Amazon's internal sales guidelines provide talking points to dispel hype about OpenAI and answer customer questions about competing AI products from Microsoft and Google.
The guidelines, obtained by Business Insider, also list nine key factors that customers should consider before purchasing generative AI models or services.
The playbook, used primarily by AWS salespeople, gives a hint at the company's AI priorities: The nine criteria include security, cost and the ability to personalize models using a technique called search augmentation generation (RAG).
Amazon Web Services CEO Matt Gurman. Amazon; Insider
The nine factors are:
Customization: The ability to tailor an AI model to specific requirements (e.g., the style of the model's output). Personalization: The ability to use in-house data to make the output of an AI model more relevant and customized (i.e. fine-tuning or RAG). Accuracy: How close the generated output is to the desired objective. Security: Implementation of appropriate measures to protect data and privacy. Monitoring: The ability to identify issues such as drift, bias, or degradation in output quality. Cost: Overall expense, including the initial investment and ongoing costs related to training, deployment, maintenance, and infrastructure. Usability: The ease of use of the model, integration capabilities, and availability of support. Responsible AI: The ability to adhere to ethical guidelines for models, address bias, provide explanations for output, and build in safeguards against misuse. Innovation: The service provider's perceived status as an innovator compared to other providers.
The guidelines instruct AWS salespeople to focus not too much on the popularity of AI chatbots, but on the underlying models and cloud infrastructure required to build AI services, as AWS's strength lies in cloud infrastructure, not consumer AI chatbots. However, the company is working on a competing product to ChatGPT.
One of the guideline documents mentions AWS' “value proposition” to leverage when selling to potential customers, including the ease of building and customizing AWS' AI services, as well as robust security and privacy features. The document also mentions AWS' “cost-effective infrastructure,” including its own AI chips and AWS-built AI applications such as Amazon Q.
“It's no secret that generative AI is an extremely competitive space, but AWS is a leader in the cloud, and customer adoption of our AI innovations is a major driver of our continued growth,” an AWS spokesperson told BI in an email.
“AWS offers more generative AI services than any other cloud provider, which is why we generate billions of dollars in revenue from our AI services alone,” the spokesperson added. “Generative AI is still in its early stages, and with so many companies offering different services, we work hard to equip our sales team members with the information they need and help customers understand why AWS is the best, easiest, and best-performing place to build generative AI applications. To interpret this as anything more or to misrepresent our leadership position is an incorrect assumption.”
Do you work at Amazon? Have any tips?
Reporter Eugene Kim can be reached on encrypted messaging apps Signal or Telegram at +1-650-942-3061, or by email at [email protected]. Please contact him from a non-work device. For more tips on sharing information safely, see Business Insider's Sources Guide.
