Fred
Fred is an Italian roadmap-validation and decision-intelligence platform. Product, research, design and agency teams run one focused study, gather participant evidence, review AI-assisted signals and confidence, then share an inspectable decision case before engineering commits.
What is Fred?
Fred is a roadmap-validation and decision-intelligence platform published by Fred The User Research Shepherd SRL, based in Ozzano dell'Emilia, Italy. Its stated purpose is narrow and deliberate: help a product team settle one risky roadmap decision before it reaches engineering, design or budget commitment.
The unit of work is the Decision Evidence Sprint, a structured eight-step workflow that runs from the decision under test and the roadmap hypothesis, through the evidence collected and the signals and contradictions it produces, to a confidence level, a recommendation, the specific risk reduced and the next validation step. Around that spine sits a four-part loop: frame the roadmap risk, run the right research workflow, synthesise the signals with reviewable AI, then share the decision case.
Method coverage is broad for a single workspace. Teams can run unmoderated usability tests, moderated usability sessions and user interviews through the built-in User Sphere, first-click tests, prototype tests, closed, hybrid and open card sorting, tree testing, preference tests, five-second tests, and surveys with conditional logic. Insight Signals adds AI-assisted reaction cues, prosody-related context, click and task-flow evidence, and webcam eye tracking that needs no specialised hardware. Fred is explicit that these signals support human interpretation and are never automated certainty.
Analysis and output are where the product argues its case. AI-assisted thematic analysis groups qualitative material into themes that stay editable, auditable and tied to the source evidence. A report builder, with an AI-drafted starting point, turns that into a document stakeholders can inspect, and a searchable research repository keeps past studies from disappearing into folders and slide decks. A public library of thirty-six templates covers checkout drop-off, pricing-page clarity, site-structure card sorts, marketing message validation and more.
Customer-enabled integrations connect Google Analytics, PostHog, Google Calendar, Jira, Slack and OneDrive. AI processing runs on Amazon Bedrock in European AWS regions, with the application and its Supabase/Postgres database in Frankfurt. There is no product API and no mobile application.
What it does
- Frame one roadmap decision as a testable hypothesis instead of a generic research request
- Run the matching study: moderated or unmoderated usability tests, interviews, card sorts, tree tests, preference and 5-second tests, surveys
- Collect participant evidence — sessions, responses, task outcomes, behavioural and attention signals, qualitative context
- Surface converging signals and contradictions with AI-assisted synthesis that stays open to human review
- Attach a confidence level and its stated limitations to the recommendation
- Publish a stakeholder-ready decision report where every finding links back to its source evidence
- Keep past studies searchable as reusable research memory for the next validation cycle
When to use Fred / When not to
A quick filter to help you decide if Fred is the right fit.
When to use Fred
- Product managers who need to defend a roadmap bet with source-linked evidence before engineering capacity is committed
- UX and user researchers who want to cut synthesis and reporting drag without losing the trail back to the raw session
- Product and service designers validating flows, concepts, navigation and prototypes before decisions harden into handoff artefacts
- Research agencies and consultants standardising client delivery around repeatable studies, evidence and stakeholder-ready reports
- European buyers with a data-residency requirement, who need a vendor that documents EU hosting, a public DPA and a public sub-processor list
When not to use Fred
- Teams that want to drive research programmatically: Fred publishes no product API and no developer documentation
- Anyone expecting a mobile or desktop client, or a browser extension — Fred is a web workspace only
- Occasional or one-off users, since there is no permanent free tier and entry starts at 89.00 EUR per month for a single project
- Organisations that require a recognised security certification, as the vendor states it holds no SOC 2 or ISO 27001
- Anyone hoping to use behavioural or AI-assisted output as a stand-alone basis for hiring, education, insurance, credit or policing decisions, which the terms prohibit
How to use Fred
A typical end-to-end flow, from setup to results.
