From GIS context and survey-grade positioning to asset management and field capture, Trimble is making the case for one connected asset lifecycle — a system in which precise location, operational software and shared data reinforce one another, and where the AI layered on top is valued less for novelty than for producing consistent, trustworthy results.
At Trimble’s booth during this year’s Esri User Conference in San Diego, the conversation about artificial intelligence started with a crucial distinction. “When we’re talking about artificial intelligence for the asset life cycle, whether it’s construct or operations and maintenance, if you get a hallucination, that could cost millions of dollars,” said Matthew Sprague, a senior Trimble leader in construction, operations, and maintenance technology, who works on the Unity Suite. “ChatGPT, Gemini, you ask the same question two times in a row, and it’s going to give you slightly different answers. That’s non-deterministic. Where we’re honing in ours is that we want it so that they ask, but it’s going to give you the exact right answer, and the same answer every single time.”
The distinction matters because of what the answer is used for. A general-purpose assistant giving a differently worded but roughly equivalent response twice is a minor annoyance in most contexts; in an asset management workflow, where the output can determine whether a bridge inspection gets scheduled or a work order gets issued, the same question needs to return the same answer, not a plausible-sounding variation on it. “The answers need to be very pragmatic and straight and straightforward,” Sprague said, “and that’s why we get knocked a little bit because we’re slower to release our AI features. We don’t want to rush tomorrow.”
Sprague tied the caution directly to the company’s institutional reputation rather than treating it as a purely technical stance. “We’ve been around for 40-plus years, so our customers know that when we release a new piece of hardware or software, it’s been tested. It’s gone through rigor, QA, and it does what it’s supposed to do,” he said. “Some newer customers that are looking at competitive things are like, hey, why don’t you have this? Because we’re a little bit more methodical behind what we’re doing.” Trimble is not staying out of generative AI entirely — Sprague described a coming, unified message spanning the company’s geospatial devices, Unity, SketchUp, and estimating products rather than each business unit introducing AI features independently. “We don’t want geospatial devices to talk about AI, and we don’t want Unity to talk about AI, and SketchUp to talk about AI, and estimating to talk. We want, what is our holistic message going to be? That’s going to be coming soon.”
One Database, Not Two
The clearest illustration Sprague offered for what Unity Suite actually solves came from a live demonstration rather than a spec sheet. Unity Maintain — the product line built on the foundation of Cityworks, the GIS-centric asset management platform Trimble has been integrating since 2020 and has since folded together with AgileAssets, Pavement Express, and Trimble Unity Work Management. During a demonstration for a city agency, a GIS manager watching the session realized what that actually meant. “All of a sudden he went, wait a minute, are you telling me I don’t have to manage two GIS databases?” Sprague recounted. Under the arrangement the manager was used to, an agency’s authoritative GIS data and its separate maintenance system had to be kept manually in sync — an update in one system requiring a notification to update the other. “He stood up, he goes, we should buy it,” Sprague said. “I think they in particular had three GIS databases that they had to kind of keep.”
Unity Maintain’s core function is scheduling and tracking the recurring operations and maintenance work that consumes most of an asset’s life — testing fire hydrants on a set interval, managing maintenance routes and work order packages, or running a dedicated pavement management and predictive-condition module that Sprague said pairs directly with Trimble’s own road-scanning hardware, the MX60 and MX90 mobile mapping systems, to shift agencies from reactive to predictive maintenance budgeting.
Permit, and the Plumbing Underneath It
Sitting alongside Unity Maintain is Unity Permit, which Sprague described as originally built for permitting, licensing, and land management but flexible enough to function as a general workflow generator — capable of running something as narrow as a municipal dog-licensing process or as involved as a full building permit review, and including public-facing interfaces for citizen requests such as graffiti reports. This year, Trimble is co-marketing Unity Permit directly with Esri, and has also partnered with Evolve, a plan review solution, to automate the workflow that follows a permit submission — moving agencies away from paper-based or manually routed plan review and toward an automated process connected to Unity Permit.
