Rubber hits the road on AI as companies move on from experimentation
Amazon Web Services believes generative AI is the most transformational technology since the cloud, with IDC predicting local spending will reach almost half a billion by the end of the year.
Amazon Web Services believes generative artificial intelligence is the most transformational technology since the cloud, with the potential to impact every industry worldwide.
IDC predicts generative AI spend in Australia to reach almost half a billion by the end of 2024.
Today, we are seeing generative AI deliver significant benefits in customer experience, software engineering, and research and development. As organisations continue to build confidence in deploying generative AI workloads, we expect to see more and more use
cases being rolled-out.
That’s why we have just announced Amazon Bedrock, our managed service that offers a choice of AI foundation models through a single API, is now available in the AWS Sydney region. It will help customers in the public sector and highly regulated industries innovate with generative AI and provide more choice on where they can run and store generative AI applications.
Apart from consumer-facing implementations, organisations are also seeing benefits from generative AI when it comes to driving employee productivity and enhancing worker and customer experiences. This includes call centres that use generative AI to summarise call transcripts and flag urgent cases to empower agents to respond to calls more efficiently and quickly.
Aussie online beauty retailer Adore Beauty is using AWS generative AI services to automatically create product descriptions, allowing their catalogue team to easily scale while keeping their site content fresh to enhance customer satisfaction.
Our customers are also finding value in using generative AI coding companion tools like Amazon CodeWhisperer to accelerate software development. One major Aussie bank customer has already improved its developer productivity by over 15 per cent using Amazon
CodeWhisperer to provide real-time code recommendations when writing software. This figure is expected to increase as thousands of its developers become more familiar with our generative AI tool. The bank’s developers will also retain full control over the final code quality and security, while automating code documentation.
As more organisations move from experimenting with generative AI to fully deploying it this year, our focus remains on helping businesses “productionise” their proofs-of-concepts.
Organisations may struggle to evolve their PoCs into production-ready solutions. Despite having numerous POCs, they lack the necessary structure, data strategy, cloud security and governance, and a skilled workforce to effectively measure the productivity gains, or a return on investment.
The companies that do this successfully have a thoughtful data management strategy, implement responsible use of technology – such as content moderation to detect biases in data, while also balancing human and AI judgement. They also select the specific foundation models that suit their needs, all while managing security and privacy implications, and often with the support of technology partners.
However, the interest is there. According to IDC, 75 per cent of organisations in Australia have established spending plans for generative AI in the next 18 months. A third of these organisations are already testing models and have a PoC in place.
Anthropic – one of Amazon Bedrock’s foundation model providers – recently announced Claude 3 which includes three models: Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus.
Data from Anthropic shows that Claude 3 Opus, the most intelligent of the model family, has set a new standard, outperforming other models available today — including OpenAI’s GPT-4 — in the areas of reasoning, maths, and coding. All three models are now available on Amazon Bedrock.
Claude 3 Haiku is already getting a lot of attention from our customers. It’s designed for near-instant responsiveness and seamless generative AI experiences that mimic human interactions. Potential use cases could be engaging customer interactions with quick and
accurate support in live chats, or fast knowledge extraction from unstructured data. Swann, a global leader in security devices, switched from Claude 2.1 to Claude 3 Haiku in Amazon Bedrock in only a couple of hours, and the speed of responses from its Swann HomeShield service is 30 per cent faster and significantly more accurate.
At Amazon, we like to say every day is Day 1, and our approach to innovation emphasises customer understanding – starting with their pain points and developing impactful solutions for them by working backwards. Day 1 thinking encourages curiosity, agility, experimentation, and embracing failures as learning opportunities. We encourage Aussie businesses to adopt similar principles when experimenting and integrating generative AI in 2024.
Louise Stigwood is director of enterprise at AWS Australia and New Zealand.
More Coverage
Dick Smith praises crackdown on scams
The outspoken businessman has welcomed new legislation that will impose fines of up to $50m on social media companies, banks and telcos that fail to actively detect and prevent scams.
Apple TV+ finally comes to Android phones
Apple has finally released its streaming platform on rival Android devices, and is offering users a seven-day free trial amid pressure to open up its tightly-controlled systems.