The rise of agentic AI will transform marketplaces from passive aggregators to active facilitators of transactions. The future is Agent-Led Marketplaces.
I agree with your thinking with the role AI will play in the advent of ALMs, but I think you can take the thought exercise even further.
To date, marketplaces have largely served to achieve better *utilization* of supply by the owners of that supply. A suitably large marketplace will be deemed successful by suppliers if they can achieve what they consider some acceptable rate of utilization. You would know better than I what that rate is - I suspect it varies by the type of supply.
But in my mind, the big role for AI to play in marketplaces is to optimize the actual supply *itself*, not just its utilization. What does that mean? Well, let me give some concrete examples and then draw a more abstract generalization.
I'll use farm equipment as a concrete, complicated example because it's in my wheelhouse. Farm equipment tends to have high cost and low utilization. Farmers own (or more accurately, make payments to the bank on) this equipment so that it's available when and where they need it nearly instantaneously. In some parts of the world, like India, there are already marketplaces for farm equipment which enable those who cannot afford such equipment to "lease" it on a short term basis from those who do. Hold that thought for a moment.
A second example, even closer to my heart, would be a drift boat for fly fishing. I would love my own drift boat, but I'm not a fishing guide, and I know it would sit idle in my driveway for probably 300+ days a year. Now you could imagine a traditional marketplace for sporting goods that might let me find an available drift boat for the day and lease it. But does AI enable a better model?
What if AI could understand the usage nature of a particular type of supply, then aggregate a pool of demand for any particular unit of supply that optimized (for some definition) that unit almost perfectly? In my farming example, it might look like a group of farmers geographically oriented south to north, such that in the Spring when the southernmost farmer was done with the equipment it could be utilized further north when the next farmer is ready for it (this exists today btw in some form through service providers). In the Fall this would work the same way in reverse (North to South). In my drift boat example, maybe AI could optimize a group of us fly fishers who live close by who like to (or are able to) fish on different days of the week, thus optimizing the boat such that it is out on the water nearly every day.
Now let's take this asset utilization exercise to its extreme... If you could be assured of an asset's utilization, and understood the initial cost, appreciation or depreciation, financing, maintenance, etc.- two new scenarios present themselves. Either the marketplace *itself* should own (and manage) the assets, or the supply and demand sides of the marketplace merge into a single entity - forming a sort of AI-driven collective. There's even a hybrid approach where the marketplace supports both scenarios.
I think this could be the next generation of marketplaces because you add the entire TAM of the market value of the assets themselves - not just their rental value. You can also optimize the inventory. In my drift boat example, a group of fly fishers could all be part owner of the drift boat itself - not just paying to use it for a day. If I wanted to "sell" my ownership, the marketplace AI could facilitate finding my perfect "buyer" to replace me in my collective. Alternatively, the AI driven marketplace could optimize the inventory of drift boats, managing that inventory as an asset class itself in addition to the rental income of that asset.
So, to generalize, what if the AI driven marketplace doesn't have a distinct supply and demand side at all, but rather reduces the problem down to pure asset optimization? That seems like the real opportunity for AI, as "optimization" is such a wildly different equation depending on the asset and/or demand characteristics.
This is helpful perspective. My one enduring gripe — and it's an aesthetic gripe entirely, for what that's worth — is how aggregation begets homogenization. Many such examples, and I think Faire is a clear one: Now every small retail store from Paris, TX to Copake, NY carries the same slate of products (hello, Graza olive oil). So I hope there's a way for agent-led marketplaces to drive more heterogeneity and novelty. Perhaps this is done by better gauging buyers' preferences (my assumption is that everyone's optimal outcome is different) and connecting that to "unsung supply."
I agree with your thinking with the role AI will play in the advent of ALMs, but I think you can take the thought exercise even further.
To date, marketplaces have largely served to achieve better *utilization* of supply by the owners of that supply. A suitably large marketplace will be deemed successful by suppliers if they can achieve what they consider some acceptable rate of utilization. You would know better than I what that rate is - I suspect it varies by the type of supply.
But in my mind, the big role for AI to play in marketplaces is to optimize the actual supply *itself*, not just its utilization. What does that mean? Well, let me give some concrete examples and then draw a more abstract generalization.
I'll use farm equipment as a concrete, complicated example because it's in my wheelhouse. Farm equipment tends to have high cost and low utilization. Farmers own (or more accurately, make payments to the bank on) this equipment so that it's available when and where they need it nearly instantaneously. In some parts of the world, like India, there are already marketplaces for farm equipment which enable those who cannot afford such equipment to "lease" it on a short term basis from those who do. Hold that thought for a moment.
A second example, even closer to my heart, would be a drift boat for fly fishing. I would love my own drift boat, but I'm not a fishing guide, and I know it would sit idle in my driveway for probably 300+ days a year. Now you could imagine a traditional marketplace for sporting goods that might let me find an available drift boat for the day and lease it. But does AI enable a better model?
What if AI could understand the usage nature of a particular type of supply, then aggregate a pool of demand for any particular unit of supply that optimized (for some definition) that unit almost perfectly? In my farming example, it might look like a group of farmers geographically oriented south to north, such that in the Spring when the southernmost farmer was done with the equipment it could be utilized further north when the next farmer is ready for it (this exists today btw in some form through service providers). In the Fall this would work the same way in reverse (North to South). In my drift boat example, maybe AI could optimize a group of us fly fishers who live close by who like to (or are able to) fish on different days of the week, thus optimizing the boat such that it is out on the water nearly every day.
Now let's take this asset utilization exercise to its extreme... If you could be assured of an asset's utilization, and understood the initial cost, appreciation or depreciation, financing, maintenance, etc.- two new scenarios present themselves. Either the marketplace *itself* should own (and manage) the assets, or the supply and demand sides of the marketplace merge into a single entity - forming a sort of AI-driven collective. There's even a hybrid approach where the marketplace supports both scenarios.
I think this could be the next generation of marketplaces because you add the entire TAM of the market value of the assets themselves - not just their rental value. You can also optimize the inventory. In my drift boat example, a group of fly fishers could all be part owner of the drift boat itself - not just paying to use it for a day. If I wanted to "sell" my ownership, the marketplace AI could facilitate finding my perfect "buyer" to replace me in my collective. Alternatively, the AI driven marketplace could optimize the inventory of drift boats, managing that inventory as an asset class itself in addition to the rental income of that asset.
So, to generalize, what if the AI driven marketplace doesn't have a distinct supply and demand side at all, but rather reduces the problem down to pure asset optimization? That seems like the real opportunity for AI, as "optimization" is such a wildly different equation depending on the asset and/or demand characteristics.
This is great. I haven't seen a lot of discussion on how AI fundamentally changes marketplaces, this is definitely an interesting perspective.
I am excited to dig in more. It feels like Christmas.
This is helpful perspective. My one enduring gripe — and it's an aesthetic gripe entirely, for what that's worth — is how aggregation begets homogenization. Many such examples, and I think Faire is a clear one: Now every small retail store from Paris, TX to Copake, NY carries the same slate of products (hello, Graza olive oil). So I hope there's a way for agent-led marketplaces to drive more heterogeneity and novelty. Perhaps this is done by better gauging buyers' preferences (my assumption is that everyone's optimal outcome is different) and connecting that to "unsung supply."