Plain answers to the questions careful buyers ask: what this measures, how, what it can't do, and where it belongs in your decisions.
Quadrant is built on the DISC model, one of the most widely used frameworks for describing workplace behavior. Everyone is a blend of four styles; Quadrant surfaces your strongest two and how they shape your communication, decisions, and work under pressure.
Results-driven. Decides fast, pushes through resistance, hates stalling.
People-powered. Persuades, energizes a room, sells the idea.
Cooperative and consistent. Holds the crew together, finishes what it starts.
Analytical and careful. Gets the details right the first time.
On top of style, the full assessment measures five character signals:
The gap between how someone rates themselves and how they answer behavior-based questions. That gap is where blind spots live, and it is the part a resume can never show you.
The DISC portion asks what's most and least like you, which resists gaming far better than agree/disagree scales.
Self-ratings are paired with concrete-situation questions, with consistency checks (candour flags) that signal when answers look idealized.
Your scores are worked out by fixed rules, then the report itself is written by Anthropic's Claude from those scores against a fixed prompt. Every report has the same sections in the same order. The sentences are generated, not pulled from a library, so two people who score alike will not get word-for-word the same report.
Read-a-person works differently: it infers a likely style from text you already have (an email, call notes, a bio). Every result carries a confidence level. It is asked to list the cues it read the style from, and usually does, but that list is not guaranteed on every result the way the confidence level is. It is an educated inference from limited signals, and never more than that.
Honesty matters more to us than big claims. DISC describes behavioral style, not intelligence, skill, experience, integrity, or job performance. A profile is a snapshot shaped by how someone saw themselves that day. Quadrant's instrument is grounded in the established DISC framework, but it is a practical business tool. It has not been through the formal validation studies a clinical test goes through, and we don't claim it has.
Confidence on any single result should be moderate; patterns across conversations, references, and real work are always stronger evidence than any assessment.
Quadrant informs decisions; it never makes them. Never reject a candidate, promote, discipline, or dismiss anyone based on an assessment result alone. Use profiles to ask better interview questions, communicate better, and structure fairer processes, alongside skills, experience, references, and human judgment. That's both good practice and what employment and human-rights law expects.
Use the ideal-profile and interview-guide features to make interviews more structured and consistent between candidates, not to screen people out automatically. Apply the same process to every candidate for a role. A style "mismatch" is a conversation to have, not a verdict; plenty of people succeed in seats that don't match their natural style, with awareness and support. If a candidate asks about their results, share them.
Assessment results belong to the company that gathered them. They live in that company's own area of Quadrant. The people that company invites in can open them, and so can Quadrant, which operates the platform. We don't sell data. No other company signed in to Quadrant can search or open a profile you gathered. The next section is the whole picture, including the one route by which something you gathered could still reach another company.
Read this part first: the talent network is switched off. We designed a shared talent pool across the companies hiring on Quadrant. We are not running it. The job that would copy people out of a company's account, talent-sync.js, is written to stop at its first line unless a server switch called NETWORK_SYNC is set to "on": with the switch anywhere else it logs that it is paused and copies nothing. It is still on a daily schedule in netlify.toml, so it does get called and does nothing. Both files are in the code you can read. And there is no screen in Quadrant that lets one company search another company's people. If you are deciding whether to buy Quadrant, do not count the talent network as something you are getting today. Count it as something we intend to build.
An earlier version of this page said the rules below were enforced in code. That was not true, and we would rather correct it in public than leave it standing. Here is the part that is real. The tool that emails role invitations only ever reaches inside one company's own records, and it will send on one of two grounds and nothing else. Either the person approached that company themselves — applied to it, or handed it their details — within the last 180 days. Or they ticked a box asking Quadrant to email them about other roles, and we hold the timestamp for it. The second is a yes. The first is not a yes: it is us treating a person's own approach to a company as permission for that company to write back, for a limited time. A record somebody else put there, including anything copied in from another company, gets neither ground and cannot be emailed at all. The rest of the rules below are written down and not wired in. Until they are, we are not moving data.
What is true today about your people. What you gather stays in your company's area of Quadrant. Your team members, your applicants, the people you interviewed, the people you hired: no other company signed in to Quadrant can search them or open their records. Every read of a person in the product is scoped to the org of whoever is asking. People Purpose can see all of it, because we operate the platform, and the next paragraph is the one route by which something you gathered could still reach another company. We don't sell data, and we don't offer your staff to other employers.
One thing does still cross between companies, so here it is. People Purpose does recruiting as well as software. We keep our own pool of people who came to us directly, and when someone in it suits a role a client is filling, a person here puts that candidate into that client's pipeline by hand, one at a time. That is us doing recruiting work, not two clients trading databases. There is a tail on it you should know about: while the sync job was running it copied some client records into that pool, and those copies are still there. We are working through them. Until that is finished, this is the one route by which something you gathered could reach another company. Ask us about your own account and we will tell you straight.
What has to be true before anything is shared. Only people a company turned down. Only people who answered yes themselves. There is exactly one place that question is ever asked: a box on the assessment intro that starts unticked. A tick is written down with the date. Leaving it alone writes nothing at all — it is silence, not a no — and nothing asks a second time, so somebody who did not notice the box is never asked again. The rejection email that was meant to be the second ask is written and nothing sends it. That is a gap in the product, not a rule we are enforcing quietly. Never anyone marked as a team member. Never anyone a company screened, interviewed, or hired, because one screening call makes that person theirs for good. Only companies that switched it on. And a searching employer would see a first name, a trade, and roughly how far the person is from the job, with no email address, no phone number, no home address, and never the name of the company that turned them down. When the code enforces that list instead of this page describing it, this page will say so, and it will name the files you can check.
Companies that take part will be able to search the pool as well as contribute to it. That is the trade, and it belongs in the client agreement rather than buried here.
Some of the work is done by outside services on our behalf. The AI that writes the reports, drafts job descriptions, interview questions and review scorecards, and answers questions inside the app, is Anthropic's Claude. Email is sent through Resend. Anyone who takes the free assessment can have their own results emailed to them. Marketing emails are consent-based with one-click unsubscribe, and the series behind the sign-up boxes on this site is eight emails long, not open-ended. Questions or removal requests: rod@disc-quadrant.com.
We'd rather you ask than wonder. We'll walk you through the instrument, the reports, and how other companies run it.