licili
licili is a German AI text-analytics platform that turns customer feedback and open-ended survey answers into themes, drivers and sentiment. It pulls reviews from 29 connected sources, then guides teams from insight to concrete action.
What is licili?
licili is a German software-as-a-service platform that applies AI text analysis to customer feedback and open-ended survey answers. Published by licili UG (haftungsbeschränkt) in Göppingen, it addresses consumer-facing organisations that already gather more feedback than any team can read by hand, and its declared purpose is to unlock that material without manual effort.
The platform layers four complementary readings over the same corpus. Theme analysis surfaces topics and sub-topics by itself. Driver analysis isolates the themes that genuinely move star ratings and the Net Promoter Score. Context and sentiment analysis flags positive and negative outliers. Relevance analysis ranks individual verbatims by how much they actually say. In the vendor's own competitive write-up these sit on three levels, running from aggregated indicators down to sorted individual comments.
What separates licili from a generic chatbot summary is its modelling approach. It uses unsupervised machine learning: no coding frame has to be supplied, topics are proposed by the system, nothing needs training beforehand, new themes are picked up on their own, and no coder bias creeps in. licili describes its engine as a text understanding AI aimed at maximum comprehension of the text rather than at generating the most plausible answer.
Feedback arrives through 29 named connectors. Scrapers cover public review sources such as Amazon, Google Reviews, Trustpilot, Trustedshops, Yelp, X, Instagram, Facebook, the App Store, Google Play, Otto, MediaMarkt and Check24. APIs cover survey tools including Qualtrics, SurveyMonkey, Usabilla, Zenloop and Getfeedback. A webhook and Excel or CSV upload complete the list.
From there the Customer Experience Hub carries findings into action: cluster the insights, derive data-backed recommendations, then define development challenges and follow them through. A Process Agent flags anomalies proactively and can attach an estimated revenue upside to a fix. Reporting is assembled from widgets, and every attribute can be filtered, from language and star rating to operating system and sentiment.
licili claims 83% less time spent analysing feedback, a positive return on investment within 45 days and a 28% average rise in customer satisfaction, across more than 500 B2C organisations. Named references include Decathlon Deutschland, Thieme Gruppe, STABILO, FitX and Stadtwerke Krefeld.
What it does
- Analyse large volumes of customer feedback automatically, with no manual coding work
- Surface themes and sub-themes on its own, through unsupervised machine learning
- Identify which themes actually move star ratings and the Net Promoter Score
- Detect positive and negative outliers with context and sentiment analysis
- Rank individual verbatims by how much substance they carry
- Aggregate feedback from 29 connected sources, plus Excel and CSV uploads
- Turn findings into recommendations, development challenges and tracked measures
When to use licili / When not to
A quick filter to help you decide if licili is the right fit.
When to use licili
- Customer experience managers in consumer-facing companies who have to act on feedback, not merely count it
- Insights and market research teams processing open-ended survey answers at volume
- Product teams turning recurring customer wishes into a prioritised development backlog
- Brand and competitive intelligence analysts benchmarking their position against rivals' public reviews
- GDPR-sensitive German and European organisations that need feedback analysed under German hosting
When not to use licili
- Teams that still need to collect feedback: licili analyses existing data and ships no survey tool of its own
- Small operations with thin volumes, since the vendor puts reliable results at roughly 100 to 150 texts and profitability at 300 a month
- Developers looking for a documented public API or a mobile app: neither exists, and the terms permit browser access only
- Buyers who need a published price or a short commitment, given the 24-month minimum contract term
- Non-German-speaking teams: site, interface and contract are in German, and only the help centre is bilingual
How to use licili
A typical end-to-end flow, from setup to results.
