You can now say goodbye to Alexa and Siri. You’ll need to be an experienced facilities manager to fully appreciate the new digital assistant.
BrainBox AI announced on Tuesday a virtual assistant for buildings called ARIA (Artificial Responsive Intelligence Assistant), the maker of a AI-powered facility management solution.
ARIA is powered by AWS Bedrock. It is designed to increase building efficiency, by seamlessly integrating into the daily processes of building management.
ARIA is a system designed for commercial and retail buildings. It has the ability to predict operational issues, while keeping a close eye on blind spots in a building.
Jean-Simon Venne is the cofounder and chief technology officer of BrainBox AI. He said, “We took an AI agent that was autonomous and we plugged it in to our existing technology stack. It already had predictive capability.”
“It will tell you not only what happened in the last year, but what is going to happen over the next few months,” said he, speaking with TechNewsWorld. Then, it will give you some advice on how to prepare for the immediate future.
He added that it was “sort of like an ultimate adviser” which would tell you how to handle situations in the future.
Morale Booster
According to BrainBox the combination between its core AI for HVAC technologies and ARIA could have a significant effect on building operations, with HVAC energy savings of up to 25 percent and greenhouse gas emissions up to 40 percent.
ARIA may also affect the morale and productivity of facility managers. Building and facility managers are faced with a lengthy list of tasks. [that] Venne explained that they are always frustrated. By the end of their shift, they have only time to complete 20% of their task list. They feel like they are failing because there is too much to do.
He explained that ARIA was a virtual assistant which allowed him to accomplish a large amount of work faster. “You can get a lot of the work done that was previously impossible,” he said.
BrainBox said that ARIA gives facility managers a 360o view of the data in a building, focusing on its specific systems and components for accurate and holistically informed strategic decision-making.
ARIA’s design incorporates two-way interactions, according to the company. The company continued that facility managers and building owners are not only encouraged to complete specific tasks, they are also informed about future action which could lead to a more efficient and effective building management.
ARIA boasted that its clients could “call” ARIA by text or voice, and take their interactions seamlessly from desktops to mobiles without missing a beat.
What’s more, it added, ARIA’s generative AI engine works 24/7 to help clients prioritize and optimize their buildings — a feature that transforms building management from reactive to proactive, increases its value and directly contributes to an organization’s sustainability efforts.
The video below shows ARIA in action, showing its ability to improve efficiency, reduce energy cost, and improve operational workflows within commercial and retail spaces.
Avoiding AI hallucinations
One of the most common concerns about AI tools that generate answers is that they may “hallucinate”, or give answers that are good-sounding but inaccurate or even bizarre. This behavior can be explained by a number of factors.
This means that while LLMs are very good at mimicking the language patterns they encounter, they may not be able to understand the full meaning of the text. The result can be a text that’s grammatically perfect but is factually inaccurate or nonsensical.
Hallucinations can be influenced by the way an LLM selects words. Some techniques emphasize fluency rather than accuracy, leading to unreal but creative outputs.
In addition, LLMs receive training based on huge amounts of data scraped directly from the Internet. This data may contain errors in fact, biases and other strange things. The model can identify these patterns and generate outputs reflecting them, even though they may not be accurate.
BrainBox solves the problem of hallucinations by restricting how and what ARIA uses. Venne explained that Gen AI was an empty bucket. You need to plug it in a data sandbox. It can then create something exciting.”
“We’re going to plug it into our existing stack of tech, which includes HVAC information and other types for buildings,” he explained. “We built ARIA atop the stack so that it could access this ocean of information, and analyze data trends.”
“The sandbox of AI is limited by our data set,” said he. “You could ask it, ‘Why did Napoleon lose the battle of Waterloo?’ but the response would be something like ‘Can we talk about your building portfolio? I’m a non-historian.
Built on Bedrock
Venne said that the translation of data has been a major challenge for BrainBox in the six years it has developed its AI-powered solutions.
He said: “It is rare to get perfect data when you plug into these systems.” You can get false readings from sensors. You can have gaps — a thermostat not giving you any readings for an hour. We had to be capable of extracting clean data and making corrections to the data when they are faulty.”
“Once a set of super-structured, clean data is available, tools such as generative AI are much easier to use,” he added. “We only trained our AI on data we had full control over. We have validated the data and are confident that it’s accurate.
AWS Bedrock is a key component of ARIA. Venne said, “Bedrock lets us stay flexible without committing to a technology that we might be stuck with in six-months.”
“Amazon Bedrock makes it easier for applications to leverage high-performing foundation models from leading AI companies like Anthropic, Meta, Mistral, and others, via a single API call — all in one secure, fully-managed service,” said AWS Vice President and Global Head of Startups Howard Wright.
Wright, speaking to TechNewsWorld, said that Amazon Bedrock allows applications such as ARIA to become the most intelligent and sophisticated building manager assistants possible by leveraging insights and selecting models most suitable for a particular use case or task.
He explained that Amazon Bedrock makes it easy to build and scale generative AI applications using foundation models.
“Startups can quickly experiment with and evaluate top foundation models for specific use cases — from Claude 3 to Llama 2,” he noted. “From there, the startup can private customize them using their own data by using techniques such retrieval-augmented generation and fine-tuning. [RAG] “Build agents that perform tasks using the startup enterprise systems and data sources.”
“With buildings being responsible for almost 40% of greenhouse gas emissions globally and 27% of those emissions coming from the energy used to heat, cool, and power them, and the global stock of buildings expected to double by 2050, the challenge of heightening building sustainability isn’t just necessary — it’s urgent,” Wright added.
BrainBox AI solves this problem by using AI-driven technologies trained on Amazon Bedrock. This allows them to rethink their energy systems and minimize the effects of the climate crises.
