Step into the future of building information and safety with our exclusive interview featuring the innovative mind behind our upcoming new feature, Goldenai. In this insightful article, we interview Ian Yeo, the dynamic CEO at the helm of our innovative platform. The spotlight of our conversation shines on Goldenai, the latest feature set to revolutionise how users engage with our platform. Goldenai is not just a feature; it’s your new personal assistant, meticulously designed to enhance your digital experience. This advanced tool promises to streamline navigation, provide tailored support, and bring a new level of efficiency and personalisation to your interactions with our platform. Join us as we delve into the inspiration, technology, and aspirations behind Goldenai.
Before getting into finite detail, we wanted to ask Ian his overall thoughts about Goldenai. Have a read of what he said below:
“I’m incredibly excited about this new feature. The results we’ve been seeing are astonishing, surpassing our initial expectations. This feature not only provides easy access to the content of your building information but also offers valuable insights into your data. We’re all too familiar with the ‘Information Tax’ – the cost incurred every time we search and disseminate information throughout a building’s entire lifecycle. Goldenai is designed to be your solution to this challenge, acting as your personal assistant.
The development journey over the past six months has been intensive, focusing on delivering this innovative solution to our customers. What’s been truly remarkable is the rapid evolution of the technology underpinning this feature since its inception, expanding the possibilities and potential it holds.”
On that note, let’s get started with the interview:
Is our Operance data sent to an external AI provider?
Rest assured, your data with Operance remains entirely secure and is never shared with external AI providers like OpenAI (ChatGPT). We strictly host your data within the UK or EU cloud infrastructure, ensuring it stays within a controlled and protected environment. This means that external AI entities neither retain nor utilise your information for their model training purposes.
Regarding the use of your data for training large language models, we adhere to a stringent policy of non-disclosure and non-utilisation. All data is securely stored within our AWS-provisioned cloud infrastructure, which prohibits its use as training material for any large language models. Your privacy and data integrity are our top priorities, ensuring a secure and trustworthy experience with Operance.
Will the AI answers be accurate?
You can expect a high level of accuracy from the AI-generated answers, though it’s important to understand that they may not always be 100% accurate. Our extensive testing has shown that the AI’s responses are generally more accurate than direct human answers. This is because our AI is specifically programmed to base its answers solely on your Operance data. However, like any advanced system, there might be rare instances where an answer isn’t wholly correct. To help you quickly verify the information, we will provide direct links to the specific Operance data that the AI used to formulate each response. This feature is designed to give you added confidence and clarity in the accuracy of the answers provided.
What types of questions provide the best answers?
To obtain the most effective answers, it is advisable to ask specific questions. For example, a query like ‘provide me with my structural risks’ will yield more precise results compared to a broader request such as ‘provide me with all my risks.’ The rationale behind this is the AI’s design: it analyses a selected subset of your Operance data for each question. By narrowing your inquiry, you enable the AI to concentrate on a specific dataset, resulting in more targeted and accurate responses.
What Operance data is used for generating the AI answer?
When generating AI responses, we utilise the complete set of Operance data available for each project. This comprehensive data includes documents in open file formats (PDF, Word, Excel and more), metadata, and structured building data. However, it’s important to note that our AI currently does not analyse proprietary file formats. This means formats like DWG drawings and Revit model files aren’t included in the analysis. But if these drawings are exported as PDFs, they are then incorporated into the AI’s data pool for generating answers.
Does the AI use information in scanned PDFs from paper documents and image files?
Certainly, our AI system is capable of using information from scanned PDFs and image files. We employ the latest Optical Character Recognition (OCR) technology to initially analyse these scanned documents. This advanced OCR analysis ensures that all your Operance data, including information from scanned sources, is comprehensively utilised by the AI to provide the most accurate and relevant answers possible.
How does Operance AI work?
At Operance AI, we have implemented a two-step process to ensure that the answers you receive are both accurate and relevant. Initially, we meticulously analyse the content of your question. This allows us to extract the most pertinent information from your Operance data. Following this, we utilise the insights gathered in the first step to formulate and provide you with an answer that is directly tailored to your query. This thorough approach is designed to ensure that the information you receive is not only precise but also perfectly aligned with your specific needs.
How do I know what information has been used to provide answers?
To ensure transparency and clarity, after each response, we include convenient links that direct you to the specific pieces of Operance data used to generate the answer to your question. This feature allows you to easily review and understand the exact information that informed the AI’s response, giving you a clearer insight into how your query was addressed.
If I need tabular information can this be exported to a spreadsheet?
“If you require information in a tabular format, you can easily export it to a spreadsheet. First, request the data in a table format through Operance. Once the table is displayed, you can quickly transfer this data to Excel by selecting the entire table and using simple keyboard shortcuts. Press ‘Ctrl + C’ to copy the table, and then ‘Ctrl + V’ to paste it into your Excel spreadsheet. This method offers a fast and efficient way to move your data into a spreadsheet for further analysis or record-keeping.
Does Goldenai come with an additional cost, or is it included?
Goldenai is currently available as a beta feature, and if you’re granted access, there’s no additional cost involved (subject to fair usage policies). We are still evaluating the outcomes and implications of this feature, which will guide our decision on future pricing. However, it’s worth noting that it is unlikely there will be an extra charge for using the chat feature as it stands now.
Is there a log or history of conversations that we can reference back to?
Yes, a log or history of your conversations with Goldenai will be available for reference. However, this feature will be introduced in stages over a few weeks. In the initial release, the history log won’t be present, but rest assured, it’s on its way and will be added shortly! We’re excited to roll out this functionality to enhance your experience with Goldenai.
Is there any chance of a data leak using Goldenai?
Rest assured, the security of your data when using Goldenai is our top priority. We adhere to stringent security measures to prevent any data leaks. When a query is made, only the specific data needed for that query is transmitted to the Large Language Model (LLM), which is securely hosted within the EU. Importantly, this data is not utilised for any training purposes, ensuring that your information remains confidential and secure at all times.
Do the conversations get logged as part of the Golden Thread?
While conversations in Goldenai are not directly logged as part of the golden thread audit trail, users have the flexibility to transfer valuable information from these interactions into their projects. Once this information is incorporated into a project, it becomes part of the audit trail. This approach is akin to how tasks are managed: the tasks themselves aren’t logged in the audit trail, but any subsequent changes or actions that arise from these tasks are recorded. This ensures that all critical information and modifications are accurately tracked and documented.
That is a wrap! In conclusion, Goldenai represents a ground-breaking advancement in the field of building information and safety. Thanks to Ian Yeo, our CEO here at Operance, we have gained valuable insights regarding what should be expected from Goldenai as we progress towards full release. Read more about our upcoming Goldenai feature here.