- Book a demo with the sales team, or start with the free Potenzialanalyse self-assessment
- Register a free account and analyse a starter batch of texts at no cost
- Tell the Customer Success team which feedback sources you want connected
- Let them wire up the scrapers and APIs: licili says the integration is finished right after the first week
- Or upload existing datasets as Excel or CSV files, attributes included
- Let the AI propose the themes, since no coding frame and no learning data are required
- Work through the four readings: themes, drivers, context and sentiment, relevance
- Filter by language, star rating, operating system, sentiment or any other attribute
- Build reports and monitoring dashboards from the widget library
- In the CX Hub, derive recommendations, define challenges and track the measures taken
Pros & Cons
Pros
- No training phase: unsupervised learning means no coding frame, no learning data and no coder bias
- Twenty-nine feedback sources unified in one place, public reviews included
- Connectors are wired up by licili's own Customer Success team, reportedly within the first week
- Analysis runs past sentiment all the way to prioritised actions and measure tracking
- German hosting, a GDPR compliance claim and a contractual data-processing agreement
- Unlimited users on the account, a deliberate choice by the vendor
- A bootstrapped, profitable publisher with a dozen named enterprise references
Cons
- No published pricing whatsoever: there is no pricing page, and the terms point to an unpublished price list
- A 24-month minimum commitment, renewed in 12-month blocks with one month's notice
- Terms dated 18 February 2020 that contradict the product FAQ on the free quota, 500 texts against 1,000
- No permanent free plan: once the test phase expires, no further analysis can be started
- No public API and no developer documentation, no mobile app, and browser-only access imposed by the terms
- Availability guaranteed at only 95% monthly on average, with up to 100 hours of planned maintenance and 100 MB of storage
- Nothing published on model training against customer data, on any opt-out, or on security certifications
Pricing & Plans
licili does not publish its prices. The site carries no pricing page, and the terms of service refer to an order form and a price list that are not made public, so the entry-level cost can only be obtained from the sales team. A free trial is offered, though the two sources disagree on its size: the terms describe a three-month test phase covering 500 free texts, while the product FAQ advertises free registration and up to 1,000 free texts. There is no permanent free plan, since no new analysis can be started once the test phase expires. Paid licences are invoiced monthly or annually in advance depending on the licence model chosen, with a 24-month minimum term and an annual increase capped at 5%.
- licili operates no public pricing page
- Free test phase
- described as three months and 500 texts in the terms and as up to 1
- 000 texts in the product FAQ
- Paid licence model invoiced monthly or annually in advance
- priced individually through an order form and an unpublished price list
- Additional text quota purchasable from the account settings
Data, GDPR & hosting
A consolidated view of how licili handles your data.
GDPR overview
licili shows a DSGVO-konform badge on its homepage and names licili UG (haftungsbeschränkt), Göppingen, as the controller. Its privacy notice runs to sixteen sections built around the GDPR: definitions, cookies, registration, legal bases under Article 6, legitimate interests, retention, and every data-subject right taken in turn, from access and rectification through erasure, restriction, portability and objection to automated decision-making and withdrawal of consent. The terms provide that the parties conclude a data-processing agreement meeting the applicable legal requirements, and a signed agreement with HubSpot Ireland is cited by name. Google Analytics runs with IP anonymisation. Two things are missing: no data protection officer is named anywhere on the site, and no independent security certification such as ISO 27001 or SOC 2 is claimed. No Article 27 representative is designated, which is expected of a controller established in Germany.
Who owns the data?
Under licili's terms of service the customer remains the holder of all rights in the data it uploads, including ownership and copyright exploitation rights. licili receives only a limited licence to reproduce that data as far as this is necessary to deliver the contracted service. It may hold a copy in a fallback data centre and may alter the structure or format of the data to clear a fault. The application itself stays licili's property. The customer alone answers for the accuracy and quality of what it uploads. licili may have the contract performed by a subcontractor, and may name the customer as a reference once the contract is signed.
Reuse rights
Customers may export their data from the service at any time and without asking permission. licili is not obliged to hand it over actively, so the export has to be triggered before the contract ends. Use of the service itself is confined to the customer's own internal purposes and those of its affiliated companies: use by or for a third party is prohibited, whether paid or unpaid, and sub-licensing is not allowed. Employees and contractors acting on the customer's behalf do not count as third parties. Usage must stay within the functional scope set out in the service description, and uploaded content must not breach applicable law, public morals or third-party rights. On its product-development page licili states that sensitive product data is handled to high standards without transfer to other large language models. No clause, however, says whether licili trains its own models on customer data.
