# Specific > Specific is a software (SaaS) which lets people create and launch conversational surveys that feel like natural chat interactions (AI Surveys). These AI-driven interviews can ask follow-up questions in real time, capturing deeper insights than traditional survey forms. The platform then uses GPT-based AI to summarize responses, distill key themes, and allow teams to interact with the data conversationally-similar to ChatGPT, but tailored to your user feedback. ## Docs - [JavaScript SDK](https://www.specific.app/documentation/js-sdk) - [API Documentation](https://public-api.specific.app/docs/introduction/welcome) ## Interactive demos (survey examples and use cases) - [Interactive demos](https://www.specific.app/interactive-demo): list of interactive examples of AI surveys, where people interested in these surveys can try them. ## Changelog - [Changelog](https://www.specific.app/changelog): new features, fixes and improvements to Specific. ## Information about the product Specific offers two types of survey products: - **In-product surveys:** AI Survey installed in a software product, which displays as a widget. Usually for SaaS products, apps or websites. - **Landing Page Surveys:** AI Survey on it's own landing page, sharable with a link. ### Who It's For Specific is great for everyone who needs to needs to do a survey. Primarily: - product managers - UX / user researchers - growth teams - customer experience professionals (CX) - entrepreneurs ### Example Use cases **In-product surveys:** - Validate new features or product-market fit - Measure Net Promoter Score (NPS) - Understand why users churn or hesitate to upgrade - Conduct market research for new ideas - Gather contextual feedback without scheduling interviews - ROI validation **Landing Page Surveys:** - Internal company poll for employees - U.S. political poll - Public opinion poll - Coworking space feedback - Kindergarden parent feedback ### Problems It Solves Traditional surveys and user interviews are often time-consuming, impersonal, and lack contextual relevance. Specific addresses these challenges by: - Delivering surveys at the right moment inside a product, based on user behavior - Engaging respondents in a natural, conversational format that increases participation and quality of answers - Automatically analyzing responses with GPT-based AI to surface actionable insights - Allowing teams to filter, segment, and interact with the data to uncover deeper understanding ### Key Features - **AI-powered Survey Creation:** Instead of filling out a form, teams can simply chat with an AI to create their survey. It's faster, easier, and typically results in a better survey than building one manually. - **Chat with GPT about responses:** Enables teams to chat directly with GPT about survey results-like ChatGPT, but with full context of user conversations. - **AI Summaries:** Automatically distills every response, including multi-select choices, into core insights using GPT-powered AI analysis. - **Advanced Targeting:** Allows triggering in-product interviews for specific users based on their identity and behavior within the product. - **Data Integrations:** Supports syncing customer data through a JS SDK or public API for targeting, and exporting results to fit existing workflows. - **Expert-made Templates:** Offers a library of ready-to-use templates created by research experts for quick survey setup. - **Code & No-code Events:** Supports triggering different interviews without needing to modify the product's source code. - **Localization:** Enables surveys to run in multiple languages simultaneously, allowing users to respond in the language they use in-app. - **Survey on a Landing Page:** Provides a sharable survey via a unique link to a standalone landing page-no embedding required. - **Easy Installation:** Can be embedded into a product or website with a simple one-time installation. ### How Specific works and what it can do Specific is built on a structured yet flexible survey engine that powers each AI-driven conversation. At the heart of every survey is a combination of predefined logic, customizable behavior, and GPT-powered interactivity that enables richer, more natural feedback collection-without sacrificing control or structure. #### Survey Anatomy Each survey consists of one or more questions, arranged in a defined order. Question types include: - **Open-ended questions** - great for qualitative insights - **Single-select multiple choice questions** - ideal for structured answers with optional probing - **NPS questions** - with tailored follow-up logic for promoters, passives, and detractors - **Ending message** - a required final message that can end the survey or leave room for continued conversation All questions are static in wording but come alive through dynamic conversation. Each can optionally include **follow-up logic**, where the AI asks additional questions in real time, based on the respondent's previous answers. The number, tone, and purpose of these follow-ups are fully customizable-survey creators can define exactly what kind of information the AI should collect, when it should stop, and what it should avoid (e.g. not asking about discounts). For example, an open-ended question can instruct the AI to: - Ask "why" to understand motivation - Clarify ambiguous phrases - Explore real-life use cases or unmet needs These follow-ups are not scripted-they are generated dynamically using GPT, allowing the AI to explore responses in a way that feels natural and human. #### Survey Creation & Customization Surveys can be created manually or by chatting with the AI survey builder, which transforms plain-language prompts into structured surveys. The platform supports a range of custom settings: - **Tone of Voice** - Define how the AI should speak to respondents (e.g., professional, casual, brief, friendly, etc.) - **Language Support** - Set the survey's default language, and optionally enable automatic multilingual support. When enabled, respondents see the survey in their app's language, without needing to manage translations manually. - **Question Logic** - Configure follow-up intensity (e.g., persistent probing or one-off follow-up), maximum follow-up depth, and branching behavior for different answer types (especially in NPS). #### Delivery & Targeting Specific supports two delivery methods: - **In-product widget** - Embedded directly into a SaaS app or website, appearing as a small, non-intrusive chat interface - **Sharable landing page** - A standalone URL that hosts the AI survey, useful for email distribution, public polls, or internal feedback For in-product delivery, targeting options are highly customizable: - **Placement** - Choose where the widget appears (e.g., bottom right, center with overlay) - **Timing** - Delay appearance by a set number of seconds or only show after a number of visits - **Event Triggers** - Show surveys based on code/no-code triggers in the product (e.g., when a user completes an action) - **Frequency Controls** - Define how often a respondent can see a survey or re-take it (great for recurring NPS or feature usage checks) You can also define whether the survey should collect responses continuously or stop after a fixed number. #### Conversational Experience When a user responds to a survey, the AI (referred to as the `aiAgent`) conducts the conversation. It: - Delivers the question in a natural, chat-like format - Dynamically generates follow-up questions based on context - Adapts its language and tone based on your defined settings - Responds conversationally even at the end of the survey-acknowledging input, expressing gratitude, and optionally continuing the chat if enabled The AI is designed to feel like a smart product researcher, not just a form filler. #### Insights & Analysis After responses are collected, the platform uses GPT-based AI to analyze the data: - **Summarizes individual responses**, even for complex open-ended answers - **Distills key themes** across all responses - Lets teams **chat directly with GPT about the results**-you can ask things like "What are the top reasons for churn?" or "How do power users describe our product's value?" This allows teams to explore their user feedback as if they had a research analyst on-demand-without needing to export data or build dashboards.