Winston Francois helps robotaxi, ADAS, and freight autonomy companies build reporting infrastructure that turns raw telemetry into clear decisions – for your board, your ops team, and your regulators.
The Problem
Sensor Data Volume Without Business Clarity
AV systems generate terabytes of sensor logs, LiDAR captures, and edge-case recordings every day. Most analytics teams can describe what the vehicle saw. Far fewer can tell the CEO which safety metric is lagging, which fleet segment is unprofitable, or which regulatory threshold is at risk. The gap between raw data and decision-ready insight is where AV companies lose time and investor confidence. You need a reporting layer that speaks to operators and executives, not just ML engineers.
Regulatory Reporting That Drains Engineering
NHTSA Standing General Orders, CARB AV testing reports, and state-level permit disclosures each require structured data in specific formats on strict timelines. Most AV companies handle this with spreadsheet exports built by engineers who should be working on the vehicle stack. Every compliance report that takes a week to produce is a week of engineering capacity burned on data wrangling. A dedicated reporting architecture makes this a scheduled export, not a quarterly scramble.
Investor Dashboards Built for the Wrong Audience
Series A and B investors in the AV space want to see miles driven, disengagement rates, geofence expansion, and fleet utilization trends – not raw log counts. When your investor update requires a data analyst to manually pull from five systems and reconcile timestamps, you are one bad month away from a credibility problem. The narrative your data tells should be consistent, reproducible, and fast to generate. That requires infrastructure, not ad hoc queries.
No Single Source of Truth Across Simulation and Physical Fleet
AV development cycles run parallel tracks: simulation environments, closed-course testing, and public road deployments. Each generates its own data schemas, stored in different systems, often managed by different teams. When leadership asks a simple question – 'How did our behavior in simulation compare to real-world performance last quarter?' – the answer takes days to construct. Without a unified data model, every cross-environment analysis is a one-off project that never compounds into organizational knowledge.
How We Help
Winston Francois begins every engagement with a data audit. We map your existing telemetry pipelines, business intelligence tools, and reporting workflows to identify where data is produced, where it is consumed, and where it breaks down between those two points. This is not a theoretical exercise. We interview your data engineers, your ops leads, and whoever currently owns the investor update to understand what questions are being asked and how long they take to answer.
From that audit, we design a reporting architecture that matches your operational reality. For most AV companies at Series A through growth stage, this means a centralized data warehouse or lakehouse structure that ingests from simulation platforms, physical fleet telemetry, and third-party sources like permit databases and weather APIs. We define the schema conventions and event taxonomy so that data from different vehicle generations and test environments can be compared without manual reconciliation.
On the strategy layer, we establish a metrics framework built around the questions that matter to your three main audiences: operators who need daily fleet performance views, executives who need weekly trend summaries, and investors and regulators who need periodic structured disclosures. Each audience gets a purpose-built reporting surface – not a single dashboard everyone ignores because it was designed for nobody in particular.
Execution includes building or rebuilding the core dashboards, automating your regulatory report exports, and wiring alerting logic so that safety-relevant thresholds surface immediately rather than appearing in a weekly review. We document every data definition in a shared glossary so that 'disengagement' means the same thing in your board deck as it does in your NHTSA filing.
Measurement is built in from day one. We define SLAs for report freshness, alert on pipeline failures, and run quarterly reviews to confirm that the metrics your dashboards surface are still the metrics your business decisions actually require. Analytics infrastructure that is not actively maintained decays. We build the maintenance cadence into the engagement from the start.
What we deliver
- Data and reporting infrastructure audit with gap analysis
- Metrics framework mapped to operator, executive, and investor audiences
- Core dashboard builds for fleet operations and safety KPIs
- Automated regulatory report exports (NHTSA SGO, state permit filings)
- Unified data model documentation and shared metrics glossary
- Alerting configuration for safety-relevant threshold breaches
- Quarterly analytics review and metrics refresh process
AV companies that treat safety telemetry and business analytics as the same problem end up with dashboards that neither their engineers nor their investors trust. The most operationally mature AV teams maintain separate but connected reporting layers – one for the vehicle stack, one for the business – with a clear data contract between them.
Our Methodology
Winston Francois runs analytics engagements in 90-day sprints. The first 30 days are diagnostic: we audit your data estate, interview stakeholders across eng, ops, and finance, and deliver a prioritized gap analysis with a concrete architecture recommendation. No generic frameworks – this is scoped to your stack, your team size, and your current reporting obligations.