- Create a workspace from the self-serve signup; Researcher accounts start on a 15-day free trial
- Set up a project, the top-level container that will hold studies, evidence, insights and reports
- Frame the roadmap risk to be validated: demand, usability, comprehension, navigation, concept direction or priority
- Create a study inside the project and add the research methods that fit that question
- Configure the participant path: recruit from the built-in tester panel or build a reusable participant pool
- Launch the study and share the live link with participants by copy or email
- Review results as they land — task evidence, transcripts, analytics, behavioural and attention signals
- Run AI-assisted synthesis at study or project level, then edit and audit the themes against the source evidence
- Build the report, generate an AI draft if useful, and share it live or export it
- Optionally connect Jira, Google Calendar, PostHog, Google Analytics, Slack or OneDrive from the settings screen
Pros & Cons
Pros
- Every finding stays connected to its source evidence, clips and participant context, which is the whole point of the product
- EU data residency is documented rather than asserted: AWS and the Supabase/Postgres database in Frankfurt, Amazon Bedrock in European regions
- Legal documentation is unusually complete for a small vendor — privacy policy, terms, public DPA, sub-processor list, AI transparency statement, all dated and versioned
- The vendor is candid about what it does not have, stating plainly that it claims no SOC 2 or ISO 27001 certification
- AI limits are framed honestly: behavioural and prosody signals are presented as review material, never as proof of what someone feels
- Prices are public and three of the four plans are payable self-serve, with a 15-day trial and monthly, six-month or annual billing
- Broad method coverage plus webcam eye tracking means fewer separate tools for a research team to stitch together
Cons
- No product API and no developer documentation; the only OpenAPI file served from the docs site is an unmodified vendor demo template
- Web only — no mobile app, no desktop client, no browser extension
- No permanent free plan, and the 89.00 EUR per month entry tier is capped at one project and one study, with 1 GB of storage
- The most distinctive AI features — summaries, insights, sentiment cues, report generation, thematic analysis — plus eye tracking are reserved for the Team and Enterprise plans
- Prices are not in the pricing page itself but rendered by an embedded checkout widget, so the amount and currency shown depend on the visitor's location
- No contact page at all: the only commercial channel is a third-party meeting-booking link
- The site never states whether customer research data may be used to train models, and documents no opt-out mechanism
Pricing & Plans
There is no permanent free plan. All three self-serve plans include a 15-day free trial, after which the lowest price point is the Start-up plan at 89.00 EUR per month. Six-month and annual commitments reduce the effective rate, and the Enterprise plan is quoted by the sales team. Prices are served by an embedded checkout widget, so the currency and amount displayed depend on the visitor's location; the figures quoted here are the euro rates published by the Italian publisher.
- 5 seats
- 1 project
- 1 study per project
- 5 methods per study
- 1 GB of storage
- 45-minute session limit
- 5 report exports per month
- AI assistant included
- 15 seats
- 5 projects
- 10 studies per project
- 15 methods per study
- 15 GB
- adds first-click and five-second testing
- tester panel
- participant rewards
- 25 seats
- 15 projects
- 30 studies per project
- unlimited methods
- 50 GB
- 90-minute sessions
- adds conditional surveys
- hybrid and open card sorting
- custom seats
- projects
- studies and methods
- unlimited session time
- 100 GB baseline storage
- SSO and advanced security
- multi-team rollout support and governance setup
Data, GDPR & hosting
A consolidated view of how Fred handles your data.
GDPR overview
GDPR implementation is concrete and unusually well documented for a vendor of this size. The publisher is established in Italy, so no Article 27 representative is required or named. The privacy policy and the AI transparency statement are in force from 9 July 2026, the terms from 19 June 2026, each versioned and dated. Fred publishes a public DPA that can be incorporated into a contract, with an annex of technical and organisational measures, and a public sub-processor list naming twelve providers with their purpose and location. Access, rectification, erasure, restriction, objection, portability and the right to complain to a supervisory authority are all set out. Breach notification to affected customers is promised without undue delay. Consent is managed through CookieScript. No SOC 2 or ISO 27001 certification is claimed.
Who owns the data?
The customer keeps ownership and control of research data. Fred acts as a processor when a customer instructs it to host, organise, analyse or report on workspace research data, and only as an independent controller for its own website activity, business contacts, billing, security logs and support records. The customer decides why and how a study runs, who takes part, and what is uploaded, and remains responsible for lawful instructions and for participant notice outside Fred's native consent flows. Customers may request export or deletion of the data they control, subject to product capability, written agreement and applicable law. Fred may keep limited records for billing, legal duties, backup integrity, security investigation or fraud prevention.
Reuse rights
Customers can reuse their own research material freely inside and outside the platform: findings, clips, reports and exports belong to the customer's workspace, and export and deletion are available on request under the product capability and the written agreement. Fred processes that material only to run the service the customer asked for — study setup, participant workflows, scheduling, evidence storage, AI-assisted synthesis, reporting, repository search, support, billing and security. Provider integration data is never sold and never used for advertising, retargeting or unrelated marketing. Data received through Google APIs is additionally excluded from advertising and from training generalised AI or machine-learning models. Reuse limits come from the customer's own side: the terms forbid unlawful, infringing or clearly unrelated material, and forbid presenting AI-assisted output as ground truth about a person's feelings, intent, health, identity or legal status.
