Workings introduces a private way to capture how something was made and verify that process with others. It directly addresses the shortfalls of current Blockchain IP evidence, AI Detection, and Point of Creation tools while also supporting the modern user with various productivity and marketing benefits.
Blockchain IP evidence tools - not designed for the AI age
Blockchain IP evidence tools are used to support IP claims: disputes often turn on proving a creation existed at a particular point in time, which can be established by cryptographically anchoring the evidence. Existing tools achieve this proof by requiring manual upload of files. However, this evidence says nothing about how the file was made.
The shortfall here is that patents require a human inventor, and AI-generated work risks being refused copyright protection without demonstrable human authorship. Establishing that authorship calls for high-fidelity evidence of how the work was made - which is exactly what Workings captures.
Read more about how Workings is uniquely placed to meet the evidentiary challenges of IP developers.
AI Detection is fundamentally flawed
Much vaunted "AI detectors" look for patterns in a final output that are more likely to have been written by a machine than a human. The major problem here is that these patterns are not a reliable way to detect AI. They are just combinations of words that could be composed by humans or AI. The stakes are extremely high for this kind of accusation with reputations on the line.
A false positive is ruinous. A false negative, scandalous.
Suppose, for illustration, a detector with a 1% false-positive rate. Across 40 essays that's roughly a one-in-three chance that a given student is falsely accused at least once during their degree (1 − 0.99⁴⁰ ≈ 33%) … to say nothing of the ones who cheat and aren't caught.
False negatives are an increasingly common phenomenon with a whole product category called “AI Humanizers” dedicated to helping people avoid detection.
These fundamental flaws are addressed by Point of Creation tools which come at the authorship problem from a different, and more reliable, angle...
Point of Creation Tools
Point of creation tools record the process of creation as it happens. To date they operate at different points along a spectrum of compromise between the privacy and integrity of data. Here’s the overview of the different approaches in the market today:
Privacy | Integrity | ||
Private? | Full Story? | Verifiable? | |
User-Controlled Key-loggers | Yes | - | Yes |
Spyware / Bossware | - | Yes | - |
Examination Software | - | - | Yes |
Local Screen Recorders | Yes | Yes | - |
Workings | Yes | Yes | Yes |
Workings solves the compromise between Privacy and Integrity.
A perfect point-of-creation tool needs to do three things at once:
a) capture everything that went into the work.
b) prove the record hasn't been altered since.
c) never hand control of that data to anyone else.
Every other approach on the market trades off between these. Workings is the only point-of-creation tool that ticks all three boxes. Here's how it works.
1. Workings captures rich, high-fidelity data
Many point-of-creation tools record only keystrokes, which means they are inferring authorship from typed characters and typing rhythm rather than observing it directly. That inference breaks down in an ordinary case: a writer who types their own words but pauses to type an email or read source material. Without the context of what the user is working on there has no way to know what the keystrokes relate to. Further more it has no way to see what the user is referring to as they work. So while confirming that human typing is taking place it cannot be tied to a deliverable without taking further measures, like locking down the user down to a single tab.
Workings captures local high-fidelity screenshots as well as optionally collecting keystrokes. These can be used to fully describe the process involved in developing the final output. Analysis of those workflows provides a detailed record of how work was developed. Screenshots reveal the app that was in use at the time that text was written; the context of the text, the research sources being consulted, the AI tool that was open in another window, the reference document being copied from. That surrounding context is what lets a reviewer see how the work was actually made. For example, the difference between genuine human-led research and writing and someone copy-pasting AI output.
That same high-fidelity record is good for more than proof. Because Workings already captures how a piece came together, the process itself becomes an asset. Privately, you can study your own workflow to sharpen how you work, that analysis never leaves your device. If you choose to, you can turn the process into shareable content like provable timelapses, differentiating your work online as visibly, verifiably human. What you share is always your decision, made deliberately.
2. Workings is compatible with all applications
Some point-of-creation tools operate only within a specific editor (e.g. within GDocs, or a proprietary editing window). This means that the user has to change the way they operate to meet the requirements of the tool, and even then, the tool doesn’t know what is going on outside of the editor.
Workings collects evidence of everything visible on screen as the output is being created (except from any apps the user explicitly excludes - noting that any report generated will identify any apps that were used but excluded from the analysis).
A further technical challenge that Workings has overcome is in its ability to identify the relevant workflows that contributed to generating the final output. Load a document locally and the story of its creation, along with any gaps, will be revealed to the user.
3. Workings seals workflow evidence at the point of creation
Some point-of-creation tools capture evidence but don’t anchor it - meaning there is no way to prove when it was created or that it hasn't been altered since.
The moment you create your work, Workings records the evidence in a tamper-evident hashchain on your device, then anchors it to a public blockchain. The result is a record that your work existed by a specific point in time - verifiable by anyone, independently, and impossible to alter after the fact, including by us.
The hashes underpinning it stay strong against near-future quantum attacks: even a quantum computer only halves a hash's effective strength, leaving the ones Workings uses far beyond practical reach.
This means you have a tamper-evident record of how the work was made as it happened, plus independent, public proof that it existed by a specific point in time and hasn't been altered since.
4. Workings - your data stays under your control
Some point-of-creation tools stream your keystrokes, biometric signals, camera snapshots or behavioural metadata to remote servers. In this way they are compromising your privacy and forcing you to trust a third party with sensitive data. Workings is built the opposite way.
Capture and processing happen entirely on your device, where your data is encrypted and stored locally. Your raw keystrokes, screen content and behavioural signals never leave it.
What gets anchored to a public blockchain is a "salted" fingerprint: a one-way hash wrapped with a random key generated on your device and never transmitted. On the chain it's just an anonymous number - it reveals nothing about your work to anyone, including us, because the key and your evidence never reach our servers. What makes this useful is that you hold the key: supply it and the record becomes independently verifiable proof; withhold it and it stays private. Private by default, provable on demand - and that choice is always yours.
To prove a record, you share it. Whoever holds it can verify it against the chain themselves, without us. To forget one, you sever it: your key is destroyed, and the link between you and that work goes with it. What you share, you've chosen to share; everything else, you can erase. Nothing happens to your work without your say so.
This data-minimising, privacy-by-design architecture makes it far easier for you and your institution to meet GDPR and FERPA obligations. You decide what to keep, what to prove, and who sees it.
Like any app, Workings also handles the basics needed to run it - account, authentication and opt-in analytics. None of this touches your raw evidence. Read our Privacy Policy for more details.