We turn noisy, scattered records into signal you can act on.
TrueSignal builds and operates data intelligence and AI-first businesses — one division that structures the regulatory and market data compliance teams are stuck reading by hand, and one that puts narrow, accountable AI agents to work for the businesses enterprise software skips.
A world where consequential decisions run on verified signal, not on whichever PDF was easiest to find.
Every product we build starts from the same complaint: the information that a filing, a bid, a payment, or a phone call depends on already exists — it's just scattered, unstructured, or sitting behind software nobody built for the person who actually has to use it. We close that gap, permanently, one vertical at a time.
Build and operate businesses that turn scattered records into structured, accountable intelligence.
TrueSignal Data extracts, normalizes, and entity-resolves public and licensed records into reference datasets compliance and analytics teams can query with a confidence score attached. TrueSignal Labs builds and runs narrow AI agents for the specific workflow of businesses that enterprise software has priced out or simply never noticed. Two divisions, one standard for evidence before we ship.
Five habits, not a poster
None of this is aspirational copy — it's the operating discipline every product on this page was actually built under, including the one instance where it cost us a live, paying release.
Verify, then build
Every dataset starts with a live check of the real source — its actual terms, its actual schema, its actual rate limits — never the version we assumed going in. A plan built on an assumption is a plan built on sand.
Integrity outlasts revenue
When a portfolio-wide compliance review found one live, paying product had outrun what its source's terms actually allowed, we took it offline the same week — before a regulator or a customer had to tell us.
Build for who gets skipped
The funeral home with no receptionist. The Tier-2 clinic with no automation budget. The property manager re-reading the same tenant email for the fourth time. We start where the well-funded tools stop looking.
Add value, never just resell
We never redistribute a source unchanged. Every product's entire job is the cleaning, entity resolution, and cross-referencing layered on top — the same reason a buyer pays us instead of the government portal directly.
Show your work
Every record ships with a confidence score, not just an answer. Every gap is marked NULL and documented, never quietly filled in. Trust is a design decision, not a slogan on a homepage.
Small, deliberately
We would rather ship one fully verified vertical than ten half-checked ones. Every product on this page passed through the same five-stage discipline before a single row reached a customer.
TrueSignal Data
Regulatory, trade, and institutional reference data — extracted, entity-resolved, and confidence-scored so compliance, fintech, and GovCon teams stop rebuilding the same cleaning pipeline in-house. Every release publishes at refdatum.com.
Problem: India's export/import flows sit in DGFT's annual portal exports with no stable entity identity across filings.
- HS-code-normalized trade flows, cross-validated against UN Comtrade
- Entity resolution across 40+ real spelling variants per exporter
- Confidence-scored records, versioned by release date
Problem: Compliance teams screen against OFAC, UN, and EU lists separately, in three formats, with no shared identity across them.
- Exact-ID and phonetic entity resolution across sources
- Alias graph with per-match confidence tiers
- Daily delta feed of adds, removes, and amendments
Problem: India's public tenders are technically open data, but nobody ships vendor win-rates or price benchmarks from them.
- Buyer, vendor, tender, and award linked as one graph
- Vendor win-rate and price-benchmark analytics
- MSME sole-proprietor names flagged for manual review, never assumed
Problem: EU tender data is fragmented across 27 national schemas with no cross-border bidder identity — a French SME and its German subsidiary read as two strangers.
- CPV-normalized notices from TED / SIMAP
- Cross-country bidder and winner de-duplication
- Award-value trend and incumbent-vs-new-entrant flags
Problem: India's school and university records live in disconnected UDISE+ and AISHE systems with no single institution master.
- Entity-resolved institution master across 33+ states
- Enrolment, infrastructure, and establishment-year aggregates
- Institution-level only — no student or teacher records, by design
Problem: 3.1M+ Indian corporate identities, no canonical ID across MCA filings, name variants, and group structures.
- A live, working entity master — genuinely built and shipped
- Taken offline by our own compliance review, not a regulator's
- Held here on purpose, as proof value #2 isn't just copy
TrueSignal Labs
Narrow, accountable AI agents — built and operated end-to-end for one specific workflow and one specific buyer, not sold as generic software and left for a small team to configure alone.
Problem: funeral homes, pest control firms, and marinas lose real customers to missed calls — and no receptionist software is built around how these businesses actually work.
- 24/7 call answering that sounds like a person, not a phone tree
- Appointment booking & message-taking tuned per business
- Setup built together, not a dashboard left for the owner to figure out
Problem: small accounting, construction, and translation firms want proven AI tools but don't have the hours to configure and tune them.
- Done-for-you setup of vetted AI tools, not a new platform to learn
- Ongoing tuning and support — never a one-time install
- Straightforward monthly pricing, no long contracts
Problem: subscription and D2C businesses lose revenue to failed UPI Autopay and card charges, and are priced out of recovery tools built for companies ten times their size.
- Automatic detection of failed UPI Autopay, card & mandate charges
- WhatsApp-first recovery cadence with a one-tap UPI link
- A dashboard that shows exactly how much revenue came back
Problem: property management teams re-answer the same "where's my deposit / maintenance request" emails all day, where one wrong reply is either a data leak or a missed fraud signal.
- Deposit, maintenance-status & lease-renewal intent classification
- Template-only replies validated against the property management system before send
- Built-in rent-diversion & landlord-impersonation fraud detection
One discipline, applied twice
TrueSignal Data's discipline is a literal pipeline. TrueSignal Labs applies the same instinct to a messier problem: a workflow, not a dataset. Neither division ships on a guess.
Extract
Pull raw records from the verified source only, checksummed and logged — never overwritten, never assumed.
Normalize
Clean, typed, snake_case fields. Ambiguous blanks become an honest NULL, never a guess.
Enrich
Entity resolution, lineage graphs, and cross-source reconciliation — the actual value we add.
QA
A hard gate: row counts, key integrity, and confidence-score distribution all checked before anything ships.
Publish
Versioned release, changelog, and a confidence score on every record a buyer can filter on.
Verify the source before we write a line of extraction code
- Read the real terms, live — not the category assumption. A source shaped like public data isn't always licensed like one.
- Probe the real technical shape — pagination limits, rate limits, and schema drift a vendor's own documentation rarely mentions.
- Ship the derived layer only — resolved, normalized, versioned — never a raw pass-through of someone else's file.
Separate drafting, validation, and the decision to act
- A Drafter proposes the response — from an approved template, never free-form disclosure.
- A Validator checks that draft against the real system of record and a restricted-field list before it's eligible to send.
- A Reviewer combines every upstream signal into a route: auto-send, human review, or a distinct fraud-review path — never guessed past.
India-first, built to extend
Every product starts from a real, local advantage — Indian payment rails, WhatsApp distribution, a specific regulator's data — before it's ever generalized to a new geography.
India
Home market. 4 of 5 live data products, plus RecoverTrue's UPI-Autopay-native recovery flow.
European Union
TrueSignal Data's first international release — TED procurement data across all 27 member states.
Southeast Asia
Exchange & legal-entity reference data in planning, built on OpenFIGI and GLEIF's open-licensed backbone.
United States
AnswerKept's home market — niche, underserved home-service verticals, sold city by city.
Global
QuietRollout and VerInbox operate wherever the buyer does — VerInbox ships in four languages at launch.
Building a compliance, data, or AI-ops product of your own?
We license data through TrueSignal Data and take on a small number of implementation partners through TrueSignal Labs each quarter.