Days 31 through 60 are build and integration. We instrument the highest-priority reporting gaps, establish the data model conventions, and deliver the first working dashboards with real data. By end of month two, your ops team has a daily fleet view and your executive team has a weekly summary that does not require a data analyst to compile.
The final 30 days focus on automation, documentation, and handoff. Regulatory exports are scheduled and tested. Alert logic is tuned against historical thresholds. Every data definition is documented. Your internal team owns the system at the end of the engagement, not us. We do not build reporting infrastructure that creates dependency on our continued presence.
How We Work
We work with AV companies at Series A through growth stage where data complexity has outpaced internal reporting capacity. Engagements typically run 90 days with a defined scope, though we offer ongoing retainers for companies that want a fractional analytics strategy partner beyond the initial build.
Days 1 through 30 are discovery and architecture. We access your existing systems, conduct stakeholder interviews, and produce an infrastructure design with data model conventions and a prioritized implementation plan.
Days 31 through 60 are build and integration. Core dashboards go live with real data. Regulatory exports are wired and tested in staging. Your team reviews and signs off on metric definitions before anything ships to a board or regulatory body.
Days 61 through 90 are hardening and handoff. Alert configurations are tuned. Documentation is complete. We run a structured knowledge transfer with your internal team so that ongoing maintenance does not require re-engaging us. Ongoing retainer options are available for companies that want quarterly audits and metrics refresh cycles.
- Investor-ready fleet performance and safety dashboards produced on a scheduled cadence without manual data pulls
- Regulatory report exports completed in hours rather than engineering-days
- A unified metrics glossary that aligns definitions across engineering, operations, and finance
- Alert infrastructure that surfaces safety threshold breaches in real time rather than in post-hoc reviews
If your autonomous vehicles company needs data, reporting & analytics leadership, we should talk.
Expand your marketing team output with our experts
Let us take a custom approach to your growth goals by assembling and leading the best-in-class marketing team to support your next stage.
Frequently asked questions
What data sources do you typically integrate for AV analytics?
The most common sources are vehicle telemetry systems, simulation platform outputs, fleet management platforms, permit and regulatory databases, and financial systems for unit economics tracking. We also frequently integrate weather and mapping APIs when safety analysis needs environmental context. The specific sources depend on your stack – we start with an audit rather than assuming a standard set applies to your company.
How do you handle the scale of sensor data AV companies produce?
We distinguish between data that needs to be stored in full fidelity for safety audits and ML training, and data that needs to be aggregated for business reporting. Most business reporting questions do not require raw sensor logs – they require well-defined event summaries and computed metrics. We design your reporting layer to operate on aggregated event streams rather than trying to run BI queries against petabyte-scale raw storage. This keeps reporting fast and cost-efficient without compromising your ability to drill into raw data when needed.
Can you help us meet NHTSA Standing General Order reporting requirements?
Yes. SGO compliance reporting requires structured incident data in specific formats on defined schedules. We build the data pipelines and export automation to make SGO filings a scheduled process rather than a manual quarterly project. We also document the data lineage so you can demonstrate to regulators exactly how each reported figure was computed. We are not legal counsel and do not interpret regulatory requirements – we build the technical infrastructure to fulfill requirements your legal team has already defined.
How long does it take to get working dashboards from the start of an engagement?
Initial dashboards with real data typically go live within 45 to 60 days, depending on the complexity of your existing data infrastructure and how many source systems we are integrating. The first 30 days are diagnostic and architecture, which is necessary to avoid building dashboards on a broken data model. Companies that try to skip the diagnostic phase typically end up rebuilding dashboards when they discover the underlying data has definition or quality issues.
Do you work with companies that already have a data team?
Yes, and most of our engagements involve an existing data team. We typically operate as a strategy and architecture layer that works alongside your internal data engineers rather than replacing them. The most common pattern is that the internal data team owns the vehicle and simulation data pipelines while we design the business reporting layer and define the data model conventions that bridge the two. We document everything so your team can maintain and extend the system without ongoing dependency on us.
What does an analytics engagement with Winston Francois cost?
A 90-day analytics engagement typically runs $15K-$35K depending on scope, number of data sources, and the complexity of regulatory reporting requirements. Ongoing retainers for quarterly audits and metrics refresh are scoped separately. We do not offer hourly billing – we scope to outcomes and charge a fixed project fee so you know the total cost before we start. We can discuss scope in a strategy call and provide a written proposal within a week.
