List Pretrained Models

2 mins read - Updated on Oct 28, 2026

Kimola has 15 pretrained models: ready-made AI models that analyze customer feedback without any setup. 14 of them are classifiers, which tag every piece of feedback with labels, and 1 is an extractor, the Entity Extractor, which picks out the names a text mentions. This list gives each model's purpose, its number of labels and a link to its page on kimola.com, where you can read the labels and try the model. A new Report starts with two of them: the Automatic Classifier and the Sentiment Classifier.

Pretrained models are available in every package, including the free Starter package. Their labels are set by Kimola; to classify feedback with labels you define, create a custom model instead (see Custom vs. Pretrained Models). This list was last checked on October 28, 2026.

Classifiers

A classifier tags every piece of feedback with one or more labels, such as a topic or a sentiment. Kimola's pretrained classifiers fall into three groups: general classifiers that work on feedback from any business, industry classifiers with the topics of one sector, and two classifiers for news topics and hate speech. Each model's name below links to its page on kimola.com/pretrained-ai-models.

General Classifiers

These three classifiers work on feedback from any business. The first two are the ones a new Report starts with.

Automatic Classifier (no fixed labels) is a content classifier that finds the themes in the data itself and names them after what customers say. It needs no list of labels, so the number of themes depends on each dataset. Use it when you look at a dataset for the first time and don't yet know what customers talk about.

Sentiment Classifier (2 labels) detects whether each piece of feedback is positive or negative, in any consumer conversation. Its labels: Positive, Negative.

Consumer Feedback Classifier (12 labels) sorts what consumers say about a product, a service or a brand into common topics. Its labels: Quality, Features & Functionality, Ease of Use, Pricing & Value, Customer Service, Purchase Experience, Delivery & Logistics, Availability & Stock, Brand Perception, Marketing & Communication, Competitor Comparison, Sustainability & Ethics.

Industry Classifiers

Each industry classifier comes with 12 or 13 labels for the topics customers of one sector talk about. Choose the one closest to your feedback. If none of them fits your sector, a custom model gives you the labels you need.

E-Commerce Feedback Classifier (12 labels) is made for online shopping feedback. Its labels: Website & App Experience, Product Information, Pricing & Discounts, Checkout & Payment, Delivery Speed, Packaging & Condition, Order Accuracy, Returns & Refunds, Customer Support, Stock Availability, Loyalty & Rewards, Trust & Security.

Mobile App Feedback Classifier (12 labels) is made for the reviews and feedback of mobile app users. Its labels: Installation & Setup, User Interface & Design, Ease of Use, Performance & Speed, Bugs & Crashes, Features & Functionality, Updates & New Versions, Privacy & Security, Notifications & Alerts, Pricing & Subscriptions, Customer Support, Compatibility.

Mobile Game Feedback Classifier (12 labels) is made for the feedback of mobile game players. Its labels: Gameplay & Mechanics, Graphics & Visuals, Sound & Music, Story & Characters, Performance & Stability, Updates & New Content, Progression & Rewards, Multiplayer & Community, Fairness & Balance, Ads & Monetization, Fun & Engagement, Customer Support.

Software Feedback Classifier (12 labels) is made for the feedback of software product users. Its labels: Ease of Use, Features & Capabilities, Integration & Compatibility, Performance & Reliability, Setup & Implementation, Customer Support, Pricing & Licensing, Training & Documentation, Security & Compliance, Updates & Roadmap, Scalability & Flexibility, Return on Investment (ROI).

Online Course Feedback Classifier (12 labels) is made for learners' comments on online course platforms. Its labels: Content Quality, Instructor Performance, Practical Value, Learning Materials, Video & Audio Quality, Platform Experience, Ease of Understanding, Updates & Relevance, Pricing & Value, Certification & Recognition, Pacing & Workload, Community & Interaction.

