SOCRATES is a structured dialogue application designed to deepen your thinking, not replace it. It does not answer for you — it helps you examine assumptions, test evidence, and follow your reasoning further. Analysed across five cognitive dimensions.
8–15 minute sessions · Five cognitive dimensions · Measured over time
Every productivity tool is built to do cognitive work for people. The user’s role is to approve outputs.
SOCRATES inverts this. The system does not complete the thinking for the user. It keeps the user inside the reasoning process.
Summarizes so you don’t read. Drafts so you don’t write. Decides so you don’t think. Each use is optimised to do more of the cognitive work for the user.
Uses questions to surface assumptions, test evidence, and make reasoning easier to examine. Each session is designed to make reasoning visible, measurable, and improvable.
A 20-second look
SOCRATES never gives you the answer. It asks the question that surfaces the gap in your reasoning — then asks why you didn’t catch it yourself.
The Method
Real dilemmas from ethics, strategy, evidence analysis, and logic. Each designed to surface specific reasoning patterns. The user reads, reacts, and commits to a position.
A city reports a 30% crime drop after expanding police presence. Your team calls it proof the policy works.
Your company's AI hiring tool is 15% more accurate but produces racially disparate outcomes.
Your startup's biggest competitor just raised $50M and pivoted into your core market.
Each intervention is designed to surface a specific kind of reasoning move, drawn from cognitive science and dialogue research. The system identifies the point in your reasoning that benefits from being examined and asks the question that opens it up.
Powered by advanced language models, guided by carefully designed Socratic dialogue patterns, SOCRATES adapts in real time to what you say — targeting the assumption you haven’t examined, the evidence you haven’t demanded, or the perspective you haven’t considered.
After every session, a structured analysis scores the session across five reasoning dimensions, using the user’s own words as evidence.
"You shifted from accepting a correlation as causal evidence to identifying three confounding variables — a meaningful move in Evidence Evaluation."
The Toolkit
Each intervention is designed to draw attention to a specific kind of reasoning move. Developed from cognitive science research, selected in real time based on what the user just said.
Surfaces beliefs the user treats as self-evident. "What are you assuming must be true for that to hold?"
Tests whether conclusions rest on sufficient evidence. "What would change your mind?"
Forces engagement with opposing views. "How would someone with the opposite experience see this?"
Traces consequences. "If that's true, what else must be true? Are you comfortable with all of it?"
Demands specificity. "You said 'most people.' How many? On what basis?"
Makes users observe their own reasoning. "You changed your position twice. What triggered each shift?"
Planned Validation Roadmap
These studies have not yet started. We intend to pre-register formal studies before data collection begins, and to report results transparently, including null or negative findings. Every outcome claim will be treated as a hypothesis until supported by data.
n=200, SOCRATES vs. waitlist control, Watson-Glaser pre/post, 6 weeks.
n=400, three-arm RCT: SOCRATES vs. unstructured AI dialogue vs. waitlist, 8 weeks.
n=300, observational cohort, W-G at 3, 6, and 12 months. Growth curve modeling.
Study designs and thresholds may be refined before pre-registration.
The Architecture
Most software is designed to maximize time-on-platform. SOCRATES is designed to graduate users. The four-phase curriculum progressively transfers the questioning skill from the system to the user until the tool is no longer needed.
Full scaffolding. The system drives all questioning. The user begins noticing gaps in their reasoning for the first time.
SOCRATES prompts the user to question themselves first. Self-questioning with scaffolding. Partial withdrawal begins.
The user initiates their own questioning. SOCRATES validates and refines. The system confirms the transfer is taking hold.
The user reasons independently. Critical questioning has become habit. SOCRATES is available but no longer necessary.
Market
Leaders, strategists, and decision-makers who want to think more clearly under pressure. SOCRATES is designed to help users examine assumptions and evidence before committing to a decision.
Critical thinking is one of the most requested skills in higher education and one of the hardest to teach at scale. SOCRATES is designed to give students structured one-to-one reasoning practice through any browser.
People who read, reflect, and want to get better at thinking. Each session is built as a short reasoning practice — eight minutes that, used regularly, are intended to support reasoning development over time.
Pricing
We're in early access. Everything is free while we refine the system with real users. Create an account to start training.
Paid tiers coming soon — beta users get early-adopter pricing
About
Every major technology system today is optimised to do cognitive work for people. They summarise so you don't read. They draft so you don't write. They decide so you don't think.
SOCRATES inverts this. It is a structured dialogue system that withholds the answer, asks the harder question, and leaves the cognitive work where it belongs — with the user. The system begins by asking all the questions. Over many sessions, the user begins asking their own. Eventually the Socratic voice becomes internal, the system steps back, and the user graduates. It is, by design, built to become unnecessary.
A stonemason's son in Athens developed a method of teaching he never wrote down. He asked questions that forced people to examine what they actually believe, discover where their reasoning breaks down, and construct something better. Cognitive science has since confirmed why it works: it develops metacognition, the single strongest predictor of intellectual performance.
Language models are the first technology that can support structured Socratic dialogue at any meaningful scale. The same capability that allows a system to process any question can be directed toward a different purpose: asking questions instead of answering them, with enough care to support genuine reasoning practice. For the first time, the method can reach a wide audience through any browser.
Building SOCRATES requires three disciplines in combination: Socratic pedagogy and reasoning education, computational linguistics and language model systems, and cognitive science. The founding team works at their intersection. The six intervention types derive from formal study of argumentation theory. The validation strategy reflects fluency with educational research methodology and psychometric standards. The architecture reflects direct experience with language model behaviour across prompt design, output evaluation, and model calibration.
