Cities, counties, towns, and special districts across California and across the United States rely on StreetSaver as the pavement management system of record for their local street and road networks. Developed and maintained by the Metropolitan Transportation Commission (MTC) of the San Francisco Bay Area, StreetSaver is purpose-built for the scale, budget realities, and reporting requirements of municipal and county agencies the city engineers, public works directors, and pavement managers who are responsible for keeping local streets functional, safe, and fiscally sustainable on budgets that are never large enough for the task.
But one question is coming up consistently across StreetSaver agencies: how do we get current, complete, and consistent PCI data into StreetSaver more frequently without the cost and scheduling burden of traditional pavement surveys?
This guide is written specifically for city engineers, pavement managers, public works directors, and GIS analysts who are already working within the StreetSaver ecosystem and are evaluating how AI-powered road condition assessment fits into not on top of their existing pavement management workflow.
Before covering where AI fits, it is worth being precise about what StreetSaver actually provides for municipal pavement management because this determines exactly where an AI condition data layer connects and adds value.
StreetSaver is a web-based pavement management system (PMS) developed by MTC and used by over 100 agencies across the United States, with particular concentration in California where Measure B, SB 1, and other state and regional funding programmes require structured pavement condition reporting. Unlike general-purpose asset management platforms, StreetSaver is built entirely around the pavement management workflow: collecting PCI data, analysing network condition, modelling deterioration, generating maintenance and rehabilitation recommendations, and producing the funding justification reports that city councils and boards of supervisors require.
The analytical foundation of StreetSaver is the Pavement Condition Index (PCI) the ASTM D6433-based 0–100 condition scale that quantifies pavement surface quality from failed (0–10) to excellent (86–100). StreetSaver stores PCI scores at the pavement section level, tracks condition history over multiple survey cycles, projects deterioration under defined maintenance scenarios, and generates maintenance and rehabilitation recommendations based on the condition-cost curve logic that MTC has refined over decades of pavement management practice. PCI is not just a number in StreetSaver it is the input that drives every prioritisation decision, budget analysis, and council report the platform produces.
StreetSaver’s M&R (Maintenance and Rehabilitation) recommendation engine assigns treatment recommendations to each pavement section based on its current PCI, surface type, and the agency’s defined treatment decision tree. Crack sealing, slurry seal, cape seal, overlay, reconstruction each treatment option has a cost, a PCI range of applicability, and a projected post-treatment condition outcome. The recommendation engine identifies the right treatment for each section at the right time in its deterioration cycle, enabling agencies to maximise the pavement life gained per dollar of maintenance investment.
StreetSaver’s budget analysis module models the network-level condition outcome under different annual maintenance spending levels. What is the average network PCI in five years if the city spends $2M per year on pavement? What if it spends $3M? What is the minimum annual investment required to maintain the current average PCI? These scenario analyses are the primary output that pavement managers use to develop Capital Improvement Programme (CIP) budget requests and communicate road infrastructure funding needs to city councils and finance departments in terms that non-engineers can evaluate and act on.
The needs analysis workflow in StreetSaver identifies the pavement sections across the network that are candidates for each treatment type in the current budget cycle, estimates the cost of treating all sections at their recommended treatment, and compares that total need against available funding. The resulting prioritised project list ranked by cost-effectiveness, PCI, or agency-defined priority criteria is what drives the annual paving programme and the CIP project schedule.
For California agencies, StreetSaver is the platform most widely used to produce pavement condition reports required for SB 1 (Road Repair and Accountability Act) reporting to the California State Controller’s Office, Measure B and other regional measure compliance reports required by Metropolitan Planning Organizations (MPOs), and CDBG and HSIP funding applications that require structured pavement condition documentation. Compliance with these reporting requirements is directly linked to an agency’s eligibility for state and regional road funding making StreetSaver’s reporting output a financial necessity, not just an operational convenience.
StreetSaver integrates with GIS platforms to display pavement condition data, maintenance history, and M&R recommendations spatially. The StreetSaver GIS viewer enables city engineers and pavement managers to view PCI heatmaps, identify geographic clusters of poor-condition streets, map planned maintenance projects, and generate the visual condition reports that council presentations and public communications require. Condition maps generated from StreetSaver PCI data are standard deliverables in city pavement management annual reports and budget presentations.
StreetSaver tracks maintenance activities applied to each pavement section treatment type, application date, pre-treatment PCI, post-treatment PCI, and cost building a treatment history that informs deterioration model calibration and treatment effectiveness analysis. Understanding how long a slurry seal or overlay is actually performing on specific street types in the local climate is essential for making StreetSaver’s deterioration projections and M&R recommendations accurate for the agency’s specific network conditions.
