What happens to companies in a world where most people have a tireless and very capable personal agent at their fingertips?
Nowadays, most companies look for a recurring and “owned” relationship with their customers. Booking.com probably loses money on the first flight a customer books with it (because of CAC) but makes it up as the customer buys again and again from it.
And a lot of that recurring relationship happens because of friction: By default, users don't want to go through the hassle of signing up for a new hotel website if the one they're already signed up with offers a good enough deal. Most users also don't even go through the hassle of shopping around, comparing prices, etc. Friction is also why users mostly don't call all their service providers and threaten to cancel so as to get a better price on their services, something they know is possible (e.g., everybody knows there's a 90% chance they can get the annual fees on their credit cards waived if they call and threaten to cancel).
A personal agent won't really care about friction, and future, more capable and efficient agents, powered by more capable and efficient models, will care even less. So if a user tells them (if they even need to tell them at all, because I think they'll most probably be proactive about it) to get the best deal on a cable/internet provider, the agent will do whatever it takes to get them that best deal, even if it means talking to a rep for a couple of hours, or accessing the website to get a quote from 100 different IP addresses to figure out dynamic pricing, etc.
Also, a personal agent doesn't care about me having a recurring relationship with providers. It will get the best objective deal, period. What does that do to recurring relationships? Of course, my agent won't ignore loyalty programs: it will factor them in. But even then, that factoring might mean spending just enough money on credit card A for certain benefits, and then pivoting to credit card B, and then C, once a given objective is met (for example, once I achieve my desired status with an airline).
I think objectivity is an important point from this reflection: my commercial relationships will definitely get more “objective.” They will depend more on the cold, hard qualities of products and services, and less on my laziness, or the opacity of the offerings, or my unconscious brand loyalty, etc.
Does that mean it will be 100% objective? Probably not. But it definitely does mean it will be much more objective.
Game theory
Another interesting consideration is the game theory of how companies will respond to this.
Nobody wants to lose direct relationships with customers. The obvious first choice would be to just work on blocking agents altogether, forcing customers to log in to apps themselves (captchas, etc.). But customers will prefer to have their agents do most things for them. Nobody looks forward to logging into a banking app to send a transfer or browsing Amazon for a good deal on toilet paper [1]. The question is how companies get from here to there. What if some player caves in first? As in some player opens up its services and rolls out the red carpet for ChatGPT or Claude or Siri or Copilot or Poke or Instinct or Town [2]? How will competitors respond? They will probably have to rush and cave in as well…? So the end state will almost certainly be, love it or hate it, a mostly agent-mediated economy.
Margins
A natural consequence of what we discussed is that things are going to get way more competitive in the future, as agents pit companies against each other for the best objective deals. This means margins for consumer companies (and probably b2b companies as well, albeit less [3]) will probably shrink. If companies nowadays make some money betting on customer inertia, they will make less money in the future.
[1] Of course, every company will build their own agent as well. The bank will have its own agent, through which users might do what they need to do within that bank. They will try to have users use this agent and not the agnostic personal agent. But how many agents will a user want to have to talk to? Isn't there going to be one agent that handles almost everything? Which companies will earn the right with consumers to have their own agents? And what will it take to earn that right?
[2] I'm willing to bet that the agent is going to be for most regular people, like my mom, native to the OS, meaning from Google (bc of Android), Apple (bc of iOS), and Microsoft (bc of Windows).
[3] Companies already pay people to get the work done and find the best deals. They will use their own agents, though, so some of this dynamic will port.
