Cartier is a bad choice, says Claude

My new Instinct shopping agent is only buying mundane groceries. For now. Please let me know if it steals my identity and sends you pleading messages for money 🙂

 

My faith in tech is at a low point after Google Maps stranded me on a flyover for 90 minutes on my way to meet Katja Forbes, the author of “Machine Customers”. More on that in a bit.

I asked Claude what is the best gold ring to buy.  It analysed the ring based on purity, quality, resale, making charges and wastage, and told me to buy a plain gold hallmarked ring from a known jewelry chain. When I asked why it did not consider Cartier it went full south Indian grandma and huffed that Cartier was merely 18K not 22K, and I was paying for brand not gold. Claude graciously made an exception for gifting and er, status. Hinting – again channeling its inner south Indian soul – that I was a shallow person who need a brand to make me feel good.

 

Let’s be real.  Humans have wants. And those wants are often not based on functional attributes. I need a phone, but I want an iPhone.

 

(Santa, are you listening?)

 

What do marketers have to do differently to address the new buying landscape of shopping agents? This is urgent – Instinct invitations are the most in-demand freebies on my whatsapp groups. Here are my top 3 suggestions.

 

And then, because I could not help it, a bit of philosophy on where are we going with this complex commercial landscape of choice.

For all their desire for consciousness, agents assess products on known parameters. As marketers we have to make sure that the parameters favourable to our brands are in the consideration set, and backed up with data that makes sense to the AI agent. This is way beyond adding a machine generated FAQ to every webpage, which most of us already do.

 

In the case of Cartier, they need to provide more data on the gifting value of the product, auction resale data, and testimonials from local people.

 

Lest you think that I am using AI agents to buy Cartier rings before breakfast on a Friday, I ran a more practical query asking for a milk recommendation based on purity and nutrition. Though I did not ask for organic, Claude surfaced Akshayakalpa saying they had the best data on my two parameters and brought in 3rd party lab testing and sourcing traceability as additional parameters.

The agents are looking for information, just like humans. But they need structured data backed up by proof. They also do not suffer from information asymmetry – they can search everywhere for information.

 

As marketers we need to build this marketing highway. It may be data rich webpages meant for machines not humans, it may be reddit posts, google reviews, news articles – whatever works best for the hungry AI hordes.

 

We also need to rethink our gatekeeping – for example one of our recurring purchases for elder care comes from a site called Karein. My bot was barred by captchas from setting up an account. It also could not access BigBasket or Amazon. That pushed my purchases to sites that were easier to transact with ie Blinkit and Licious.

 

There are also other more deep aspects that Katja Forbes writes about such as all agents look the same to the seller but the owner of the agent may have special needs or attributes that need to be taken into account at the point of fulfillment. How do we plan for that?

 

B2B agentic buying is similar but the parameters are different and need even more hard data.  And guard-rails.

End user preferences are still the final call. If the users want added protein organic milk or Cartier the search will be narrowed down to those specific categories. So in our desire to accommodate Agentic AI we should not forget about building human preferences – that can still override the agents 🙂

 

If you want to bypass the Agentic AI digital chokehold on access, you can use the physical world to build preferences. Conferences for B2B and sampling for B2C fulfil this need.

 

Old-fashioned brand storytelling is needed more than ever – addressing the latent needs make a bottle of Evian taste better than the tap.

I grew up with precisely one brand of cheese in our local store – Amul Cheese Cubes. 

 

Wants and needs were the same, Cheese =Amul.

 

I could of course continue to live that way – cheese-and-glace cherry on a toothpick anyone? – but if I want to enjoy the huge number of choices now available, I practically need super intelligence to help me figure out what is best for me. I am on the side of choice, and we need to decide how we wish to live – consciously reduce the choices we make to avoid being overwhelmed in some areas and dopamine-maxxing in others.

 

Yes, Google Maps occasionally gets me lost or hopelessly delayed. This is due to information asymmetry – if the underpass is flooded humans can see it in real time, but Google Maps needs to wait for a digital update. On the other hand Google Maps allows me to confidently navigate places that I know nothing about – whether it is Lichtenstein or Lepakshi.

Today is Mahatma Gandhi’s birthday. He had a difficult relationship with machines. He wanted machines-as-tools, that helped reduce the labour of humans, not displaced them with no alternatives. That in his view was the difference between the Singer Sewing Machine and the Power Loom – one allowed the individual to practice their craft in an easier way but remain connected with the output, while the other reduced the human to a button puncher.  We are once more facing that seismic transition.

 

Can we use our new power to make hand-crafted handbags from KadamHaat available? Or make ethically sourced skincare from Birdsong Life a habit? Or scale the power of resilient water and livelihoods with IIT-IIT? I hope so. And yes, I’m an investor in the first two and helping the third with communication.

 

Gandhi was also a big proponent of self-sufficiency. I am speaking to students at Kristu Jayanti University as part of Times of India’s Leaders on Campus series on AI-powered marketing. One angle is that rather than see themselves as individual contributors they can – with the power of AI – position themselves as an Agency-of-One, delivering results that were previously impossible for a single person. In some ways AI makes entrepreneurship more accessible for more – if you are able to crack the science and art of discovery.

 

Five years ago, in the midst of lockdowns, I wrote Issue #475 titled “Make it Easy for Me” – that thought is often how most innovations emerge.

 

Have a great weekend!