Data retention & training
Hosting summary
licili states on its homepage that the service is hosted in Germany, and the controller is established in Göppingen. Beyond that badge the documentation stays vague. The terms say only that licili delivers the service up to the internet connection of the data centre in which it is operated, without naming that centre, and reserve the right to use both its own IT systems and those of third parties, as well as to keep data in a fallback data centre whose location is not stated. Backups run on 24-hour cycles. The service description linked from the terms is itself hosted on IONOS's German cloud, which is an indication rather than a declaration. The marketing site is a separate matter: it is served by Webflow behind Cloudflare, and its address resolves to a US anycast edge node, which is a CDN location and not where customer data is processed. The privacy notice names third parties used by the website — HubSpot Ireland, Google Ireland, Automattic, Pinterest Europe and Sentry — but publishes no subprocessor list for the application itself. No hosting security certification is claimed.
Things to keep in mind
Risks and trade-offs to weigh before adopting licili.
- A 24-month minimum commitment with no public pricing means committing before you can compare offers
- Annual price rises of up to 5% count as accepted unless you object in writing within four weeks of the announcement
- All customer data is deleted 60 days after the contract ends, and licili is not obliged to hand it over: an export you forget to trigger is gone for good
- Availability is contractually only 95% on a monthly average, with up to 100 hours of planned maintenance per month and 100 MB of storage for your content
- Nothing is published about whether customer data feeds model training, nor about any way to exclude it
- Public reviews are gathered by scraping: honouring each source platform's own terms stays your responsibility, and the verbatims may carry your end customers' personal data
- Themes come from an unsupervised model that the vendor does not vouch for below roughly 100 to 150 texts, and a dashboard is never a substitute for listening to customers directly
Setup & Integrations
Technical difficulty
Low, and deliberately so. licili's own Customer Success team wires up every connector, scraper or API alike, and says the integration is finished right after the first week. No data science skill is needed, since the model is unsupervised: no coding frame and no learning data have to be prepared. A non-technical route exists too, uploading an Excel or CSV file with its attributes. The tool runs in the browser, with nothing to install. One customer praises the absence of a long setup phase; another notes that an adequate volume of data must keep flowing in.
Deployment
Integrations
Supported languages
Behind licili
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement licili.
Frequently asked questions
What exactly does licili do?
Do I have to train the AI or supply a coding frame first?
Which feedback sources can be connected?
Who sets up the integrations, and how long does it take?
How is this different from summarising feedback with ChatGPT?
How much feedback do I need for it to be worthwhile?
What does licili cost, and can I try it for free?
Where is the data hosted, and who owns it?
What happens to my data when the contract ends?
Is there an API or a mobile app?
Should you pick licili?
licili is a focused, mature vertical tool rather than a general-purpose AI assistant. It does one thing — read large volumes of customer feedback and open-ended survey answers — and it does it with an unsupervised model that spares the buyer the usual setup tax of a coding frame and training data. That, together with 29 ready-made connectors wired up by the vendor's own team, is what makes the promise of results within a week credible. The analysis does not stop at a sentiment score either: driver analysis ties themes to star ratings and NPS, and the CX Hub carries findings through to defined challenges and tracked measures. For a European buyer, German hosting, a contractual data-processing agreement and a bootstrapped, profitable publisher with a dozen named enterprise references are real reassurance.
The reservations are commercial rather than technical. Nothing about the price is public, and the contract asks for a 24-month minimum commitment, so you are asked to sign before you can compare. The contractual documentation is dated February 2020 and openly contradicts the current product FAQ on the size of the free trial, which is the kind of detail worth settling in writing before signing. Teams wanting a public API, a mobile app or an English-language interface will not find them, and nothing is published about whether customer data feeds model training or how to exclude it.
In short: a strong fit for a consumer-facing organisation with a real volume of feedback and a structured CX or insights team, provided the pricing and contract terms are negotiated with eyes open.
- Choosing a selection results in a full page refresh.
- Opens in a new window.