$1 Party, page-2

Alot of maybes, potentially, and likelys

Fact -

Strategy refresh – long term growth and revenue diversification Appen plays a key role in powering both deep learning applications and generative AI. Today, most of the company’s revenue comes from deep learning projects, specifically in model testing and relevance work. This half, Appen re-positioned itself to take advantage of its full growth potential and to capture all available opportunities in the generative AI space. Appen has a comparative advantage over its competitors by using its 28 years of experience in crowd-based data collection, annotation and model evaluation and adapt this expertise to generative AI. Appen’s growth vision captures a full set of AI data services for deep learning and generative AI as illustrated in diagram 1.

Our growth vision is to move beyond AI data into industry vertical AI solutions and expand our addressable market, which will help to diversify our customer base, deliver margin improvement, increase revenue visibility and deliver strong return on investment. Generative AI will create a significant growth opportunity for Appen, and with the market expected to grow from $8 billion in 2021 to $111 billion in 20301 . Generative AI is expected to be driven by adoption by both technology and non-technology companies, presenting Appen with a significant opportunity to diversify its revenue as enterprise companies develop and adopt generative AI. Generative AI will require human feedback in the loop for base models to perform at levels comparable to humans and to minimise hallucinations, bias and toxicity and many base models will not work out-ofthe-box and fine tuning and evaluation is required as well as real customer interactions. In addition, new approaches are needed for model evaluation particularly when measuring against risk and compliance guidelines. In 1H FY23 Appen launched a new set of Large Language Models (LLMs) products that provide a comprehensive data platform for enterprise customers to generate meaningful value when they build and integrate generative AI models in their organisations, and include: - Fine tuning – allows models to understand and respond to their intended context, whilst limiting their ability to respond outside the intended context, allowing enterprises to be able to customise their models from Nvidia, OpenAI, Cohere, Google, Anthropic, Reka and open-sourced channels. - Assurance – models must be continually monitored to ensure that performance is maintained and to identify and correct any negative customer experiences, mitigating the risk of hallucinations, bias and toxicity. 1 Acumen Research (December 2022)
Appen is also focused on increasing its channels to market and is collaborating with Nvidia, Reka and other professional service providers. This provides important access to enterprise customers and will enable Appen to reach more clients. Appen has made strong early progress and has multiple projects underway that relate to generative AI model development and evaluation, for both large tech and enterprise customers.

People are expecting miracles when it had a strategy refresh in May - next May will be full year of strategic change.

No matter what one believes Appen will continue to evolve because it does have a front row seat to A.I developments.