We believe one of the most valuable applications of language modelling right now is helping people think better. Not think faster. Not think less. Better. Every outcome claim we make about supporting reasoning development will be treated as a hypothesis until measured against established psychometric instruments. We intend to report results transparently, including null or negative findings.
The unexamined life is not worth living.Socrates, in Plato's Apology
Get in touch: noetic.profectus@gmail.com
Research
Every outcome claim SOCRATES makes about supporting reasoning development will be treated as a hypothesis until tested against established instruments.
We do not make outcome claims without data. We intend to pre-register formal studies before data collection begins, using established psychometric instruments — including the Watson-Glaser Critical Thinking Appraisal — rather than inventing our own metrics.
Our planned validation roadmap spans three studies over 30 months, each with planned evaluation targets. If the data does not support our hypothesis, we intend to report that transparently and adjust the product accordingly.
SOCRATES measures reasoning across five empirically grounded dimensions, scored on a 1–5 scale after every session:
Careers
We are a small team building a structured Socratic dialogue application designed to make its users more capable, not more dependent. If that matters to you, we want to hear from you.
We're building at the intersection of Socratic pedagogy, language model systems, and cognitive science. We value intellectual honesty, rigorous thinking, and the discipline to publish null results. Titles matter less than what you can do and how you think.
We're especially interested in people with backgrounds in computational linguistics, educational psychology, LLM evaluation, or full-stack engineering — but the best people we've worked with have come from unexpected places.
No open roles listed right now, but we're always looking.
Send a note explaining what you'd build and why to:
noetic.profectus@gmail.com
Contact
Questions about SOCRATES, partnership opportunities, or press inquiries.
noetic.profectus@gmail.com
We welcome collaboration with researchers in cognitive science, education, and computational linguistics.
noetic.profectus@gmail.com
Noetic Profectus Ltd.
Legal
Last updated: March 2026
When you use SOCRATES, we collect the email address you provide at signup, your dialogue session transcripts (for scoring and improvement), and basic usage data (session timestamps, completion rates). We do not collect personal information beyond what you voluntarily provide.
Your session data is used to generate cognitive scores, track your progress over time, and improve the SOCRATES dialogue system. We do not sell your data. We do not share identifiable data with third parties. Aggregate, anonymised data may be used in published research.
Your data is stored on secure, encrypted servers. Session transcripts may be processed by third-party model providers solely to provide the service, subject to their applicable data-processing terms.
You may request access to, correction of, or deletion of your personal data at any time by contacting noetic.profectus@gmail.com. We will respond within 30 days.
SOCRATES uses only essential cookies required for the application to function, plus privacy-respecting product analytics (PostHog) to understand how the system is used. We never record full prompts, transcripts, spoken content, email addresses, or other personal content — only safe metadata such as page views, named button clicks, mode (typed or voice), message counts, and event types. We do not use advertising cookies, autocapture, or session-replay tools.
For privacy-related inquiries: noetic.profectus@gmail.com
Legal
Last updated: March 2026
SOCRATES is an educational reasoning-practice application that uses Socratic questioning to support critical thinking practice. It is not a substitute for professional education, therapy, or mental health treatment.
You agree to use SOCRATES for its intended purpose: reasoning practice through structured Socratic dialogue. You agree not to attempt to reverse-engineer, copy, or redistribute the system, its prompts, or its scoring methodology.
All content, methodology, scoring systems, intervention frameworks, and interface designs are the intellectual property of Noetic Profectus Ltd. Your session transcripts remain yours; the system's responses, scoring algorithms, and analytical methods remain ours.
SOCRATES is currently in beta. The service is provided as-is during the beta period. Features, scoring methods, and availability may change. We will provide reasonable notice before any material changes.
SOCRATES is an educational reasoning-practice tool, not a diagnostic instrument. Cognitive scores are designed to track relative improvement, not to assess clinical capacity. Noetic Profectus Ltd. is not liable for decisions made based on SOCRATES scores or feedback.
For questions about these terms: noetic.profectus@gmail.com
Legal
Last updated: March 2026
Noetic Profectus Ltd. is committed to compliance with the General Data Protection Regulation (GDPR). We process personal data only with lawful basis, typically consent or legitimate interest in providing our service.
We collect only the data necessary to provide the SOCRATES service: your email, session transcripts, and cognitive scores. We do not collect location data, device fingerprints, or browsing history.
Session transcripts may be processed by third-party model providers solely to provide the service, subject to their applicable data-processing terms. We do not use your identifiable session transcripts to train third-party AI models. De-identified or aggregate session data may be used to improve, evaluate, and validate SOCRATES, as described in this policy. Third-party model processing is governed by the applicable provider terms.
Active accounts: data retained for service delivery and longitudinal tracking. Deleted accounts: personal data removed within 30 days. Anonymized, aggregate data may be retained for research purposes.
You have the right to: access your data, rectify inaccuracies, request erasure, restrict processing, data portability, and object to processing. To exercise any right, contact noetic.profectus@gmail.com.
For data protection inquiries: noetic.profectus@gmail.com
Pick a scenario. Spend eight minutes with it. You should feel the difference within the first two questions.
Start an 8-minute sessionOccasional updates on new scenarios, product changes, and validation work.
Thank you. We'll be in touch.
Something went wrong. Please try again.