Data retention & training
Hosting summary
Core application servers and customer data are hosted in Europe. AWS infrastructure and the Supabase/Postgres database layer are configured in Frankfurt, Germany. AI-assisted processing runs through Amazon Bedrock in European AWS regions, with Stockholm, Sweden as the primary configuration and Frankfurt as a fallback or related path. The contracting entity for AWS is European. Cloudflare provides edge protection, in a European configuration for Fred's traffic where available, and is not a system of record for research data. The published sub-processor list places Stripe, Microsoft, PostHog, Sentry and Freshdesk in the European Union, Google in Ireland and the consent platform in Lithuania, while HubSpot data sits in Frankfurt; Meta is described as United States or European Union depending on the configured path. The publisher notes that some operational metadata may still be handled by providers with international operations, under provider terms, Standard Contractual Clauses where applicable and supplementary safeguards.
Things to keep in mind
Risks and trade-offs to weigh before adopting Fred.
- Behavioural, attention, eye-tracking and prosody signals are sensitive processing; the vendor pushes responsibility for participant notice and consent back onto the customer whenever collection happens outside its native flows
- AI-assisted reaction and sentiment cues invite over-reading — the vendor warns against treating them as statements about what someone truly feels, and that warning is easy to forget once a report is in front of stakeholders
- A confidence level printed next to a recommendation can manufacture false certainty; the underlying sample and limitations still have to be read, not skipped
- The pricing page shows no amount in its own markup: the figure and currency come from an embedded checkout widget and vary with the visitor's location, so budget against a quote rather than a screenshot
- Nothing on the site states whether customer research material may be used to train models, and no opt-out is documented — treat this as an open contractual question, not as a reassurance
- No minimum age is published anywhere, including for the tester pool, which matters when studies may recruit participants directly
- The publisher discloses no incorporation date, no team page and no named director, and takes part in startup ecosystem programmes; assess vendor longevity accordingly before storing years of research memory there
Setup & Integrations
Technical difficulty
Low. Fred is a web workspace with self-serve signup, nothing to install and no development work: there is no API to integrate. Even webcam eye tracking needs no specialised hardware. The only concept to absorb before starting is the project, study and method hierarchy, and studies are shared with participants by a plain link. Optional integrations with Jira, Google Calendar, PostHog, Google Analytics, Slack and OneDrive connect through OAuth from the settings screen. Public user documentation covers the whole path. SSO and a formal security review are Enterprise matters, negotiated with sales.
Deployment
Integrations
Supported languages
Behind Fred
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What problem does Fred actually solve?
Which research methods does Fred support?
Is there a free plan?
Where is my data hosted?
Which AI does Fred use?
Does Fred train models on my research data?
Is there an API?
What security and privacy documents are published?
Which tools does Fred integrate with?
How do I get in touch with the vendor?
Should you pick Fred?
Fred is a focused product, and its focus is its main argument. Rather than selling itself as one more UX research suite, it positions the whole workflow around a single question a team has to answer before it builds, and it keeps the chain from raw session to recommendation intact so a stakeholder can inspect the reasoning instead of trusting a summary. For a product or research team that keeps losing evidence between the study and the decision, that is a genuinely useful frame.
Two things make it credible beyond the marketing. The first is the legal and infrastructure documentation: a public DPA, a named sub-processor list, a dated AI transparency statement and clearly stated EU hosting are more than most vendors of this size publish, and they make Fred straightforward to take through a European procurement review. The second is the restraint — the vendor says outright that it holds no SOC 2 or ISO 27001 certification, and repeatedly frames its behavioural and prosody signals as material for human review rather than as findings.
The reservations are practical. The entry plan is expensive for what it allows: one project and one study, with the AI features that carry the marketing reserved for higher tiers, so the real starting point is further up the price list than the headline suggests. There is no API and no mobile client, which rules out any automated pipeline. And two gaps deserve a direct question before signing: the site is silent on whether customer research material may be used to train models, and the publisher discloses no incorporation date, team page or named director.
Worth evaluating if you run recurring product research in Europe and value traceability over breadth. Take the 15-day trial on the tier you would actually buy, not the cheapest one.
- Choosing a selection results in a full page refresh.
- Opens in a new window.