Introducing Atmosphere

Stephen Hardy
Tuesday, Jun 16, 2020

Here we introduce Atmosphere – Ambiata’s new continuous intelligence service – and explain how it fits into our renewed values and purpose.

Founded in 2012, Ambiata are one of the early pioneers of ML and AI in the Australian market. By successfully combining our core strength in data science from CSIRO’s Data61 with commercial acumen, we have partnered with many large Australian enterprises in the Financial Services, Retail, Telco and Public sectors to embed ML and experimentation into their business processes.

What we have learnt

Over the last 8 years, we have witnessed the evolution of the data landscape and its effect on local organisations:

Data has been recognised as a core asset - companies are seeing the need to develop in-house capability to manage and exploit their own data. It is no longer considered strategic to send data to third parties, as the value derived from understanding that data must be internalised.
An agile culture is required to adapt to changing technology - the constant release of new data processing technologies, the adoption of cloud, and now the trend towards hybrid- and multi-cloud have led to new architectures and approaches. Adapting to this landscape requires an agile culture, a willingness to reinvent, and the capability to deploy new approaches quickly.
Experimentation is not optional - digital natives continue to lift the bar, disrupting traditional business models through rapid digital experimentation at-scale, pushing the boundaries on personalisation for their customers, which in turn is increasing customer expectations from other service providers.
Despite all this, activating internal data to generate returns remains difficult - even after massive investment in data infrastructure, putting that data to work effectively is still limited by insufficient internal capability, and the coordination of data-driven decisioning between disparate groups.
Finally, there is an increasing need for governance - businesses must ensure that they manage the risks inherent in deploying ML and AI, including bias, fairness, transparency and privacy. The need for data and model governance for companies using AI, as well as external factors such as the emerging Australian government AI ethics framework and strengthened privacy requirements, have brought these topics front-of-mind for organisations operating at scale.

During the same period, we developed new ways to activate data and deploy ML and AI at scale. We have also demonstrated how to apply the scientific method in business to derive knowledge about customer decisioning, and to discover what drives effective data science outcomes to the business.

Based on what we have learnt from our customers and partners over the last 8 years, and what we have discovered and developed ourselves, Ambiata has been busy for the last 12 months developing and deploying a new solution that brings a leading-edge data science capability into the Australian data ecosystem.

We have taken the capabilities inherent in our multi-tenanted ML platform, combined it with our rigorous experimentation framework, and deployed it using a modern DevSecOps architecture to make an on-prem or your-cloud AI solution.

We call this “Atmosphere”.

Atmosphere is already deployed in production driving better customer outcomes in our parent company, IAG. IAG is Australia’s largest general insurer, with brands including NRMA Insurance, CGU and WFI.

Atmosphere embodies our purpose and our values.

Our purpose

Ambiata’s purpose is to help organisations activate their internal data for better customer outcomes.

In setting this purpose, we also want to remain true to our strong track record of successfully combining deep scientific expertise alongside commercial focus, ensuring that our clients generate great returns on their data investments. Today, we call this our ‘Scientific Return-on-Data’.

Our values

Our rich history alongside our market knowledge has been encapsulated in our values:

Scientifically Better - we embed ML and AI inside a rigorous scientific framework to identify cause-and-effect relationships to drive true customer understanding.
Validated & Scalable - we deploy our technology in replicable, scalable, and secure architectures. Our relationship with Insurance Australia Group has allowed us to demonstrate this on Australia’s largest general insurance data set.
Safer is Smarter - we believe in explicitly optimising customer outcomes and deploying ethical AI systems. This is in our DNA - our co-founder, Dr. Tiberio Caetano helped establish the Gradient Institute, a non-for-profit research institute we partner closely with, whose purpose is to progress the research, design, development and adoption of ethical AI systems.

Introducing Atmosphere, our continuous intelligence service

Atmosphere was created to address the key pain points that organisations currently face in deploying and maintaining automated decision making systems.

By choosing Atmosphere, our clients will be able to run closed-loop ML and AI experiments which activate existing data, learning from customer behaviours quickly, efficiently and safely to drive better customer outcomes. This “continuous intelligence” approach leads to quicker outcomes that not only offer a better return-on-data, but also helps businesses become smarter, creating more valuable data assets over time.

Some key features of our continuous intelligence solution include:

Data scientists can deploy custom ML models into production using a modern DevSecOps deployment pattern;
The real business and customer impacts of deployed algorithms are measured and optimised through our rigorous experimental framework;
Our user interface helps break down silos within the organisation to aid collaboration and drive the creation of intelligent digital services;
An in-built capability using state-of-the-art ML for efficient continuous learning from both online and internally held data;
Deployment on client premises or in their cloud to ensure they maintain control of their data and processes.

Alongside Atmosphere, we also have a team of data scientists who are experts in combining ML and experimentation. We understand that integrating ML and AI into our clients’ business is a journey, and we are able to provide support each step of the way, and help upskill internal teams in ML, AI and experimentation.

Where we are today and what’s next

Atmosphere represents the culmination of all our previous experience and lessons learnt (often hard-earned!) whilst implementing ML solutions at-scale, from our people past and present.

In launching this new and enhanced version of our continuous intelligence solution, we have built upon the experimental techniques that Ambiata has developed and refined over many years.

Excitingly, we have already deployed Atmosphere in a real-world environment with our partners at IAG where it is busy putting personalisation into production.