The slow march of AI in commerce: Why marketers should focus on what’s in sight

A recent podcast featuring an AI safety advocate discussed the race to AGI and its existential risks. The advocate predicted transformational changes within just a few years – something echoed by many marketers when talking about commerce. For example, many believe years of development will be compressed into urgent forecasts, and Google Search will cease – replaced instead with fully automated shopping run by AI agents.

It’s interesting when you look at this through the lens of Amara’s law: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” It’s something many in marketing and commerce will be able to relate to, as gains tend to be made when brands focus on what is actually in sight. Customers are buying today, next week and next month. How they might buy in 2030 is a strategic question, which sits apart from the immediate work.

The persistence of shopping habits

Shopping behaviour tends to evolve slowly and often takes two steps forward, one step back. The in‑store retail bounce‑back after COVID made that plain. Many DTC brands have struggled to build sustainable models. Nike’s public reset on its DTC emphasis shows how hard this path can be. Meanwhile much hyped social shopping has grown in Europe and the United States, but growth has been modest for larger, established brands.

Where we actually stand

AI adoption looks fast when set against earlier technology cycles. That impression can mislead. Many people are already online and most tools are free, unlike those in prior cycles (PC, smartphone, broadband internet etc.). AI has started from a strong base for early growth.

Even with free products, audience data shows plenty of headroom. According to GWI, 50.1% of UK adult internet users had not used an AI‑specific tool (e.g. Chat GPT) in the past month. In contrast, Capgemini reports that 71% of consumers want generative AI in shopping journeys. Interest is clear. The gap between interest and regular use shows where we really sit on the adoption curve.

Current AI commerce features show both promise and limits. Shoppers can already use:

• Amazon to skim AI generated review highlights that condense thousands of comments into a short summary. Amazon also offers Lens Live, where a shopper points a camera at an item and sees similar products. Rufus sits in the flow for follow‑up questions.
• Walmart’s Sparky to compare options, create lists, and auto‑fill a basket to a brief such as furnishing a flat within a set budget and style.
• Instacart’s Ask Instacart to turn open questions like “what can I cook for four in 30 minutes” into a shoppable plan.
• Zalando Assistant to give style advice in local languages that links straight to available items.
• IKEA Kreativ to place products into a lifelike model of a shopper’s own room and then buy the look.
• Boots No7 shade matching to pick foundation online with AI and AR analysis that reduces trial‑and‑error.
• Wayfair Decorify to re‑style a room from a single photo and shop the generated designs.
• Google’s AI Mode for Shopping to ask conversational queries and see curated product panels drawn from a large Shopping Graph.

These tools help discovery, sizing, confidence and comparison. The act of shopping still feels familiar.  But there is more guidance on the path.

The reality of change

Rapid behaviour change needs strong incentives on both sides of the market. Consumers still care most about convenience, trust, and price from retailers. Retailers need clear commercial benefits from AI features to invest at scale. They’re still competing for the same demand. AI referrals aren’t magical from a retailer’s perspective; if anything, they risk losing direct customer relationships to AI platforms that monetise through advertising and affiliates.

Demographics matter. GWI reports that 74.7% of Gen Z used an AI‑specific tool in the last month, while only 28.7% of Baby Boomers did so. This usage points to a pattern driven by demographic shifts over time. Uptake is unlikely to be universal all at once.

A pragmatic path forward

Amara’s law has two parts. The near term calls for restraint and clear‑eyed testing. The long term points to deeper change. Commerce will transform in profound ways given enough time.

For marketers in commerce today, the task is balance. Test AI where it solves real customer problems. Experiment with personalisation, chatbots and recommendation engines. Recognise that customers are shopping today with current behaviours and expectations. The infrastructure, trust frameworks and incentives for large‑scale AI commerce are still being built.

Focus on value through the channels and methods that work now. Stay informed about new possibilities and be ready to move when they prove out. The AI shift in commerce is coming. But it is arriving more slowly and across a longer window than headlines claim. Teams that balance experimentation with pragmatism will serve today’s customers and prepare for tomorrow’s opportunities.

Frequently Asked Questions

Q: What is the main idea behind the “slow march of AI in commerce”?
A: The main idea is that while AI feels like it is advancing rapidly, real consumer shopping behaviour changes much more slowly, meaning marketers should focus on present-day customer needs rather than over-investing in speculative future AI-driven commerce models.

Q: Why do marketers tend to overestimate AI’s short-term impact?
A: Marketers often assume new technology will transform behaviour faster than it actually does, but historical patterns show that adoption is gradual and shaped by trust, habits, and practical usefulness rather than hype.

Q: How are customers actually interacting with AI in commerce today?
A: Customers are already using AI tools for tasks like product comparison, recommendations, and summarising reviews, but these tools mostly enhance existing shopping behaviour rather than replacing it entirely.

Q: Why is focusing on “what’s in sight” important for marketers?
A: Focusing on what’s in sight means prioritising current customer behaviour, proven channels, and immediate commercial opportunities, rather than building strategies around uncertain future AI-driven shopping models.

Q: What is the recommended approach to AI in commerce strategy?
A: The recommended approach is a balanced one—experiment with AI where it solves real customer problems, but continue optimising existing marketing and commerce systems that are already driving revenue today.