StreetSaver manages pavement condition data, generates maintenance recommendations, and produces funding justifications with a precision and workflow fit that is specifically calibrated for municipal pavement management. What it does not do is generate the PCI data that all of those functions depend on.
The PCI scores, condition histories, and distress records in StreetSaver are only as current as the last time the agency conducted a formal pavement condition survey and entered the results. In practice, for most cities and counties, that means:
The result: a pavement management platform whose M&R recommendations, budget analyses, and funding reports are only as reliable as the PCI data feeding them and that data is, for most agencies, several years old, collected with variable consistency, and missing the early-stage distress signals that cost-effective preventive maintenance depends on.
This is not a gap in StreetSaver's design. It is a structural limitation of how municipal pavement condition data has historically been collected. StreetSaver can only manage what it receives.
AI-based computer vision for road surveys addresses the PCI data generation problem directly and does so at a cost and operational simplicity that changes the economics of pavement condition and monitoring surveying for municipal agencies. The approach is straightforward: a vehicle-mounted dashcam captures geo-tagged, time-stamped video of the street network as the vehicle drives at normal operating speeds. That footage is processed through AI models that detect and classify pavement distress automatically, returning structured condition results including ASTM D6433-aligned PCI scores that load directly into StreetSaver.
No specialist survey vehicle. No contracted survey crew. No per-lane-mile survey cost at rates that strain pavement programme budgets. Any city or county vehicle already driving the street network a public works maintenance truck, a street inspector’s vehicle, a parks department van, a code enforcement car becomes a pavement survey platform. For municipalities where the per-lane-mile cost of traditional surveys is the binding constraint on survey frequency, this changes what is operationally and financially achievable.
RoadVision AI’s models are trained on over 100 million road images from networks across South Asia, the Middle East, Southeast Asia, Europe, and Africa, covering diverse pavement types, climate conditions, road classifications, and surface configurations. For California municipal agencies, this training breadth covers the specific surface types and distress patterns common in California’s varied climate zones: reflective cracking through slurry and cape seal surfaces, thermal cracking in inland valley climates, alligator cracking on residential collector streets, rutting on arterials with high bus and truck loading, and ravelling on aged chip seal in rural county conditions. This specificity of training coverage is what determines whether AI condition scores are accurate enough to replace contracted survey data as the PCI input to StreetSaver.
A single dashcam survey pass, processed through the RoadVision AI pipeline, returns pavement condition intelligence across the following categories all structured for direct import into StreetSaver:
Detection and classification of all ASTM D6433 distress types relevant to asphalt pavement: alligator cracking, bleeding, block cracking, bumps and sags, corrugation, depression, edge cracking, joint reflection cracking, lane and shoulder drop-off, longitudinal and transverse cracking, patching and utility cut patching, polished aggregate, potholes, railroad crossing distress, rutting, shoving, slippage cracking, swell, and weathering/ravelling. Each distress type is detected with severity level (Low, Medium, High) and quantity estimate, which are combined to produce a per-section ASTM D6433-aligned PCI score on the standard 0–100 scale. IRI-equivalent roughness is also returned per 100-metre segment, compatible with StreetSaver’s pavement condition fields.
AI distress detections are aggregated at the pavement section level as defined in the agency’s StreetSaver database not just at the raw detection point level. This means each survey output maps to a StreetSaver pavement section record with a PCI score, individual distress type quantities, and survey date, ready for direct import without requiring the agency to aggregate point detections into section-level condition records manually.
AI survey processing identifies the surface type of each pavement section dense-graded asphalt, open-graded asphalt, slurry seal, cape seal, chip seal, concrete, or composite which is relevant to StreetSaver’s M&R decision trees, which assign different treatment options based on surface type as well as PCI. Surface type identification also supports the detection of areas where a previous slurry or cape seal has masked underlying distress that will re-emerge prematurely a condition common in California municipal networks that standard windshield surveys frequently miss.
Automated detection and classification of roadway asset inventory including signs (regulatory, warning, informational), pavement markings (centreline, edge line, crosswalks, stop bars, bike lane markings, school zone markings), kerb and gutter condition, ADA ramp condition, drainage inlets, and utility cover condition. Every asset detection includes GPS coordinates, asset type, condition grade, and photographic evidence providing an updated asset inventory alongside the pavement condition survey in a single pass.
