Rflect
Swiss platform that embeds guided reflection into university curricula, turning everyday coursework into defensible longitudinal evidence of how students develop human skills, with AI support for lecturers through the Rflect Teaching Assistant.
What is Rflect?
Rflect is an assessment platform for the human skills that automated systems do not replace, built by Rflect AG in Zurich as a spin-off of the University of St. Gallen. The vendor describes it as infrastructure for teaching and assessing human skills in your curriculum at scale, and frames the problem plainly: employers, accreditors and students now want proof of critical thinking, judgement and resilience, and a final essay demonstrates none of them.
The platform combines five elements. Guided self-reflection is the backbone, with structured prompts that push students to connect course content to their own experience. Self-assessment uses scientifically validated scales at the beginning and end of a programme. Peer assessment gathers structured feedback from classmates and from outside people such as supervisors or mentors. Personal learning goals turn assessment results into concrete commitments. Peer coaching sends students off with a ready-made script and brings their conclusions back into the platform. Five competencies are highlighted throughout: collaboration, critical thinking, resilience, creativity and perspective skills.
Three features are AI-assisted. Go Deeper is an optional extension that analyses a student's written reflection and nudges them along the Gibbs Reflective Cycle. Personal learning goals can be reviewed by an AI coach to make them more concrete. The third is the Rflect Teaching Assistant, shortened by the vendor to Rflect TA: an assistant trained on the company's own library of reflection topics and pedagogical practice, which builds the reflection timeline, reviews topics, surfaces themes from shared student reflections, produces engagement reports and guides lecturers around the interface. It is a capability inside the platform rather than a separately sold product.
Rflect slots into an existing module or runs as a standalone programme-level course across several semesters, connecting to Moodle, Canvas and Blackboard through LTI 1.3. It is used as a credit-bearing component at several universities, typically worth one to three ECTS per semester. Students work from native iPhone, iPad and Android apps or any browser, while lecturers manage programmes from the browser. By design, lecturers cannot see individual self-assessment results, only aggregated class results, and never see private reflections.
What it does
- Send guided reflection prompts on a rhythm set by the lecturer across the course timeline
- Have students set, revisit and refine personal learning goals with AI feedback
- Run self-assessments on scientifically validated scales at the start and end of a programme
- Organise structured peer assessment and peer coaching sessions between students
- Build a complete reflection journey in roughly thirty minutes by talking to the Rflect Teaching Assistant
- Track cohort engagement in real time and spot students who need support early
- Export accreditation-ready evidence of competency development
When to use Rflect / When not to
A quick filter to help you decide if Rflect is the right fit.
When to use Rflect
- Lecturers in higher education who want structured reflection in their course without spending hours building it
- Programme directors and department chairs who need cohort-level visibility and early warning on disengaging students
- Accreditation and quality teams looking for direct evidence of competency development, such as AACSB Assurance of Learning
- Instructional designers and learning technologists integrating a reflection layer through LTI 1.3
- Bachelor and master students who want a documented record of their own growth rather than a single exam grade
When not to use Rflect
- Individuals: the service is sold exclusively to organisations, and direct consumer self-signup is not offered
- Anyone seeking therapy, counselling or crisis support, which the vendor explicitly excludes
- Buyers who need a published price list before starting a conversation, since no amount is disclosed anywhere
- Teams expecting to plug the tool into their own systems through a product API, which does not exist publicly
- Institutions needing an interface language other than German, English or Spanish
How to use Rflect
A typical end-to-end flow, from setup to results.
