• Taking a restaurant out of business
• Deceiving consumers with crappy products
• Today’s failing solutions are reactive
• AI-Commerce and the promise of real reviews only

Fake reviews are the scourge of e-commerce for consumers (fake positive reviews) and businesses (fake negative reviews) alike.

§Taking a restaurant out of business

A Harvard Business Review working paper done in 2016 found that a one-star increase in Yelp rating leads to a 5-9 percent increase in revenue (source). A likely similar magnitude can be expected in the opposite direction: a one-star drop will reduce revenue by 5 % to 9%. This is the kind of revenue drop that closes restaurants, with their rising costs and razor-thin margins.

Unsurprisingly, businesses facing more competition are more likely to receive fake negative reviews, typically driven by the competition (source) and things are only going to get worse with generative AI.

And let’s not forget about review bombing;when a mob, driven by a viral social media event, decides to bomb the reviews of a business that they have not done business with; just because the business did something that the mob does not agree with.

If you’re a business owner, you most likely are fighting with fake negative reviews one way or another.

§Deceiving consumers with crappy products

Consumers also suffer from fake reviews. From outright fake reviews bought by merchants to raise their visibility in Amazon Marketplace (source) to “paid reviews” written by people who did not purchase the product and got paid to review it instead.

Have you ever seen a person getting paid to issue a review and post a negative one? Even if they did, will the business publish it? I don’t think so….

§Today’s failing solutions are reactive

Today’s solutions to the problem are reactive, like this final ruling issued by the FTC banning fake reviews and testimonials. But playing a reactive game in a world where over 3 million reviews are published daily is setting yourself up for failure.

The real solution is to stop fake reviews from happening all together.

§AI-Commerce and the promise of real reviews only

As soon as Generative AI became a thing, I realized that we’re running towards a world where we can’t trust anything unless it is cryptographically signed. You need to have mathematical proof of the origins of any piece of content if you are to put any trust in it.

For example, this blog post is cryptographically signed to verify that it was published by @9f475...cc946 . The contents of this blog post are immutable and provenance is 100% clear.

The same thing applies to reviews. Reviews must be cryptographically signed to ensure their integrity (content is as the reviewer wrote it) and their originator.

So how does AI-commerce play into the whole thing? It’s all about building the right agent-native rails for commerce.

What would happen if the rails for AI-commerce allowed AI agents to:

• Cryptographically sign a transaction to certify who the buyer is

• Cryptographically sign a review to certify who the reviewer is

• Include in the review a reference to the cryptographically signed transaction

• Use identities for signatures that are universal and not siloed to specific platforms or agents

Now businesses can choose to display only cryptographically signed reviews that are linked to a purchase transaction signed by the same key—i.e., a verified buyer.

And consumers can instruct their AI shopping assistant to consider only cryptographically signed reviews that are linked to a purchase transaction signed by the same key—i.e., from a verified buyer.

Additionally, thanks to universal identities, you as a consumer, get to see all verified reviews, as opposed to today where you need to go to separate platforms to see the reviews on each platform.

And just like that, fake reviews, although still out there, become invisible to consumers. And once invisible, the incentives to continue creating fake reviews disappear.

AI-commerce represents a unique opportunity to optimize the market for buyers and sellers, removing intermediaries that add cost and noise. This will result in lower costs for sellers and more genuine options at a lower cost for buyers.

With AI assistants, product discovery is not centralized anymore and every business has the opportunity to present itself in front of the buyer, at the right time.

Synvya is building the open, agent-native rails for AI-commerce: identity, discovery, and transactions — with no middleman. Visit synvya.com and follow @9f475...cc946 for more on how we’re building a direct, merchant-owned agent channel with verified identity and low-friction payments.