Scalable&CustomizedsolutionsforyourBusiness

FrequentlyAskedQuestions

Everything you need to know about KnowDis AI search and discovery products. Can't find your answer? Talk to our team.

• ProductProducts & capabilities

  • What AI products does KnowDis offer for e-commerce?

    KnowDis offers three production-ready AI products: Gazo Search (visual/image search engine), an AI-powered multilingual semantic search pipeline, and Indifai (English-to-Hindi AI translator). Each integrates via API and is designed specifically for e-commerce platforms.

  • What is Gazo Search and how does it work?

    Gazo Search is a visual product search engine that lets shoppers find products by uploading an image - a screenshot, catalogue photo, or camera snap. The engine returns exact or visually similar product matches using computer vision AI models. It is particularly effective for fashion, home décor, and lifestyle product categories.

  • What is the KnowDis semantic search pipeline?

    The semantic search pipeline is a production-ready AI search engine built for Indian e-commerce. Key capabilities include:

    • Hinglish and Indic language query understanding
    • Automatic spelling correction and typo tolerance
    • Multi-attribute extraction (colour, category, gender, etc.)
    • Synonym expansion and semantic intent matching
    • Sub-10ms response time at production scale

  • What is Indifai and who should use it?

    Indifai is a state-of-the-art English-to-Hindi AI translation model optimised for e-commerce product descriptions and buyer-facing copy. It produces natural, easy-to-read Hindi that significantly outperforms generic translation tools like Google Translate - as validated by IndiaMART, one of India's largest B2B platforms. It is ideal for any platform serving Hindi-speaking buyers.

  • Does KnowDis offer a multimodal product recommendation engine?

    Yes. KnowDis's recommendation engine uses multimodal signals - text, image, and behavioural data - to surface the most relevant products for each shopper. It is designed to increase add-to-cart rates and reduce search abandonment by personalising results at the query level.

• Search and AISearch, AI & language

  • Does KnowDis support Hinglish and Indic language search?

    Yes. Our search pipeline is built specifically for how Indian shoppers search - including Hinglish (mixed Hindi-English), Roman-script Hindi, and regional language transliterations. This is a core differentiator: the pipeline was trained on Indian e-commerce query data, not generic English corpora.

  • How does the search handle spelling mistakes and typos?

    The pipeline includes a multilingual spelling corrector as the first stage of query processing. It handles phonetic misspellings, key transpositions, and abbreviated queries - and corrects them before semantic processing begins. This significantly reduces zero-result searches and improves conversion.

  • What industries and product domains does the search support?

    KnowDis's search pipeline is domain-adaptive and has been deployed across fashion and apparel, home décor, electronics, B2B industrial parts, auto components, and pharmaceutical catalogues. The models adapt to product-specific vocabulary without requiring manual synonym lists.

  • Is the visual search limited to fashion, or does it work across categories?

    Gazo Search was originally optimised for fashion and apparel - where visual search has the highest shopper intent - but is applicable to any visually-driven category including home décor, jewellery, electronics, and furniture. We recommend a discovery call to assess fit for your specific catalogue.

  • How does KnowDis's search compare to Elasticsearch or Solr?

    Traditional engines like Elasticsearch use keyword matching and BM25 ranking. KnowDis's pipeline uses transformer-based semantic models trained on e-commerce data, which means it understands intent rather than just matching tokens. This yields significantly better results for ambiguous, multilingual, or conversational queries - and requires no manual synonym or boosting configuration.

• IntegrationIntegration & technical setup

  • How do I integrate KnowDis products into my platform?

    All KnowDis products are available as REST APIs. Integration typically takes 1-3 business days for a standard e-commerce setup. Our team provides full API documentation, test credentials, and dedicated engineering support during onboarding.

  • What platforms and tech stacks does KnowDis support?

    KnowDis APIs are platform-agnostic and integrate with any stack - including Shopify, Magento, WooCommerce, custom-built platforms, and B2B marketplaces. Our APIs are REST-based and language-agnostic, supporting Python, Node.js, Java, PHP, and others.

  • How fast is the API response time?

    Our search pipeline delivers responses in under 10 milliseconds at production scale. We process over 7 million+ API requests per week across all client platforms, with consistent uptime and performance SLAs available for enterprise clients.

  • Do you support on-premise or private cloud deployment?

    Yes. In addition to our hosted cloud API, KnowDis supports on-premise and private cloud deployments for enterprise clients with specific data residency or security requirements. Please contact us at info@knowdis.ai to discuss deployment options.

  • What does the onboarding process look like?

    Our standard onboarding has four steps:

    • Discovery call: We understand your catalogue, search use cases, and traffic patterns
    • API access: We provide sandbox credentials and documentation
    • Integration support: Our engineers work with your team during integration
    • Go-live: We monitor performance and fine-tune models against your data

• CompanyAbout KnowDis AI

  • Who are KnowDis's customers?

    KnowDis works with e-commerce platforms across India and Southeast Asia, ranging from large B2B marketplaces like IndiaMART to fashion and lifestyle brands. Our solutions are used at scale across industries including fashion, industrial parts, pharmaceutical, and consumer electronics.

  • What makes KnowDis different from other AI search vendors?

    Several things set KnowDis apart:

    • India-first: Our models are trained on Indian e-commerce query data, not Western corpora - making them materially better for Hinglish and Indic language search
    • Research-grade team: Our ML team includes researchers from IITs, IISc, DSTI Paris, and Carnegie Mellon
    • Custom architectures: We build search systems around your catalogue and users, not off-the-shelf components
    • Production proven: 7M+ API requests/week, sub-10ms response time, deployed across diverse domains

  • Does KnowDis have academic partnerships?

    Yes. KnowDis has active research partnerships with the Yardi School of AI at IIT Delhi and IIT Bombay. These partnerships inform our model architecture choices and give us early access to cutting-edge AI research in natural language processing and computer vision.