Where the Curve Bends
Unity Construct addresses the earlier end of the same asset lifecycle: planning, design, and the execution of construction itself. Sprague pointed to a specific dataset to explain why Trimble treats this phase as disproportionately important — a Woolpert analysis showing that organizations using an end-to-end asset lifecycle management approach can reduce an asset’s total cost of ownership by as much as 40 percent, a figure that also anchors Trimble’s own published guidance on the subject. The underlying logic, laid out in that material, is that owners retain the greatest influence over an asset’s lifetime cost during planning, design, and early construction, and that influence collapses once the asset reaches a certain point of completion — the same published analysis puts roughly 75 percent of an asset’s total cost inside the operation and maintenance phase specifically.
Sprague illustrated the stakes with two examples. A submittal review that isn’t caught correctly — the wrong grade of concrete specified for a component — can mean a bridge pylon needs replacement five years early, an expensive failure traceable to a single missed review step. Trimble’s recent acquisition of Document Crunch, an Atlanta-based AI document analysis company already used on more than 10,000 construction projects, is aimed at exactly that failure point: comparing design documents against submittal packages directly and flagging what’s been missed, rather than relying on a human reviewer to catch a discrepancy buried in a large submittal. The second example was simpler and, by Sprague’s account, more common: warranty documentation that goes missing between the team that designed a building system and the team that eventually operates it. “An HVAC unit stops working, they don’t know what the warranty, so they get a new one,” he said. “Depending on the size of it, $50,000, $100,000 — it could have been replaced for free.”
Trimble Connect as the Handoff
The mechanism tying all of this together, in Sprague’s telling, is Trimble Connect — not yet built natively into every Unity product, but positioned as the collaboration layer that will eventually run through all of them. Its more immediate role is bridging reality capture and design: point cloud and lidar data captured on site is brought into Connect, overlaid on the design model, and used by architects and engineers to mark up and update that model, producing an as-is state that then informs the next phase of planning or maintenance. “The field survey data is informing the design process,” as Sprague put it, describing something closer to a loop than a one-way handoff.
Sprague drew a direct line from that scanning and location data to Esri’s role in the stack, citing a framing Trimble used in its own Tuesday session at the conference: Esri provides spatial awareness, and Trimble provides the precision of the location itself — and neither, on its own, is sufficient. He illustrated it with a maintenance scenario from his own region: a late-winter storm floods a highway when storm drains clog, but the drains themselves are underwater and impossible to locate visually. If their positions had been precisely mapped in advance, a crew could go directly to them rather than searching. The same logic extends underground, where utility networks buried at high accuracy let a crew avoid digging in the wrong place entirely. “Having the precision location without being able to put it onto a map and have spatial analysis and spatial awareness is an incomplete data set,” Sprague said. “That’s why we work so well handed.”
The Human Drone
The layer underneath all of that — how the field data that feeds Trimble Connect actually gets captured — comes down to reality capture itself. Bryer Palmer, a Trimble product manager for SiteVision, walked through the phone-and-tablet-based reality capture workflow that SiteVision’s 3D scan tool enables: pairing the lidar sensor on a supported iPhone or iPad with Trimble’s DA2 GNSS receiver to capture a georeferenced, centimeter-accurate point cloud directly in the field, without a drone or a dedicated laser scanner.
Asked where that phone-based approach holds up on its own, and where it hits a wall, Palmer was direct about the limiting factor. “We’re still working with GNSS,” he said. “Our algorithm for collecting lidar has been advanced and iterated so many times that we’re starting to get close to taking away a lot of the error that’s produced by the lidar offset from the sensor itself and the camera. So again, it starts to be, what can you do with GNSS.” Concrete floor flatness, overhead mechanical, electrical, and plumbing installations, telephone lines, overhead poles, and very deep trenches all sit outside the tool’s comfortable range — anywhere the tolerance drops to an inch or less, Palmer said, is a place he wouldn’t take the chance with SiteVision. Indoors, the tool can tie into a placed QR code for high local accuracy, but that accuracy holds only relative to the code itself; once a user moves away from it, positional certainty degrades in the way it wouldn’t with a continuously tracked GNSS position outdoors.
