Hiya, TNS, and First Orion: The Three Labeling Algorithms Killing Your Pipeline
Hiya, TNS Call Guardian, and First Orion run the three spam-labeling algorithms on every US carrier. Here is how each one scores your calls.
Short answer
Three private companies — Hiya Inc., Transaction Network Services (TNS), and First Orion — run the three reputation-scoring algorithms that decide whether your outbound calls ring, get labeled, or get silently blocked in the United States. Hiya powers AT&T and Samsung's native dialer; TNS (via Call Guardian) powers Verizon and US Cellular; First Orion powers T-Mobile, Metro, and Boost (and invented the "Scam Likely" label). Each vendor maintains a separate reputation score and a separate business-registration portal — Hiya Connect, TNS Enterprise Branded Calling, and First Orion INFORM — and cleanliness with one does not transfer to the others. You need to be registered, monitored, and remediated with all three in parallel.
Three private companies effectively decide whether your outbound calls ring in the United States. They are Hiya Inc., Transaction Network Services (TNS), and First Orion. Each partners with a different major carrier, each runs a proprietary scoring algorithm, and each has its own registration pathway — which means "clean with one" does not mean "clean with the other two." If you are not actively managing your reputation with all three, you are managing it with none.
The short version
The three companies that label your calls are Hiya, TNS Call Guardian, and First Orion. Hiya powers AT&T's spam labeling, TNS powers Verizon's, and First Orion powers T-Mobile's (and owns the "Scam Likely" label). Each uses a separate reputation score, each requires separate registration, and cleanliness with one does not transfer to the others.
Who runs what
| Analytics vendor | Primary carrier partners | Signature label |
|---|---|---|
| Hiya Inc. | AT&T, Cricket, Samsung (native Android dialer) | "Spam Risk," "Fraud Risk," "Scam" |
| TNS Call Guardian | Verizon, US Cellular, Sprint legacy | "Potential Spam," "Potential Fraud" |
| First Orion | T-Mobile, Metro, Boost, some cable/MSO carriers | "Scam Likely," "Spam Likely" |
This table is not theoretical. Every outbound call you make in the US hits one or more of these three engines at the terminating carrier before it rings. Which one depends entirely on the carrier of the person you are calling.
How each algorithm scores a call
None of the three publishes the full scoring model — they treat it as proprietary — but based on their published whitepapers, patent filings, and annual transparency reports, the inputs overlap heavily. The differences are in weighting.
Hiya Inc.
Hiya leans heavily on consumer feedback loops. It powers the consumer-facing Hiya app as well as the native Samsung dialer, which means it pulls user-reported spam data from a huge installed base. Hiya's Q2 2025 Global Call Threat Report disclosed it flagged 13.7 billion suspected spam calls in a single quarter, giving a sense of the data volume its algorithm is trained on (Hiya Global Call Threat Report). Numbers reported in the Hiya app get penalized across the AT&T network, even if the complaint came from a T-Mobile user.
Hiya also weights:
- Call-pattern anomalies (burst dialing, 24/7 activity)
- CNAM-to-entity consistency
- Registered-caller enrollment status (Hiya Connect branded calling)
- Cross-network reputation signals via Hiya's data exchange
Remediation path: Hiya Connect, Hiya's business portal, allows legitimate businesses to register numbers, attest to ownership, and dispute incorrect labels (Hiya Registration announcement). Registration does not override behavioral scoring — it raises the floor.
TNS Call Guardian
TNS's Call Guardian leans heavily on cross-carrier traffic analysis. TNS sits on a huge portion of US carrier-to-carrier signaling traffic (it is a major SS7/SIP interconnect provider), so it sees patterns Hiya and First Orion do not: not just complaints, but the raw shape of call traffic across networks. TNS's 2025 Robocall Report found 85% of inter-Tier-1 traffic was signed with STIR/SHAKEN and 93% of that signed traffic was at Level A (TNS 2025 Report). By 2026, TNS was flagging over-attestation — up to 13% of signed traffic on invalid numbers receiving Level A — as a new risk class its algorithm weights against (TNS 2026 Report).
TNS weights:
- Call-origination-network reputation (is your carrier trusted?)
- Call completion and short-duration-hangup ratios
- Complaint density normalized against dial volume
- STIR/SHAKEN attestation level — TNS is notably strict here; Level C attestation is a significant negative signal
- Number age and history of ownership changes
Remediation path: TNS Enterprise Branded Calling allows verified businesses to register and remediate. The intake is more rigorous than Hiya's, often requiring proof of legal entity, DNS validation, and carrier attestation.
First Orion
First Orion is the algorithm that produces "Scam Likely" — the most recognizable spam label in the US. It powers T-Mobile's network, which means if your T-Mobile connect rate is underperforming, First Orion is where to look. First Orion won a Juniper Research gold award in 2025 for its SENTRY call-blocking solution, which sits alongside INFORM (its branded-calling product) across all major mobile carriers (First Orion INFORM).
First Orion weights:
- Direct complaint volume via T-Mobile's Scam Shield app
- Call-behavior fingerprints (extremely heavy weighting on hangup-before-ring patterns)
- Number-neighborhood reputation (if the 20 numbers adjacent to yours are dirty, you inherit suspicion)
- Branded calling enrollment (INFORM program)
Remediation path: First Orion's INFORM program lets businesses verify identity and request review. First Orion has historically been the slowest to review and the most conservative about removing labels once applied.
Why registering with one is not enough
This is where most agents and agencies get burned. They see a "register your number to stop spam labels" offer from one vendor, pay for it, and assume they are done.
They are not. They are clean on one-third of the US mobile network.
