Pulse is the platform every Beacon product runs on. It connects to the systems and feeds your agency already has, checks whether the data can be trusted, and tells your people what's broken, where, why, and what to do about it. Every answer traces to its source and its cost.
Your ATMS runs the network. Pulse sits alongside it and reads across every system at once, so the problems no single system can see come with a cause and a next step. It works with the ATMS you have. It doesn't replace it.
Not "camera down," but "the IP changed after an encoder swap, so update the host mapping." The diagnosis and the fix, not just the alert. Camera Reliability is built on Pulse.
Misspelled or stale DMS messages, flagged before a driver or a commissioner reads them on the road. Every message can be replayed in the context of the incident it was posted for.
Pulse checks each feed for more than structure. It asks whether a work zone in the feed matches real conditions and your published standards, and says plainly when a source can't be trusted.
Every incident is matched to the cameras, signs and devices nearby, so operators see the whole scene instead of opening five systems to build it themselves.
State DOTs and transit agencies have spent decades building ProjectWise repositories, SharePoint document libraries, and public data feeds. Beacon doesn't replace any of that. We read from your existing systems, index the content, and make it answerable.
The document stays where it's always been. The SOP stays in OneDrive. The concept report stays in ProjectWise. Beacon adds a knowledge layer on top, one that keeps in sync as your documents change and answers questions from a single interface instead of five different logins.
The first question agencies ask us is whether this requires a rip-and-replace of their existing infrastructure. The answer is no. We have never asked a DOT to move their documents to get started.
Trust, Throughput, Tokens, and Transparency. Every product Beacon builds is held to all four.
Citations are the floor, not the ceiling. Every response links to its source passage, but trust is built through continuous evaluation. 520 golden Q&A pairs run weekly against every active deployment, scoring retrieval, faithfulness, and citation accuracy against defined thresholds. When something regresses, Beacon staff sees it before operators do.
State DOT and transit agency document environments are large. One state DOT's ProjectWise environment alone contains 10.5 million documents. Beacon is built to index at that scale, across multiple data sources simultaneously, and stay current as documents change. When a file is edited, only the changed section is re-embedded. The rest of the index is untouched.
When Uber's CTO admitted he'd burned his entire 2026 AI budget in a few months, the underlying problem was invisible spend. Token spend scales with usage, and usage grows faster than most teams expect. For any organization with a budget owner to answer to, that's not a minor issue. It's a real one.
Beacon is built model-agnostic by design. The knowledge layer (your index, your citations, your evaluation history) lives independently of whichever model is doing the inference. Swap Anthropic for OpenAI or vice versa without rebuilding anything. That's not just a technical convenience; it's leverage. You can credibly take your business elsewhere, which keeps providers accountable to your terms.
State DOTs and transit agencies hold decades of irreplaceable operational knowledge: how to price a complex estimate, which responder to call on which corridor, what a pattern of incidents actually means on that stretch of road. Every API call to a frontier model without proper data controls is, at some level, a transfer of that institutional knowledge to a third party whose incentives don't align with yours.
AI model providers have a structural incentive to absorb as much intelligence from their customers as possible. It sharpens their models, which they then lease to your competitors. The question agencies should be asking isn't just "is our data private?" It's "who owns the intelligence our operational data contains, and who does it compound for?"
Beacon's architecture is designed so the answer is always: you. The retrieval index lives within a perimeter your IT team controls. The model sees only the retrieved passage. Your operational knowledge stays yours, and compounds for your agency rather than for a model provider.
Beacon runs a structured evaluation suite on every deployment, every week. 520 golden Q&A pairs, questions with known correct answers, run through the full pipeline and score against six RAG metrics. When something regresses, we know before you do.
The evaluation layer covers the failure modes that actually matter in government AI: hallucinated facts, broken citations, stale sources, and answers that are technically retrieved but not actually relevant. Each metric has a threshold. Drops below the threshold trigger immediate review.
Operators see a summarized trust score. Beacon staff sees the full metric breakdown. When we find a regression, we fix it and tell you what changed, not at the next quarterly review but within the week.
Government AI procurement requires vendors to be honest about where they actually stand on compliance, not to oversell certifications they haven't earned. Here's where Beacon actually stands.
We've done the architecture work that makes compliance achievable: threat modeling, NIST SP 800-53 family mapping, a four-role authorization model with explicit no-escalation invariants, and an audit log with hash-chain verification. The foundation is real. The certifications take time.
For agencies with immediate procurement requirements, the path we've used is a public-data-only pilot that requires no DIT or state security review to start, giving agencies six months to work the compliance path in parallel while already seeing value.
Architecture mapped against NIST SP 800-53 control families. 24 threat models documented with attacker profiles, mitigations, and residual gaps.
Every operator interaction logged with tamper-evident hash chain verification and configurable retention to match agency records policies.
In progress for NCDOT engagement. GCP Assured Workloads architecture scoped. Working through state-specific security review processes.
On the roadmap. Not something we'll claim until the audits are done. If your engagement requires it, let's talk timeline. The architecture is built to support it.