- A lecturer or authorised representative registers the institution through the Try for Free link; individual consumers cannot sign up on their own
- Accept the institutional terms with the checkbox at registration, then agree an Order Form covering seats, modules, AI features, term and price
- Connect Rflect to the learning management system over LTI 1.3 with Moodle, Canvas or Blackboard, or simply use browser access
- Define the competencies to develop and map the programme learning outcomes into the platform
- Build the reflection timeline, either manually or by asking the Rflect Teaching Assistant, reachable from the smiley at the bottom right of the lecturer dashboard
- Set the privacy setting for each topic: private, shared with the lecturer, shared with the class, or shared with the class anonymously
- Pass the student terms and the privacy notice to users, who accept them at first login
- Students answer prompts each week from the mobile app or a browser, and can enable push notifications when a new reflection opens
- Students who want AI feedback on their private reflections give a separate in-app opt-in, which they can withdraw at any time
- Lecturers monitor engagement, intervene early where needed, and export evidence for accreditation
Pros & Cons
Pros
- Unusually strict privacy architecture: private reflections are technically unreachable from institutional dashboards and from the vendor's own internal administration tools
- The whole AI chain stays in the EU on AWS Bedrock, with an ephemeral prompt cache capped at one hour and no permanent storage of prompts or responses
- An explicit contractual commitment not to train AI models on reflection content or lecturer-created content
- Complete, dated legal documentation published in the open: institutional terms, student terms, DPA and technical measures, with all four sub-processors named and their own agreements linked
- Real portability: students export what they wrote without their institution's approval, and keep a thirty-day window after an account is deactivated
- Integrates into the existing LMS through LTI 1.3 rather than adding another isolated tool
- Verifiable adoption with named universities, and an Innosuisse study with the University of St. Gallen reporting significant gains across all five Inner Development Goals dimensions
Cons
- No public pricing at all: no amount, no grid, not even a range, so evaluation starts with a sales call
- Sold only to organisations, so an individual lecturer or independent trainer cannot buy access
- No documented product API, which limits custom integration to what LTI 1.3 covers
- The interface is available in German, English and Spanish only
- Vendor liability is excluded as far as the law permits, including for data loss, and the service is provided as is without warranties
- Students have no direct contractual recourse against Rflect: their only counterparty is their own institution
- Focused firmly on higher education, with corporate and secondary-school use still at pilot stage
Pricing & Plans
No price is published. Rflect is sold on a quotation basis: the vendor offers pilots as well as department-wide and university-wide licence agreements, with the amount depending on the number of students and programmes and on the level of support. All arrangements include the full platform functionality. Contractual prices are stated in Swiss francs, exclusive of VAT, with the billing cycle, licence period and price set out in an Order Form and invoices payable within thirty days. Because no amount is disclosed anywhere on the site or in the terms, the starting price, currency and billing unit fields are deliberately left empty together rather than estimated. The service is free of charge for students, since the institution pays the licence. A Try for Free link appears on every page, but no trial duration or condition is published.
- Pilot
- Department-wide licence agreement
- University-wide licence agreement
Data, GDPR & hosting
A consolidated view of how Rflect handles your data.
GDPR overview
GDPR implementation is documented in concrete, dated form. The privacy notice, last updated 30 March 2026, states it is designed to meet both the GDPR and the revised Swiss Data Protection Act. A Data Processing Agreement under Article 28 GDPR and Technical and Organisational Measures under Article 32 GDPR are published as downloadable documents dated 16 June 2026, the DPA forming an annex to the institutional terms. Data subject rights are enumerated, breach notification to the controller is committed to within 24 hours, and transfers outside Switzerland and the EEA rely on EU Standard Contractual Clauses with the Swiss FDPIC addendum. The data protection contact is privacy@rflect.ch. One gap stands out: no EU representative under Article 27 is designated, although the vendor is established in Switzerland and serves EU universities.
Who owns the data?
The institution is the data controller and Rflect AG acts as its processor, so the university owns the data and decides how it is used. Students and institutions keep ownership of the content they upload, including reflections, answers and media. One category is deliberately carved out: reflections a student marks as private are processed by Rflect as controller and are never disclosed to the university, the lecturer or other students, and at the end of a contract they are deleted or returned to the student rather than handed to the institution. Rflect takes a non-exclusive, royalty-free right to use outputs to operate and improve the service and, in anonymised form, for performance reporting and academic research, but not for marketing.
Reuse rights
Students may export or migrate everything they have authored on their own initiative, without needing their institution's approval, and regardless of the sharing setting they originally chose. Exporting produces a copy and does not erase the institution's own records. If an institution deactivates an account, deletion of the student's self-authored data is held for thirty days so it can still be exported. Each reflection carries one of four sharing settings: private, shared with the lecturer, shared with the class, or shared with the class anonymously. Reuse is otherwise restricted: the terms forbid copying, redistribution, commercial exploitation, reverse engineering and automated scraping. Users remain solely responsible for AI outputs and are required to verify them before relying on them.