A real-world example of cross-vendor divergence (typical, not a specific case):
| DID | Hiya (AT&T) | TNS (Verizon) | First Orion (T-Mobile) |
|---|---|---|---|
| 555-0100 | Clean | Clean | Scam Likely |
| 555-0101 | Clean | Potential Spam | Clean |
| 555-0102 | Spam Risk | Clean | Clean |
| 555-0103 | Clean | Clean | Clean |
Four numbers, three of them flagged on exactly one network each. If you tested only by calling your own phone (one carrier), you would conclude everything is fine. Your prospects would disagree.
What operators are reporting
The cross-vendor pattern shows up consistently in agent-community discussions. A long-running thread on Insurance Forums, "Need help: Number getting flagged as spam," has licensed agents describing numbers that look clean on one carrier and burned on another, exactly matching the vendor-split problem (Insurance Forums thread). A BiggerPockets thread collecting real-estate operators' experience with the same issue documents the same confusion — carriers offer no single remediation path because the labels are controlled by three separate private vendors (BiggerPockets thread).
The registration maze
To be fully registered across all three you need:
- Hiya Connect enrollment — business verification, DID attestation, optional branded display.
- TNS Enterprise Branded Calling — deeper business verification, often requires carrier coordination.
- First Orion INFORM — identity verification, brand submission, periodic re-attestation.
Each is a separate onboarding, separate fee structure, separate portal, separate review timeline (typically 2–6 weeks per vendor from clean submission to active status).
On top of that, the registrations do not eliminate the algorithm — they raise the floor for legitimate behavior. If you start dialing 400 calls a day with a 2% answer rate, the algorithms will still flag you. Registration lets you remediate faster when problems occur; it does not immunize you.
The feedback-loop problem
All three algorithms incorporate real-time consumer complaints. That creates a structural challenge for insurance outbound:
- You dial a prospect who does not remember opting in.
- They hit "Report Spam" in their dialer.
- That report feeds directly into the analytics vendor tied to their carrier.
- A small cluster of similar reports on a low-volume DID can tip it into a labeled state.
TNS's own survey data showed 43% of US respondents filed a robocall complaint with a government agency in the prior year — consumers are actively reporting, and the analytics engines are listening (TNS 2026 Robocall Report).
The only defenses are (a) lead consent hygiene that keeps reports rare, and (b) continuous monitoring so you catch the flag early and rotate before the reputation decay compounds. See What a "Clean" DID Actually Looks Like in 2026 for the operational playbook.
How the three compare at a glance
| Dimension | Hiya | TNS | First Orion |
|---|---|---|---|
| Primary carrier | AT&T | Verizon | T-Mobile |
| Consumer app footprint | Huge (Samsung native) | Minimal | T-Mobile Scam Shield |
| STIR/SHAKEN weighting | Medium | High | Medium |
| Complaint-driven weighting | High | Medium | Very high |
| Review/remediation speed | Fastest | Slowest | Medium |
| Registration cost range | Low–medium | Medium–high | Medium |
What to do if you are already labeled
- Confirm which vendor flagged you. Call your own number from a phone on each of the big three carriers. Whichever displays a label is the vendor you need to work with.
- Open a remediation ticket with that vendor through their business portal. Be prepared to prove business identity and DID ownership.
- Fix the underlying signal. If you got flagged for volume or short-duration patterns, remediation without behavior change just gets you flagged again.
- Do not rotate yet. Rotating a flagged number without remediation wastes a DID and teaches the algorithm nothing.
- Audit your full pool monthly. Prevention is cheaper than remediation.
For the broader structural context, see Why Insurance Agents Have the Worst Contact Rates in B2C Sales.
FAQ
Q: Which carrier uses Hiya vs TNS vs First Orion? A: Hiya Inc. powers AT&T (and Cricket) plus the native Samsung dialer. TNS Call Guardian powers Verizon and US Cellular. First Orion powers T-Mobile, Metro, and Boost. MVNOs typically inherit the parent carrier's vendor.
Q: Can I just use a dialer that "handles" labeling for me? A: Some dialer vendors have relationships with one or two of the three analytics companies. Very few have clean relationships with all three, and those relationships do not substitute for your own business verification.
Q: What about smaller carriers — MVNOs, regional carriers? A: Most MVNOs inherit the parent carrier's analytics vendor (e.g., Cricket uses Hiya like parent AT&T). Some regional carriers pick one of the three independently.
Q: Do these vendors ever get labeling wrong? A: Yes, and they all acknowledge it. Each has a dispute process. Review timelines typically run from several days to a few weeks depending on vendor and volume.
Q: Can I pay to be exempt from labeling? A: No. Branded-calling programs improve display and raise the baseline, but no amount of payment exempts you from behavioral scoring.
Q: How often does my reputation score update? A: Hiya and First Orion update near-continuously based on public vendor guidance. TNS updates in batches but typically within the same day.
Q: Does a flagged number on one vendor eventually spread to the others? A: Not directly — the three do not share scoring data. But the underlying behavior that flagged you on one vendor is likely to flag you on the others too, so it is effectively simultaneous.
Q: Should I tell my prospects to unblock my number? A: It cannot hurt for warm prospects, but it does not fix the algorithm — it just fixes that one contact list. Scalable fix has to be at the DID/reputation level.
Q: How do I remediate a spam likely label with First Orion? A: Through First Orion's INFORM program portal. You will need to verify business identity, submit the affected DIDs, and provide context for the label review. First Orion is historically the slowest of the three to act on remediation, so start early.
Stop guessing. Run the audit.
LineAudit checks your first 20 DIDs free. You'll see exactly which numbers are flagged, mis-registered, or at risk — in 30 seconds.