The Information Quality Standard

AI checks its answer. Nobody checks the source.

Every AI product retrieves sources to answer you. Lens scores the document itself, sentence by sentence, with the evidence attached, on models we own. In the reader's hands today. One call away from every AI pipeline.

IQS · Document reportSample report

Regional health authority defends new screening policy

The policy, announced Tuesday, expands screening to adults over 40. Critics say the move is nothing more than a cash grab by an agency desperate for relevance. Officials cited two peer-reviewed trials supporting earlier detection. The outlet's parent company holds a stake in a competing diagnostics firm, which the article does not disclose. A spokesperson said implementation begins next quarter.

64 / 100
Composite · calibrated
Sourcing81 out of 100
Framing52 out of 100
Conflicts of interest58 out of 100
2 flagged passages · reasons attached
01The problem

The substrate goes unmeasured.

Trust has collapsed and the web is turning synthetic. No accepted standard measures the document itself.

49.9%

of new English-language online articles are AI-generated (Q1 2026).

Graphite, 2026
62%

of people say they cannot tell what is real from what is fake online in the news.

Reuters Institute, 2026
0

accepted third-party standards score the individual document.

The gap Lens fills
  1. Half the web is written by machines, and more of it is read by them every day.

  2. There is more information than anyone can verify, and the people publishing it have interests that bend what gets reported.

  3. Reader, analyst, or model: everyone is trusting sources nobody measured.

  4. The industry checks whether an AI's answer matches what it retrieved. Nobody checks the document itself.

  1. A bad article on a good site sails straight through a domain score.

  2. Capital creates conflicts of interest, institutions bend coverage to narrative, and writers reinforce it with technique.

  3. A measurement owned by the measured inherits the problem it was built to solve.

  4. It has to sit outside, and it has to show its work.

02The standard

A score with its evidence attached.

The IQS: a real-time, document-level quality score for public information.

Document-level

The single piece of text is the unit of judgment. Publisher reputation is a separate metric.

Evidence with every score

You see the flagged passages and a plain reason for each one.

Reproducible by design

The same document returns the same score, every time.

A measurement, not a verdict

A calibrated 1 to 100 per field. The truth call stays with the reader.

Live in production since May 2026Scoring live news today.
03One engine, every surface

The same problem, everywhere information is used.

Live

Lens Reader

The score, where people read.

Every score comes with its reasons. Free and paid tiers.

Open the reader
In development

AI Platform

A source check before the answer.

See the roadmap
Opening

Vertical add-ons

The same engine, tuned to one decision.

Legal · Marketing · Research · Education. First up: ad quality audits, in early access.

For business
04The serving path

No third-party AI in the serving path.

Our own models

Built, trained, and run in-house.

No outside dependency

Nobody else's model sits between you and the score.

Evidence, always

Every flag points to the exact passage.

05The Advisory Board

The people refining the standard.

Academics, diplomats, journalists, and career professionals refine the IQS and verify how it is applied.

What's next · coming soon

Lens Feed.

The news, scored before you read it. The highest-IQS coverage of each day's stories surfaces first.

Coming soon
06For business

Engineering the measurement layer the AI economy demands.

What is live, what is shipping next, and what is on the roadmap.

Explore the platform