HoReCa Feedback Classifier (13 labels) is made for the feedback of hotel, restaurant and café guests. Its labels: Food & Beverage Quality, Menu Variety & Options, Service & Staff, Cleanliness & Hygiene, Atmosphere & Ambience, Pricing & Value, Booking & Reservations, Accommodation Quality, Facilities & Amenities, Location & Accessibility, Delivery & Takeaway, Events & Entertainment, Brand Perception & Trust.

Example

In a multi-label Report of a hotel chain's reviews, the HoReCa Feedback Classifier might tag "The room was spotless, but breakfast had run out by nine" with both Cleanliness & Hygiene and Food & Beverage Quality.

Banking Feedback Classifier (12 labels) is made for the feedback of banking customers. Its labels: Account Management, Mobile & Online Banking, Fees & Charges, Customer Service, Transactions & Payments, Loans & Credit, Savings & Investments, Security & Fraud, Branch & ATM Experience, Product Variety, Trust & Reputation, Promotions & Offers.

Automotive Feedback Classifier (13 labels) is made for the feedback of car buyers and owners. Its labels: Performance & Power, Comfort & Interior, Design & Style, Technology & Features, Fuel Efficiency & Range, Safety, Maintenance & Repairs, Purchase Experience, After-Sales Service, Pricing & Value, Resale & Depreciation, Brand Perception, Environmental Impact.

Mobile Network Feedback Classifier (12 labels) is made for the feedback of mobile network operators' customers. Its labels: Network Coverage, Call Quality, Data Speed & Reliability, Network Stability, Roaming & International Use, Plans & Pricing, Billing & Charges, Customer Support, Promotions & Offers, Device & SIM Services, Security & Privacy, Brand Perception.

Other Classifiers

Two classifiers answer questions that aren't tied to one industry: which public topic a text is about, and whether it contains hate speech.

News Classifier (13 labels) sorts any text into news topics. Use it to categorize news articles, or to see which public topics consumer conversations touch. Its labels: Politics & Government, Business & Economy, Technology & Innovation, Health & Medicine, Environment & Climate, Education, Science & Space, Sports, Arts & Culture, Crime & Law, Society & Lifestyle, World Affairs, Finance & Markets.

Hate Speech Detector (2 labels) detects hate speech toward an ethnicity, a nationality, a religion, a sexuality, a disability or a class. It reads the context of a sentence, so a comment can be detected even when it contains no explicit slur. Its labels: Hate Speech, Not Hate Speech.

Extractors

An extractor doesn't assign labels from a list. It picks out the entities a text mentions and tells you what kind of entity each one is. Kimola has one pretrained extractor.

Entity Extractor (no fixed labels) finds the entities in a text and classifies each one into one of more than 35 categories, from people, organizations and locations to more specific ones such as food and beverage products, financial products, sports or movies. It reads each name in its context, so "Apple" in a complaint about customer service and "apple" in a review of a snack's flavor are told apart. Use it when your question is which names come up in customer conversations, such as the competitors customers compare you with.

Use a Pretrained Model

You choose pretrained models in the Classifiers step of the Report wizard. Open the PRE-BUILT MODELS tab, next to CUSTOM MODELS, and type a name in the filter to find a model. A new Report starts with the Automatic Classifier and the Sentiment Classifier, and you can add any other pretrained model or your own custom models. See Create a Report Using Pretrained Models.

The icon next to each pretrained model opens its page in a new tab. The page shows the model's description, a console where you can try it on a file (Document) or on your own text (Text), its Labels with a description of each, and More on Gallery, a list of the other pretrained models. See Test a Pretrained Model.

Whether one piece of feedback can get one label or several is set by the Report, in its Structure step (see Multi-label vs. Single-label Classification). Every model you add to a Report counts toward its queries: queries = rows × models × Dimensions (see How to Calculate Queries).

Pretrained models are used in Reports. A Funnel, and so every Space, classifies feedback with custom models only. Your own applications can also list Kimola's pretrained models and classify text with them through the Kimola API, where they are called Presets; the API is available in the Enterprise package.

Was this article helpful?

Tell us about your thoughts and experiences regarding the article.