How AI is changing Amazon PPC advertising

Rick Wong, Co-Founder of Amazon seller agency, SellerMetrics, unpacks how AI is changing Amazon PPC advertising -- and how seller can benefits

Artificial intelligence (AI) is unravelling established business processes at an unprecedented pace. Not surprisingly, Amazon advertising is not exempt from the tidal wave of new AI tools. One constant when implementing AI solutions is the staggering mix of surgical precision and mind-blowing failures.

If you have tested Amazon AI optimisation pay-per-click (PPC) tools, chances are you've experienced both moments of jaw-dropping performance and instances where you couldn't fathom how such a sophisticated tool could make such basic (and costly) mistakes.

In this article, Rick Wong, founder of the Amazon seller agency SellerMetrics and a former Amazon seller himself, provides a guide to AI adoption for Amazon PPC optimisation and shares his take on when AI can be trusted to take over, and in which instances human oversight is not just a nice-to-have, but an absolute necessity.

Scenarios where AI can already boost Amazon ad performance

• Creative production 

In September 2025, Amazon launched AI creative workflows via a new agentic tool embedded within its Creative Studio. The intuitive creative suite enables brands to turn dull, dreary assets into high-performing videos or images in just a few minutes.

While hiring a professional designer or orchestrating a photo shoot will yield better results, the Amazon Creative Studio delivers solid, on-brand results fast and at no cost. Rather than investing excessive time and resources, sellers can shorten time-to-market and deploy effective assets right away.

Why this matters: The product and lifestyle images used in your listings are key to improving click-through rates. As Wong points out: "Besides optimising listing titles, enhancing product images is the single most effective way to drive more traffic to your listings. In return, this also helps keep your cost per click (CPC) on Amazon ads in check."

• Conversational data analysis

While many Amazon operators are true Excel wizards, crunching numbers isn't everyone's favourite activity. New Amazon tools, such as the Amazon Seller Assistant, powered by Amazon's Bedrock generative AI, make it easier to discover accurate insights in data.

Analysing Amazon Search Term Reports for Sponsored Products campaigns is a great use case. Rather than sifting through troves of keywords, sellers can "talk to their data" and ask questions, such as: "Which search terms are surging?" or “Which search terms are we spending budget on but aren't generating any sales?"

Large learning models (LLMs)  already perform well on language-related tasks: spotting irrelevant search terms, grouping queries by user intent, drafting lists of negative keyword candidates, and tweaking ad copy based on seasonal changes are all tasks that existing AI solutions handle well today. Integrating this functionality natively into Amazon Seller Central will reduce barriers and boost adoption. Especially when paired with custom ad-optimisation rules, sellers can expect better, more predictable ad performance without risking control.

While Amazon Seller Assistant is still in beta and performance can be shaky at times, it will likely evolve into a powerful tool that surfaces actionable findings for Amazon brands.

Why this matters: According to Wong: "The Amazon Seller Assistant will likely not be a game changer for the most sophisticated Amazon brands, but it will level the playing field by providing smaller sellers with an effective tool to fine-tune their Amazon PPC campaigns."

Areas where AI (still) falls short

While the benefits of Amazon's new AI tools are apparent, it's also important to note their limitations.

Numerical analysis: Sellers may be tempted to use AI to analyse Amazon advertising reports. As Wong notes: "Amazon advertising reports contain thousands of rows of data. Spread across dozens of columns, LLMs can get overwhelmed. While we've tested AI systems to analyse quantitative advertising data, the results were disappointing. Simple errors were the norm rather than the exception. We can only warn sellers not to go down this path at this time."

Identifying outlier events: A marketplace like Amazon is rife with them. Sales spikes during Cyber Week or sudden demand shifts in weather-dependent categories can profoundly confuse AI models. An AI-powered advertising platform optimising toward a strict 20% ACOS target (Amazon's “advertising cost of sales," i.e., its ad-spend ROI metric) may completely miss out on outsized sales opportunities during a sudden demand spike.

On-brand assets: While Amazon's Creative Studio can create somewhat consistent creative assets, demanding brand marketers will likely argue that it's still not good enough. In practice, expect some AI creative variants to be off-brand or squishy. Ultimately, it comes down to how you view the trade-off between execution speed, consistency, and production costs.

Amazon’s AI roadmap for 2026

Wong believes it's likely that Amazon will deploy more tools to integrate seller operations even further—from product identification to sourcing and marketing. As he argues, Amazon's success is ultimately tied to its sellers' and, with shopping trends evolving faster than ever (spurred by the TikTok-ization of digital commerce), Amazon cannot afford to let its seller base fall behind.

Getting trending products onto the marketplace is a sprint, not a marathon, and even missing out by a few days can cost Amazon millions of dollars in foregone sales.

Regarding Amazon Advertising, Wong predicts the emergence of a Google-style Performance Max campaign. This automated campaign type—which autonomously invests across Google-owned channels such as Google Search, Google Shopping, YouTube, and the Google Display Network—has proven effective for direct-to-consumer (DTC) e-commerce brands.

Likewise, Meta has seen positive results with its Advantage+ autopilot campaign format. It's reasonable to expect that Amazon will develop a similar product to bridge its currently distinct ad formats: Sponsored Products, Sponsored Brands, and Sponsored Display.

In the future, Amazon will likely offer an option to auto-assemble multiformat advertising campaigns based on a seller's product catalogue and assets.

With some limitations, PPC management is AI-ready

As this article outlines, AI is not a holy grail that will turn underperforming Amazon campaigns around in a heartbeat. For that, expert advice and meticulous optimisation are still required.

But, especially for producing creative assets and mining search terms, AI-based solutions can already support Amazon ad operations in a meaningful way today.

With a solid playbook in place, Amazon sellers can leverage AI to move fast, capture new trends before competitors do, and ultimately gain an advantage in the highly competitive marketplace.