Data retention & training
Hosting summary
All application data sits with AWS in the Frankfurt region, eu-central-1, in Germany: the application itself, the PostgreSQL database on RDS in a private VPC subnet, logs and backups. AI inference runs exclusively through the EU inference profile of AWS Bedrock using Anthropic Claude, inference only, with no fine-tuning. Bedrock stores neither prompts nor responses; a prompt may sit in an encrypted, ephemeral cache inside the EU inference environment for at most one hour before automatic deletion. Four sub-processors are named with their locations: AWS in the EU, mainly Frankfurt; Friendly Analytics in Switzerland; Sentry in the EU; and Pipedrive in the EU, Frankfurt. The data processing agreement records no third-country transfer for any of them, and where one might arise the vendor commits to EU Standard Contractual Clauses with the Swiss FDPIC addendum. Note one discrepancy: the privacy notice states more broadly that data may in exceptional cases be processed in any country. Encryption is TLS 1.2+ in transit and AES-256 at rest. The certifications cited in the technical measures, including ISO 27001 and SOC 2 Type 2, belong to AWS as the infrastructure provider, not to Rflect AG.
Things to keep in mind
Risks and trade-offs to weigh before adopting Rflect.
- Reflections are inherently sensitive personal writing about struggle, doubt and wellbeing; the vendor itself asks students not to put names or student identifiers into them
- Reflection is often graded or credit-bearing, worth one to three ECTS, which creates pressure to write the expected answer rather than an honest one
- Leaning on the Go Deeper AI coach can replace the cognitive effort of reflecting instead of supporting it, which defeats the purpose of the exercise
- The platform is explicitly not a medical, therapeutic, counselling or crisis service, so a student in distress must be directed to a qualified professional, not to the app
- AI outputs may be inaccurate and the terms place the duty to verify them on the user, whether student, lecturer or institution
- Reflections shared with a lecturer or a class are processed by AI without a separate opt-in, because sharing counts as consent; students need to understand where that line falls
- Vendor liability is excluded as far as the law permits, including for data loss, and contract and master data are kept for ten years under Swiss statutory retention rules
Setup & Integrations
Technical difficulty
Technically light, pedagogically demanding. A lecturer or authorised representative registers the institution and accepts the terms with a checkbox; no development is required because there is no API to wire up. Connecting to Moodle, Canvas or Blackboard over LTI 1.3 with OIDC launch does need the institution's LMS team, but the platform also works through direct browser access with no LMS at all. Students install nothing unless they want the iOS or Android app. The real effort is instructional design rather than IT: deciding which competencies to develop and building the reflection timeline, which the Rflect Teaching Assistant reduces to about thirty minutes.
Deployment
Apps stores
Integrations
Supported languages
Behind Rflect
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Rflect actually do?
What is the Rflect Teaching Assistant?
What can a lecturer see, and what stays hidden?
Where is the data hosted?
Are student reflections used to train AI models?
Who owns the data?
Can Rflect count toward ECTS credits?
Which learning management systems does it connect to?
Can a student refuse the AI features?
How much does it cost?
Should you pick Rflect?
Rflect is a deliberately narrow tool, and that is its strength. It does not try to be a general education platform: it addresses one problem, proving that students developed judgement, collaboration and resilience, and it builds the whole product around structured reflection as the mechanism.
The most convincing part of the offer is not the AI. It is the data architecture. Reflections a student marks as private are placed out of reach of institutional dashboards and of the vendor's own internal tools, students can export what they wrote without asking their university, and the commitment not to train models on reflection content is written into the institutional contract rather than promised on a marketing page. The full legal stack, terms, data processing agreement and technical measures, is published and dated, with every sub-processor named. For a young company that is unusual, and it matters here because student reflections are among the most sensitive data a university handles.
The AI layer is a genuine time-saver rather than the point of the product. The Rflect Teaching Assistant compresses what the vendor says used to take hours into roughly thirty minutes of building a reflection journey, and Go Deeper gives students a prompt to think harder. The company is explicit that it wants the human kept in the loop.
The reservations are commercial rather than technical. Nothing about the price is public, so any evaluation starts with a sales conversation, and individuals cannot buy at all. Liability is excluded about as far as Swiss law allows, and students have no direct recourse against the vendor. Corporate and secondary-school use remains at pilot stage. For a university that already wants reflection in its curriculum and needs evidence that will survive an accreditation review, it is a serious candidate.
- Choosing a selection results in a full page refresh.
- Opens in a new window.