Detection of safety deficiencies including faded or missing pavement markings, damaged or missing signage, ADA accessibility obstructions, sight-line obstructions at intersections, and encroachments onto the travelled way. Safety findings are geo-tagged and severity-scored, providing documentation for local road safety plan development and HSIP funding applications.
Every detection record includes: distress type, severity, quantity, PCI contribution, GPS coordinates, pavement section ID, survey date, and a geo-tagged photographic evidence frame. Output files are structured for direct import into StreetSaver’s pavement condition dataset.
StreetSaver manages pavement condition data through structured section records, PCI rating histories, distress type inputs, and M&R recommendation workflows all built around the ASTM D6433 methodology. RoadVision AI outputs are structured to align precisely with StreetSaver’s data schema and import requirements, enabling AI-generated PCI data to enter the pavement management workflow without manual data transformation, distress re-aggregation, or GIS processing.
This is the foundational integration point. RoadVision AI delivers per-section PCI scores, individual distress type quantities by severity level, and survey dates in the format required by StreetSaver’s condition data import. The full ASTM D6433 distress catalogue is supported not a subset meaning the PCI calculation methodology used by the AI pipeline matches the methodology that StreetSaver applies to manually collected distress data. AI-derived PCI scores are directly comparable to historic manually-rated PCI scores in the same section record, preserving the continuity of the condition history that StreetSaver’s deterioration curve calibration depends on.
AI survey results are delivered with pavement section IDs and from/to street references that map to existing section records in the agency’s StreetSaver database. Each AI condition output loads against the correct StreetSaver section record without requiring manual spatial matching, GIS overlay processing, or section boundary re-definition. For agencies with large networks, this automated section mapping eliminates the most time-consuming step of the traditional survey data entry process.
StreetSaver’s deterioration models project future PCI for each pavement section based on observed condition trajectories and treatment history. When PCI observations are available annually or biennially rather than every three to five years, deterioration curves are calibrated on denser time-series data producing more accurate projections of when each section will reach treatment thresholds. More accurate deterioration curves produce more reliable M&R recommendations and more credible budget analysis outputs. The MTC-developed condition-cost models in StreetSaver become more effective as the quality and frequency of their PCI inputs improve.
AI survey outputs are formatted to supply the pavement condition data required for California SB 1 annual reporting to the State Controller’s Office, regional measure compliance reports for Bay Area Measure B and equivalent MPO programmes across California, and CDBG and HSIP grant applications requiring structured condition documentation. Annual AI surveys provide the annually updated PCI data that these reporting cycles require in StreetSaver’s format, without requiring annual contracted survey campaigns.
AI survey results are available as Shapefile and GeoJSON exports compatible with StreetSaver’s GIS integration and Esri ArcGIS. PCI scores, distress flags, and M&R recommendation inputs load into the GIS layer as spatially referenced pavement section features enabling PCI heatmaps, worst-streets lists, and planned project maps to be generated immediately following each survey cycle within the StreetSaver GIS environment that city engineers use for council presentations and public communications.
Once loaded, AI PCI data participates in every downstream StreetSaver workflow the agency already runs: M&R recommendation generation, needs analysis, budget scenario modelling, CIP project list development, SB 1 and MPO compliance reporting, GIS condition mapping, and treatment effectiveness tracking. Pavement managers work in the same StreetSaver environment they have always used the difference is that PCI data is current, complete, and consistently measured every survey cycle.
What does not change: The StreetSaver platform. The PCI methodology. The M&R decision trees. The budget analysis workflows. The CIP development process. The SB 1 and MPO reporting structure. The GIS viewer. The council presentation format.
What changes: The frequency, cost, coverage, and consistency of the PCI data feeding into those workflows and through them, the accuracy and credibility of every M&R recommendation, budget analysis, and funding report StreetSaver produces.
Survey frequency increases and survey cost decreases simultaneously. Traditional contracted pavement surveys cost tens of thousands of dollars per campaign and are conducted every three to five years. AI dashcam surveys conducted using existing city vehicles cost a fraction of contracted survey rates per lane mile and can be repeated annually or more frequently. For the first time, the economics of annual full-network PCI surveys are achievable for the average-sized municipal agency.
Coverage becomes genuinely complete every cycle. Traditional survey campaigns prioritise arterials and collector streets because full residential network coverage at survey contract rates is prohibitively expensive for most cities. AI processing of city vehicle dashcam footage covers every lane mile in the network on every survey pass including cul-de-sacs, alleys, parking lot access drives, and the residential grid that StreetSaver’s database contains but that has not received a formal PCI survey in years.
