AI and the Future of Search

Forum Takeaways

 
 

Longtail UX brought together industry leaders, experts, and innovators for an insightful online panel discussion on AI and the Future of Search. The event featured a distinguished panel, including keynote speaker Scott Thomson (Head of Innovation, Google Cloud ANZ), Daniel Hogben (Head of SEO for APAC, Overdose.), Richard Bedford (Head of Client Growth, Longtail UX), and Andreas Dzumla (Founder, Longtail UX). The forum attracted an exceptional audience representing leading Australian brands such as Coles, Woolworths, Temple & Webster, Myer, The Good Guys, Officeworks, Adore Beauty, Kmart, Super Retail Group, Dan Murphy’s and others.

Key Insights from the Forum

Keynote – Scott Thomson, Head of Innovation, Google Cloud ANZ.

Scott explored AI’s profound influence on eCommerce. While it’s certainly not the first time that great expectations are directed at AI – he referenced a The New York Times title cover about the “great AI awakening” from 2016 in relation to Google Translate – he highlighted many impactful near-term AI opportunities for eCommerce businesses with generative AI technology, both image and text-based. These opportunities include integrated chatbots for customer service, video search capabilities, video language dubbing to expand a brand’s reach to new audiences, and AI-generated contextual product images for impactful advertising at scale.

Google takes an “AI-first” approach. An AI-first approach in the context of  eCommerce businesses can translate to using AI to enhance planning, reduce manual tasks, and redeploy talent for higher-value pursuits. This approach goes beyond cost-saving, instead re-investing resources in growth and innovation.

The keynote offered a framework for identifying generative AI use cases within eCommerce. Assessing potential customer value against business risk allows businesses to identify those opportunities that fall into the “high value – low risk” quadrant, providing examples across various domains, specifically:

  • Merchandising and marketing: Enhancements of product attributes, Product Benchmarking, Marketing content
  • Commerce & customer service: Conversational commerce, Recommendations and beauty plans, Contact centre, help desks
  • Support & shared services: Knowledge discovery, Virtual scribe/assistant, Coding


With regards to risk, what eCommerce leaders and boards should care about are data security, not leaking; Safe, factual model outputs; Explainability and lineage; IP boundary and IP control; and regulatory integration.

The Panellists:

Scott Thomson
Head of Innovation, Customer Engineering, Google Cloud

Scott has worked at Google since 2015 in the Ads Data Platforms, Analytics and Google Cloud product areas in AuNZ and APAC. Before Google, Scott worked with Adobe across APAC on digital strategy and digital transformation. He also worked for Telstra and Sensis in digital transformation and held the role of CTO in an Australian startup focused on personalised video advertising for mobiles. He is passionate about working with and helping startups throughout AU and NZ and still dreams of running away to join another startup circus.

Daniel Hogben
Head of SEO – APAC, Overdose.

Daniel is an SEO Specialist with expertise in developing high ROI strategies for local & international e-commerce clients, with a passion for content marketing & data analysis. He continues to find opportunities for his clients to grow and reach their maximum potential.

Richard Bedford
Head of Client Growth, Longtail UX

A digital leader with 20 years experience in SEO, data, tech and analytics, Rich has worked in-house, across boutique and global media agencies, and has helped hundreds of clients achieve rankings and ROI from search engines around the world.

Andreas Dzumla
Founder, Longtail UX

Andreas is the inventor of the patented customer acquisition platform Longtail UX, that empowers marketers to leverage their website and the power of longtail search to acquire new customers and penetrate new categories faster. He is an ex-Googler and Dentsu Aegis agency GM with 20 years of experience on platform, agency and client-side in APAC, the US and Europe bringing people and technology together to deliver sustained growth and innovation.

Q&A Topics:

AI can be expected to fundamentally change how users search, moving towards a more conversational style. Businesses will need to adapt their content to multiple formats like text, images, and videos to win in the conversational search landscape. While search features like ‘Knowledge Panels’ and ‘People Also Ask’ have already increased so-called ‘zero-click searchs’ for top-of-funnel information searches, generative AI can be expected to increase this trend. This may or may not lead to a decrease in traffic numbers for informational searches – on the other hand, this can be expected to further increase the quality of website traffic from search, leading to higher conversion rates, with users self-pre-filtering with more specific searches on generative AI search.

User expectations can be expected to heighten, with users receiving more tailored search results and expecting a similar experience on the click-through to websites. This emphasizes the importance of landing page relevance. Metrics like click rates and dwell time in comparison to competitors in the same industry and for the same search query might become even more critical for ranking. On the other hand, AI provides opportunities to bring better personalization and content freshness, requiring businesses to have up-to-date inventory and audience data.

Google’s webmaster guidelines with regard to AI-generated content have been updated this year, mentioning AI-generated content as accepted. We can expect Google to evaluate AI-generated content with the same criteria as it has been evaluating human-generated content: on the basis of its relevance and usefulness. To ensure high-quality content, human touch and expertise will continue to be crucial for higher rankings, whether AI plays a role in the actual creation of content or not. Generative AI can be particularly useful for e-commerce, helping to write product descriptions and offering personalization; meanwhile, a well-optimized merchant centre feed will become even more important for paid and organic product listings.

Chatbots are unlikely to replace traditional SERPs but will add another layer to enrich the user experience, enhancing engagement. This augmentation also extends to organic search results, which could transition into reference links, potentially leading to slightly lower click-through rates. However, the upside lies in the cultivation of more qualified and high-value traffic, indicating the dual benefits of chatbots and SERPs.

Companies must ensure their AI models are free of third-party IP violations and provide indemnities in their terms of service. In regards to any 3rd party IP used in training our AI, Google takes best effort to keep training data free and clear of IP, with tools like citation checkers to block known sources of IP. Google Cloud provides indemnity in terms of service for customers on any third-party claims arising from its training data.

There are also a lot of ethical issues with AI in search, and because Google has worked on this for so long, they have given a lot of thought about the responsible way to do this. They look at toxicity in content and content bias, both from the training data set and output and ensure this is reviewed regularly and best efforts are continuously applied.

Conclusion:

While all of the exact ways in which Generative AI will change Search and Ecommerce is yet unknown, we know it will. So, as a business, you want to be prepared. The best way to prepare is to test generative AI – and there are plenty of options and great tools for eCommerce, including the ones that Google presented from Google Cloud.

From a business perspective, it is important to create the right frameworks to manage security, risk, IP implications and the right environments to encourage teams to use generative AI within the frameworks provided. In his keynote, Scott mentioned governance and security aspects; we also recommend Scott’s grid of decision criteria for high-impact areas for AI in eCommerce businesses. Rich and Dan mentioned a lot of great applications: Focus on technologies like Python, Databricks, AWS, and Salesforce for offer recommendation engines; Cleanrooms such as Snowflake, Google Ads Data Hub, and Infosum will become essential for data regulation, privacy, and sovereignty.

Generative AI can be expected to further increase user expectations on the speed of access to information: providing the exact matching products to the search query and high-quality content in text (e.g. rich product descriptions), image and video format. So, supporting information, personalisation, and generally high landing page relevance and quality on the website will become even more important than it already is.

Related Articles