At Catalyx, we lean into the power of AI with humans at the helm. We live and breathe innovation and we’re committed to making success more predictable for tangible consumer goods brands. Our consumer-centric approach gets us there, which is bolstered by cutting-edge technology, particularly artificial intelligence (AI), used to enhance our processing power.
AI for us is a tool to deliver more accurate insights and process larger quantities of qualitative text data efficiently – part of what supercharges our Qual at Quant Scale approach – all while maintaining the highest standards of quality and ethics.
As expected with AI, and even explored within our whitepaper on the topic, we’re also aware of the concerns surrounding the new technology. To demystify our use of AI and give you a glimpse behind the curtain, we are dedicated to addressing these responsibly. So, let’s dive in!
How does Catalyx use AI within our proprietary insights platform?
Data Cleaning: Our use of Natural Language Processing (NLP) and Machine Learning (ML) ensures the cleanliness and validity of data. These technologies screen open-ended responses for gibberish or copied text and remove straightliners and speeders, thus ensuring high-quality data.
Translation: AI-powered translation tools enable us to handle multilingual data seamlessly. These tools provide high-quality, context-aware translations, which our language experts then refine to capture cultural nuances accurately.
Theme Identification: NLP assists in rapidly identifying patterns and insights from complex datasets through thematic analysis and word vectorisation. This allows us to extract meaningful themes quickly and effectively.
Data Exploration with Gen AI: Our strategists utilise enterprise-grade Generative AI (ChatGPT and Claude 2) to interrogate large datasets. This significantly increases the time-to-insight and reduces manual data handling work. Gen AI models help in identifying themes and generating concise summaries from qualitative data.
What safety precautions does Catalyx take?
Human-in-the-Loop: We use AI tools in conjunction with expert oversight by our strategists and data scientists to ensure accuracy and mitigate the risk of errors or hallucinations.
Validation: All AI-generated insights undergo multiple stages of validation and cross-verification by our human experts. Models are limited to providing analysis solely on the data provided and are prompted to offer verbatim evidence to support conclusions.
Data Security: All data processed by AI tools are handled within a secure environment under strict data security protocols to prevent unauthorised access or training. We strictly control the use of data, ensuring it is only utilised for intended research purposes and not for training third-party models.
Which services do we use?
Domo/Jupyter Notebooks: Domo, a data processing platform, is utilised in conjunction with Jupyter Notebooks to operate our proprietary ML/NLP models. We host and maintain these models in-house, ensuring that no data is shared with third-party services.
Google Cloud Translation API: We primarily rely on the Google Cloud Translation API for our translation needs. Google does not use the data processed through its API for any secondary purposes. For detailed information, please refer to their Data Usage FAQ.
Claude Pro: Owned by Anthropic, Claude Pro ensures that prompts and interactions are not used for model training. For further details, review Anthropic’s Data Security policy.
ChatGPT Team: Developed by OpenAI, ChatGPT is employed under a Team account, which guarantees that the data we input does not contribute to training the model. More information can be found at the Trust Portal.
Our use of AI is designed to augment and support our human experts, not replace them. We maintain full human oversight and accountability throughout the research process, ensuring the responsible and ethical use of these technologies.
Find out how our approach can make your innovation success predictable
We’re ready to unlock your innovation’s success, and won’t settle until our work does just that.
Don’t just take our word for it; check out our new website’s Case Studies to discover how our qual at quant scale approach has supercharged the success of innovations across industries.
Follow us on LinkedIn to make sure you never miss an announcement.
Let’s talk. Click here to innovate with confidence.
(Independent research conducted by Catalyx)
