Lauren Walker, our Chief Client Solutions Officer, recently had the pleasure of joining Deepak Shukla, Senior Data & AI Sales Specialist at AWS, at the Marketing Leaders Summit in London to discuss all things AI. Here, they continue that conversation.
Lauren Walker: What tradeoffs, if any, should advertisers be considering as AI innovation heats up? When it comes to leveraging AI responsibly and ethically (in the context of consumer privacy), what best practices should advertisers keep in mind?
Deepak Shukla: Generative AI helps facilitate highly natural conversational experiences between consumers and organisations, and we’ve seen some great success stories from Amazon Web Services (AWS) customers using generative AI to reimagine how they engage with their own customers. However, it’s clear that generative AI tools need appropriate guardrail mechanisms. It is important that advertisers and marketers consider implementing guardrails for consumer applications, like a chatbot, to prevent incorrect information from being shared or conversations veering off-topic into open-ended and potentially controversial subjects that are irrelevant to a company’s brand. Advertisers and marketers should also have the ability to customize and apply safety, privacy and truthfulness protections, such as the ability to redact sensitive information like personally identifiable information (PII) to protect consumers’ privacy. Additionally, strong security measures should be in place to prevent malicious actors from attempting to manipulate AI applications into sharing harmful outputs. I understand that as you build Omni, which includes generative AI agents, you are taking these same steps.
LW: With your extensive background leading digital transformation initiatives across industries like financial services and retail, which AWS innovations do you feel will drive greater agility and efficiency for large-scale institutions?
DS: AWS holds a unique position to help advertisers and marketers create campaigns quickly and responsibly, deliver hyper-personalized customer experiences, and transform organizational productivity through our generative AI tools. A few innovations that are helping our customers drive greater agility and efficiency are Amazon Bedrock, Amazon Nova foundation models, and AWS Clean Rooms.
For example, at AWS re:Invent 2024, Amazon introduced Amazon Nova, a new generation of foundation models (FMs) that have state-of-the-art intelligence across a wide range of tasks, and industry-leading price performance. Several models will be available in Amazon Bedrock, including Amazon Nova Micro (a very fast, text-to-text model), Amazon Nova Lite, Amazon Nova Pro, and Amazon Nova Premier (multi-modal models that can process text, images, and videos to generate text). Amazon also launched two additional models – Amazon Nova Canvas (generates studio-quality images) and Amazon Nova Reel (generates studio-quality videos).
Finally, the ability to securely collaborate with partners, and without sharing raw data, is increasingly top-of-mind for advertisers and marketers. Many companies use AWS Clean Rooms, which helps them and their partners to more easily analyse and collaborate on their collective datasets in a secure manner — all without sharing or copying one another’s underlying data. With AWS Clean Rooms, you can create a secure data clean room in minutes and collaborate with any other company to generate unique insights about advertising campaigns, investment decisions, and research and development. Annalect actually worked closely with us on this by informing our clean room capability. And now, Annalect is working closely with colleagues on the Amazon Ads side of our business to drive capability development around Amazon Marketing Cloud.
LW: Having worked with clients across diverse sectors — financial services, consumer goods, and retail — what are the common challenges and opportunities you’ve seen arise when it comes to adopting AI at scale?
DS: While 2023 was the year of learning what’s possible with generative AI, late 2024 was all about moving to production and scaling. As our customers are going through this journey, they need to work backwards from production. Doing so will ensure that the proof-of-concept they’re working on is successful, and any potential blockers are identified and mitigated upfront rather than late in the process. When it comes to selecting LLMs, one size doesn’t fit all. It’s important that organisations choose the model that best fits their own distinct needs. Amazon Bedrock Marketplace is a new capability in Amazon Bedrock that enables developers to discover, test, and use over 100 popular, emerging, and specialized foundation models (FMs) alongside the current selection of industry-leading models in Amazon Bedrock. It is one offering meant to help customers explore more models in order to find the right model for their specific needs.
LW: In regard to AI and generative AI innovation, what do agencies need to be prioritizing in the months ahead?
DS: As we continue to see new innovations in this space, I recommend advertisers and marketers prioritise three things. First, prioritise high value use-cases for generative AI and identify how it will yield long-term benefits. (Annalect embedding agentic intelligence into Omni is a great example of this.) Second, ensure efforts are being made to improve data foundations and introduce modern data strategies to support AI adoption and scaling. Third, have a training plan in place for business users to become more AI-savvy in their activities.
LW: Can you share examples of how you use AI and/or generative AI, both personally and professionally?
DS: On a personal level, I’ve started using Amazon Rufus a lot while shopping on Amazon. I recently bought sports shoes for my 9-year-old son and was impressed by how well Rufus guided me through the shopping process. Professionally, I sometimes use AI when preparing for a customer meeting. I use Amazon Q Business to make sure I have our latest AWS content before I engage with our customers.