platform evolution, Samsara introduced a new Worker Safety system, also available in summer 2025, that includes a wearable device – that will be on the market later this year – built for rugged environments. It clips onto the outwear, includes an SOS button, speaker, and display, and lasts up to a year on one battery.
Instead of relying on cellular networks, the device sends signals to nearby Samsara-connected vehicles or mobile phones to reach the cloud. In an emergency, workers can record messages, transmit locations, or trigger real-time check-ins. Missed check-ins initiate automated alerts back to fleet managers who can tap into livestream data from the device.