Palmer’s own customers have taken to calling the workflow “the human drone” — an operator walking a site can go places a drone can’t, including into dense brush and low canopy that defeats aerial line of sight, and can capture a stockpile or trench condition that changes day to day, faster than any aircraft that isn’t flying every single day. “A lot of the feedback we’ve received is that convenience, the portability, the speed, is really a lot more important than dialing in on millimeters,” Palmer said. “While this can’t handle the same applications that our mobile mapping systems can, nor could an X9, it’s pretty much applicable for 90 to 95 percent of job applications, depending on what you’re using it for.” He was careful to frame the tool as a complement to drones and full lidar rigs rather than a replacement for either — useful precisely in the conditions, like dense foliage or a rapidly changing stockpile, where a drone’s own physics work against it.
Feeding the Field What the Office Already Knows
The second half of the SiteVision conversation moved from capture to context — specifically, a Trimble Connect integration bringing third-party utility data directly into the augmented-reality view a field crew sees on site. Trimble has built a set of web feature service integrations, still largely in prototype, pulling in outside GIS data through a growing list of partners. On the SiteVision side specifically, one confirmed partner is 4M Analytics, an Austin, Texas-based company that compiles subsurface utility records from across the United States into a single AI-processed map. “We’ve actually got a lot of those running, or prototyping, in the background right now,” Palmer said, “and we’re really excited about a lot of the data sets we could acquire just by having these different types of integrations.”
The integration, defined by a project boundary set inside Trimble Connect, pulls in all utility information 4M has associated with that location and grants an API key that renders it in both 2D and 3D through SiteVision. Palmer described a live demonstration of the workflow, run for a group of utility contractors in New York and Atlanta, that went well beyond expectations. “Our first demonstration went so well that we got more or less a standing ovation,” he said. “Since then, we’ve had interest from other parties in Trimble — working with civil side people, the building infrastructure side, our excavation business.” The practical case, in Palmer’s framing, is straightforward: a crew walking up to a street with no BIM design and no existing utility markup can plug into the integration and see utilities appear in their SiteVision view immediately, giving excavators an accurate starting point instead of guessing or waiting on a locate request. “It could save you hours, days, weeks,” Palmer said, “and hundreds of thousands of dollars in missing utility strikes.”
Palmer was careful to frame the data as guidance rather than a substitute for verification — “we’re only recommending this as a form of guidance, not total liability” — and pointed to the underlying openness of most utility records as the reason the integration was tractable at all: point and line data of the kind utility owners typically collect isn’t proprietary in the way a purpose-built software platform’s data model might be, which leaves room for exactly this kind of cross-company aggregation. Asked about the broader business logic of sharing data this openly, Palmer didn’t treat it as altruism. “You have to own the entire space to make it an actually observable, valuable play, which nobody does,” he said. “Right. And even as big as Trimble is, we’re not even close.”
What Comes Next
Both conversations pointed toward the same near-term venue: Trimble Dimensions, the company’s annual user conference, returning to The Venetian in Las Vegas November 9 through 11, with attendance expected to exceed 8,000 professionals across construction, geospatial, and land survey. Sprague pointed to Dimensions as where the promised unified AI message across Trimble’s product lines is likely to take shape. Between now and then, the through-line across both interviews is less about any single feature than about where Trimble is placing its bet: that the value in construction and infrastructure technology increasingly sits not inside any one tool, but in the connective layer between GIS context, precise field capture, and the software that manages an asset for the decades after it’s built — and that the AI riding on top of that connective layer has to earn trust by being boring and repeatable, not by being impressive.