PCI ratings become consistent across cycles. Because the same AI models apply the same ASTM D6433 distress detection and severity classification criteria on every survey run, PCI scores are directly comparable between survey years, between different sections of the network, and between different survey operators. The year-on-year PCI series in StreetSaver becomes a genuinely consistent time-series rather than a patchwork of assessments from different contracted crews using different rating practices.
SB 1 and regional measure reporting becomes fully compliant. With annual AI surveys providing current PCI data for the full network, California agencies can submit SB 1 annual reports and regional measure compliance reports based on current measured condition not interpolated or carried-forward values. The credibility of the agency’s condition data in the context of state and regional funding allocation improves directly.
M&R recommendations reach sections in the preservation window. StreetSaver’s M&R logic is designed to catch streets in the PCI 55–75 range where preventive treatments slurry seal, cape seal, crack seal are cost-effective. Annual AI surveys detect early-stage distress that keeps PCI records current through this preservation window, ensuring that StreetSaver’s treatment recommendations trigger while low-cost options are still viable. For agencies managing deferred maintenance backlogs, this is where AI-driven road survey frequency most directly reduces long-run capital costs.
Council presentations gain current, credible condition evidence. PCI heatmaps, worst-streets lists, and before-and-after treatment performance maps generated from annually updated AI data carry more credibility in council chambers than maps based on three-year-old survey data. When a council member asks why a particular street is not on the paving list, the answer is grounded in a PCI score from this survey cycle not an extrapolation from a survey conducted before the previous election.
The evaluation path is designed to be practical for city engineers and pavement managers working with limited IT and GIS support:
For cities and counties using StreetSaver, the practical question is straightforward: can we get current, complete, and consistent PCI data into StreetSaver more frequently and at lower cost than our current contracted survey approach, without compromising the data quality that our M&R recommendations, budget analysis, and SB 1 reporting depend on? That is the specific question this integration is designed to answer.
Any GPS-enabled dashcam producing standard MP4 or MOV video at 1080p or above. No proprietary hardware, no calibration rig, and no vehicle modifications beyond mounting the camera. A consumer-grade GPS dashcam costing under $200 mounted on the windshield of any city vehicle is sufficient to capture survey-quality footage.
No. Any city or county vehicle already operating on the street network serves as a survey platform. A public works maintenance truck, a street inspector’s vehicle, a parks department van, or a city fleet vehicle driving routine routes can capture pavement condition footage as part of normal operations. No additional survey staff time, no specialist vehicle equipment, and no separate survey scheduling is required.
AI-derived PCI scores are calculated using the full ASTM D6433 distress detection and severity classification methodology the same standard that StreetSaver uses to calculate PCI from manually collected distress data. For agencies transitioning from contracted windshield or automated surveys to AI dashcam surveys, a calibration run over a set of reference sections with known PCI values allows the RoadVision AI team to validate and, where needed, calibrate AI outputs against the agency’s existing StreetSaver baseline. This ensures continuity of the PCI history that StreetSaver’s deterioration model calibration depends on.
Yes. AI survey outputs are formatted to supply the pavement condition data that StreetSaver compiles for SB 1 annual reports to the California State Controller’s Office — including section-level PCI, surface type, survey date, and network-level condition summaries. For agencies currently submitting interpolated or carried-forward PCI data because annual full-network manual surveys are not feasible, AI surveys provide the annually updated, full-network PCI data that SB 1 reporting requires without the cost of annual contracted survey campaigns.
Yes. MTC Measure B and equivalent Bay Area regional measure compliance frameworks require structured pavement condition reporting by member agencies. AI PCI data is formatted for the condition reporting inputs that StreetSaver compiles for MTC and MPO submissions, including average network PCI, percentage of network in good, fair, and poor condition, and year-over-year condition change metrics.
Yes. The full ASTM D6433 distress catalogue for asphalt pavement is supported, including all distress types relevant to California surface treatments: reflective cracking through slurry and cape seal, alligator cracking, longitudinal and transverse cracking, rutting, ravelling, and patching. For concrete pavement sections in the StreetSaver database, AI detection covers the relevant ASTM D6433 concrete distress types as well.
The AI models operate at over 97% detection accuracy across core pavement distress categories on standard road network footage. On validation tests against concurrent contracted manual PCI surveys on California municipal networks, AI-derived PCI scores fall within five PCI points of manual ratings in the large majority of sections. Complete accuracy benchmarks by distress type, surface category, and lighting condition are available in the technical documentation provided at onboarding.
The Autonomous Road Engineer
AI that is transforming road infrastructure, from reactive to predictive.
.webp)
.webp)
.png)