Skip to content

TRUST & DATA SOURCING

Everything we tell you, your competitor published.

Vaarta is built on public record. Pricing pages, changelogs, job reqs, reviews, filings, press, executive posts — things a company chose to put in the open, collected on a schedule and kept. There is no inside track, no purchased contact list, no scraped inbox, and nothing that requires a login we were not given.

That constraint is also the product. Because every signal is a public artefact with a URL, every claim built on it can be walked back to the thing itself. If we cannot show you where a claim came from, you never see the claim — it is discarded before delivery, not flagged.

The eight sources, and what each one cannot tell you

8 CATEGORIES · ALL PUBLIC

Each verified event is built from one or more of these. The third column is the part most trust pages leave out: what the source is blind to, so you know where the record stops.

01

Website and pricing pages

Full-page snapshots, diffed against the last one, so a silent edit to a price, a plan name or a seat minimum is an event rather than a rumour.

DAILY · SNAPSHOT DIFFED

BLIND TO

Anything behind a login or a “contact sales” wall. If the number is only quoted on a call, the page cannot show it.

02

Changelogs and release notes

Shipped features, deprecations and quiet capability additions, taken from the changelog itself rather than the launch post.

DAILY

BLIND TO

Work that is built but unannounced, and anything shipped only to a private beta cohort.

03

Careers and hiring signals

Public job requisitions from applicant tracking systems: role, team, location, and the responsibilities that reveal what is being staffed.

DAILY · ATS FEEDS

BLIND TO

Internal transfers, contractor teams, and roles filled without ever being posted.

04

Review platforms

New reviews and rating movement, read as themes across a window rather than as individual opinions.

DAILY · G2, CAPTERRA, TRUSTPILOT

BLIND TO

Aggregate sentiment we cannot attribute. We will say a theme recurs; we will not tell you it is the most common complaint.

05

News and press

Coverage across outlets, deduplicated so one funding round is one event instead of nine notifications.

CONTINUOUS · DEDUPLICATED

BLIND TO

Paywalled bodies we cannot fetch. A headline alone is not enough to write a summary against, so it does not ship.

06

Social and executive communications

Public posts from company accounts and named executives speaking in their professional capacity — positioning shifts usually appear here first.

CONTINUOUS

BLIND TO

Private accounts, direct messages, and anything a person posted as a private individual.

07

LinkedIn company intelligence

Public company-page facts: headcount bands, leadership changes, office locations, and the shape of team growth.

WEEKLY

BLIND TO

Individual employee profiles as a dataset. We read leadership changes, not rosters.

08

Financial and M&A activity

Funding rounds, acquisitions, and registrations and filings — new regional entities tend to precede launches by a couple of quarters.

CONTINUOUS · PUBLIC FILINGS

BLIND TO

Private company financials. Revenue, burn and churn are not public, so we do not claim to know them.

What we will not touch, ever

Competitive intelligence has a grubby end. These are the lines, written plainly so you can hold us to them in a security review rather than infer them from a marketing page.

Credentials or logins

We do not create accounts on a competitor’s product, borrow a customer’s login, or collect anything that requires authentication. Nothing gated is in the record.

Purchased personal data

No bought contact lists, no enriched personal profiles, no email or phone datasets. Vaarta tracks companies and the public behaviour of companies.

Confidential material

Leaked documents, internal decks and anything shared with us in confidence by a third party stay out, regardless of how useful they would be.

Impersonation or pretexting

Nobody poses as a prospect, a candidate or a journalist to extract information. Every collection method would read the same to the company being tracked.

Reproducing paywalled work

We link to journalism, we do not republish it. Where a body cannot be fetched, the signal does not ship rather than being summarised from a headline.

Predictions dressed as facts

The analysis layer is instructed never to predict, and every figure in a claim must already exist in the verified record. Speculation is a different product.

SECURITY REVIEW

Ask for the specifics — retention, sub-processors, where signals are stored, what a deletion actually deletes — and you get the actual answer from someone who wrote the pipeline. Talk to us.

Public is the floor, not the promise

A public source is only the raw material. Every candidate still has to clear four gates — entity resolution, relevance, classification, summary — and a candidate that fails any of them does not become a signal — what dies at each gate. Then the analysis layer, where a model says what a set of signals implies, carries the hardest rule in the product.

Each implication must reference the signals it rests on, and those references are checked against the signals that actually exist for your workspace. An implication that cites nothing, or cites something that cannot be found, is not stored — so it never reaches a briefing, a digest or the API.

Every analysis served by the API carries the ids of the signals underneath it, and each id resolves against the signals endpoint over the REST API or the hosted MCP server. You can walk from a conclusion back to the evidence without asking us.

THIS HAS THROWN AWAY REAL WORK

34 implications, dropped wholesale

On one analysis run the model returned its citations in a format the validator did not accept. All 34 implications in that run failed the check and were dropped rather than published ungrounded. The run cost real tokens and produced nothing deliverable. That is the correct outcome: the system would rather say nothing than say something it cannot stand behind.

Those claims were not fabrications — they were drawn from real signals, and the citations pointed at real evidence. They were dropped because the validator could not mechanically resolve the citation format, and a citation it cannot resolve is a citation it cannot verify. We fixed the parser afterwards. We did not touch the rule that discarded the output.

The same claim, shipped and discarded

Every implication is rendered with the verified signals it was drawn from, linked to their sources. Here is one that ships, and one built from the same evidence that does not.

SHIPS

Entry pricing moved from $29 to $49 in eight weeks, which is consistent with a deliberate move upmarket rather than a one-off test.

GROUNDED IN 2 SIGNALS

  • Entry tier raised from $29 to $49 per month · competitor.com
  • Starter plan removed from the pricing page · competitor.com

Both signals carry the source URL they were verified against, and every figure in the claim — $29 and $49 — appears in the verified record it came from. Nothing in the sentence is unsupported, so the sentence ships.

DROPPED

Entry pricing moved from $29 to $49, and roughly 40% of their customers churned as a result.

The first half is the same well-founded claim, citing the same two signals. The second half contains a figure, 40%, that appears nowhere in the verified record. The check is mechanical: every number in a claim has to already appear in the verified signals the analysis pass was given. This one does not, so the whole claim fails. It is not shown with a caveat. It is not stored at all.

Being exact about the boundary: that check is deterministic and catches invented figures. It is not a truth oracle. A claim can carry no numbers and still be a thin reading of the evidence, which is why the analysis layer is instructed never to predict, and why every claim ships with its sources attached for you to judge.

What this does not claim

Public is not the same as complete

A gate tuned to drop unverifiable signals will sometimes drop a real one, and things that were never published are simply not in the record. We would rather miss an event than invent one.

Grounded is not the same as right

A grounded claim can still be a weak reading of a real event. Grounding guarantees the evidence exists and is linked, not that the interpretation is the best one available.

This page is a snapshot

Sources change their terms, their formats and their availability. When a source stops being collectable we remove it from the record rather than substituting a guess.

See the public record on your competitors.

Track one competitor free, no card required. Every signal arrives with the source it was verified against.