State DOTs and transit agencies generate and manage an enormous amount of public transportation data: ATSPM signal data, incident feeds, crash reports, sensor readings, permit databases, camera systems, and more. But that data lives in siloed systems that don't talk to each other.
Pulse connects them. Ask a natural language question and Pulse queries the relevant sources simultaneously, returning a synthesized answer with confidence scoring and full source attribution. Every source it drew from is visible. Every gap in the data is disclosed.
The goal isn't to replace any system. It's to make the data those systems already contain accessible to the people who need it, without a data engineering project every time.
Every state agency procurement officer asking about AI right now has the same concern: how do we know what this will cost, and how do we know it won't run away? That's a reasonable question, and most vendors don't have a good answer.
Pulse was designed from the ground up to answer it. Every query logs its token cost. The quality score shows how good the answer actually was. The cost-vs-quality chart shows how different ways of asking the same question produce different cost and reliability tradeoffs.
This isn't just for finance. It's how Beacon helps agencies decide which queries are worth automating, which should stay human-in-the-loop, and where efficiency can be captured without sacrificing quality. The dashboard is accessible to your team at tokens.beacon.dev.
Pulse is also an MCP server. Any LLM-based tool your agency already uses can query the same aggregated data layer directly, without Beacon building a custom interface for each one.
The transportation AI market is splitting into two architectures: closed copilots and open knowledge layers. Closed copilots are purpose-built interfaces that control everything. Open knowledge layers expose clean data APIs that any tool can connect to.
Pulse is the open knowledge layer for transportation. Your TMC operators can use TMC-React. Your traffic engineers can use Benchmark. Any AI tool that supports MCP, including the Microsoft Copilot many agencies already pay for, can connect to the same underlying data, without Beacon building the interface.
That's not a feature. That's a different philosophy about how transportation agencies should own and access their data. The data should travel with the agency, not with the vendor.
Pulse exposes its aggregated public data layer as an MCP. The engineering data from ProjectWise (workbooks, concept reports, historical projects) is next. Each product Beacon builds adds another layer to that foundation.
SOPs, workbooks, historical projects, concept reports, and aggregated public data, all normalized, indexed, and accessible from any MCP-compatible tool. Built to compound: every new Beacon product adds another layer.
Claude, Microsoft Copilot, Gemini, your agency's internal tools: if it supports MCP, it can query Pulse. Your operators don't have to abandon what they know to get access to what they need.
Every query your staff runs generates signal: what was asked, what was retrieved, what was useful. That signal flows back into the knowledge layer, improving retrieval, surfacing gaps in your documentation, and sharpening the system for the next query. Usage → signal → better knowledge → better answers → more usage. That's the flywheel.
That flywheel only compounds value for the organization that owns the index. Run it through an AI model provider's closed system, and the intelligence accrues to them. Run it through Pulse, where your agency owns the index, and it accrues to you. Each product deepens the same foundation: TMC-React indexes your SOPs, Benchmark adds historical project data, and Camera Reliability adds the health of every device on the network. Each one makes the next more powerful.
Via a native MCP server, that compounding knowledge is accessible from Claude, Microsoft Copilot, or any other AI tool that supports MCP, without custom integration work for each one. The knowledge travels with you, not with the model provider.
No. ATMS is how a DOT acts on its network. Pulse is how it understands it. Pulse reads from your ATMS and every other system around it, finds what's broken, and tells you why. The ATMS stays the system your operators act in.
The sources your state already makes publicly available: signal data, incident feeds, crash databases, DMS message logs, responder unit activity, and traffic sensor networks. All public, with no data sharing agreements required to get started. Private agency data (ProjectWise, internal systems) can be added via the MCP integration layer.
Pulse discloses the source and timestamp of every data point it uses in an answer. If a source hasn't updated recently, the answer says so. Quality scores reflect data currency, not just answer coherence. We don't hide the freshness problem. We surface it.
Each AI query consumes a certain number of tokens, the compute units that LLMs charge for. Pulse shows exactly how many tokens each query type costs and what quality it produces. This lets agencies model costs before procurement and set guardrails before costs surprise anyone.
Yes, where your Microsoft 365 tenant allows MCP connections. Copilot is good with the documents a person already has open. It doesn't see live feeds, and it doesn't compute across thousands of projects. Pulse does. Copilot is the front door; Pulse is the engine. Microsoft is still rolling MCP support out to the government clouds, so we'll confirm what your tenant allows before you plan on it.
They're applications built on Pulse. TMC-React is purpose-built for TMC operators querying SOPs and contacts. Benchmark is for project managers doing task-level estimates. Camera Reliability diagnoses why cameras go dark. Pulse is the platform underneath all of them: it connects the data, checks it, governs which models touch it, and logs what every query costs.
It means your agency's data becomes accessible from any AI tool that supports the MCP standard, which includes Claude and a growing list of others, Microsoft Copilot among them. You don't need internal engineers to benefit from it. Beacon handles the integration layer; your staff queries through whatever interface they already use.