The Future of AI Agents

Published on August 27, 2025 9:58 PM GMTHere is the question that I have been losing sleep over: "What would it take for personal AI or agents to find its consumer moment?". About 95% of public startups are working on agents for B2B SaaS. So far, these agents are semi-reliable and perform the best when an outcome is clearly definable and measurable such as minutes or money saved. When building agents for consumers though, we are still very early. There are a lot of tensions to navigate when building the "perfect consumer agent". Output can quickly become indeterministic in a human-AI conversation. Consumers also have a high bar for what an AI presence should feel like in their lives. They expect the personality to be realistic and not sycophantic. Context about users such as their preferred brands, budget, and allergies should be documented implicitly. Finally, the agent should be able to inform the user on why it can or can't do something (ex: this website looks suspicious to buy from). Let's do a thought exercise. I'm going to make a series of statements on what I think would need to happen for agents to become ready for mainstream adoption by consumers. Hint: the revolution begins with payments & purchases. ChatGPT is uncool culturally in my generation (I am 18). The relationship is one-way and reactive.The future of e-commerce will be driven by agents. It is only a matter of how soon we get there.While there are hundreds of use cases that agents could potentially assist in, only a few of them are high-leverage. One such use cases is paying for what you need, which is currently fragmented. Need groceries? Instacart it. Need to buy protein powder or mascara? Go on Amazon. What if I had one agent to streamline all of these purchases?To ease agents into people's lifestyles, automating repeat purchases of the essentials is a good start. With health, fitness, & beauty, we rarely deviate from our favorite brands. If you can keep me stocked proactively, that would earn my goodwill.Money spent and saved is clear and measurable, providing agents the best conditions to thrive in. So step it up a notch by helping me save. Negotiate on my behalf for bulk orders, and find price drops so that you can get my money back. This was actually the core model of Paribus, founded by Ramp CEO, and they did this in 2014 with just NLP.The last but most difficult part is proactive discovery by the agent. Once you know my past purchase history (gleaned from Plaid/Email receipts) & taste profile, find things that surprise me and fall within my constraints. The play here is that people end up comparing agents and their most "underground finds" (shareable moments ). Who has the coolest agent? If you're not happy with how your "niche" agent is, keep interacting with it until it gets you[1].The agent can't just be designed to be useful. It should feel like a helpful person in your life who just "gets" your taste[2]. I say let the agent have a name of your choosing. Also, modality can create separation of roles. Ex: the agent only texts you, but as the final authority, you can text or send a voice message to the agent.The actual agent is invisible and operates in the background (by consequence, you don't need to run it 24/7). Users only interact with the outcomes & give the final approval. This is very different from computer-using agents, which actively walks you through each step.For the user, programming the personal agent should be fun. Again, this is mainly a design question, but I think a solid way to do this is to let humans curate through their words[3].Designing the agent's flow from suggestion to execution will be interesting. Tap into existing frameworks (Apple Pay) & form factors such as a mobile app from which you can text the agent. In this agent-driven economy, I've  seen some suggest that agents & humans will pay via stablecoin from a wallet[4]. I think phone & contactless pay will reign supreme for another decade.  Plus, ~90% of online retailers now accept Apple Pay, and cash on hand is not a norm with my generation.Soon, your agent goes from buying from self to buying for others. It gathers your social context by, for instance, integrating with your Google Calendar and seeing whose birthday is coming up.  Ok. What a list. Personally, I  am excited for this future, and I'm confident that a version of what I have presented above will happen. Think about it. Your greatest strength and point of pride is knowing what you like and don't like. And if you are extra good, others trust your opinions and finds. So let the agent do the boring stuff (reorders, negotiations, returns, research). Your preferences like the type of food you eat, music you listen to, and style you buy will remain yours. Only now, these preferences are further amplified by your agent. Feel free to let me know though which parts of the list you agree or disagree with. If you want to help build around this question and push the category forward, please feel free to reach out. I truly believe in using AI to create something novel and delightful for me and my loved ones. I'm on X: @kavyavenkat4. Send hatemail to: kavyav500@gmail.com. Your grievances will be acknowledged. Two more bonus sections for you all. Downstream Implications: Your agent becomes a "status symbol". At the end of the day, you still call the shots. So the agent is just a reflection of  the values and desires you chose to encode.Can an AI emulate the taste of a human? This is a cool underlying question. And how much unpredictability (similar to 'temperature') should be programmed into the agent? I personally would love to give my personal agent pocket money ($20-$30) and see what it can surprise me with[5].Your agent will act like a paywall for your time and resources[4]. Products will have to compete for the attention of your agent(s). AI SEO and llm.txt files are examples of our digital infrastructure evolving to be agent-friendly.Agents can trade notes and learn from each other. With every user session, the agents indexes the Internet in a way that is increasingly relevant to human culture. Imagine a future where someone asks "How many people have actually bought this?", and you have real data coming from other agents to verify the purchase.  For the counterarguments I anticipate, I've put together a brief FAQ: Q: What about Perplexity Comet? A: Comet seems awesome, and if you prompt it on a workflow to execute, it will do a decent job. I do think that on a mobile phone, you don't want to see the agent go through every step in front of you (search for e-commerence should happen in the background on its own time). Perplexity's distribution is also weak. It feels like a great power user tool, but even then, the average user still chooses ChatGPT to compare items and get  recommendations. Q: Couldn't any of the foundational AI companies make their own version of this purchasing agent? A: Absolutely, and they probably will. These companies don't know how to distribute and design for cultural fit though[6]. The bigger issue is that no matter what these companies promise you, they will use your data against you (selling to advertisers). I think we are ready for a different model. Once you know the preferences of enough people, you have all the leverage. Intelligence about collective demand will matter more in a future when agents do the research, negotiation, and ordering on your behalf. Q: What is the monetization model? A: This is a good question. You can easily set up tiers where each agent has a different set of capabilities. Maybe you have some agents subscribe to each other if you want to buy the same things as your favorite human influencer or curator. Affiliate revenue and commission on transactions is also another method of monetization. Finally, I think collective intelligence is something future business would pay for. Here is a sample insight: "50 people want this product to be in the color blue". The key here is that the demand being extracted directly from what consumers actually want, not manufactured by advertisers trying to drum up interest. Q: Wouldn't this idea end up as an ad business once it matures? A: See the second part of question 2. Q: Why require the 'approval' by the user for each purchase? A: Money is deeply personal and sensitive. Most want control over where every dollar goes[7]. Some may argue that this kills the "autonomous" nature of this agent. Discovery does happen autonomously. It is the final "yes" (or Apple Pay Click) that requires the human input, and through this purposeful friction, you build trust anyway. Note that early adopters might be willing to allocate a part of their budget for serendipitous discovery and purchase by the agent [8].This post is cross-posted from Substack. ^Scott Belsky writes about how "personalization effects" are the new network effects. Valuable AI products shouldn't just collect context about you. They should use it to improve every turn of the conversation. ^The word 'taste' is thrown around a lot. For this post, I define taste as this higher-dimension algorithm that helps you choose which of the options you like. It doesn't matter if your taste is better or worse than someone else's. The agent is designed to capture YOUR taste. ^Humans are natural curators. Just look at how culturally relevant sites like Pinterest or Letterboxd and feautres like "Saved Reels" are. ^Great essay by Daisy Alioto titled "The Future of Media is Bank". She argues that agents will soon decide for us which media we consume. The media is accessed by paying for it with a model that is more flexible and spontaneous than subscriptions. ^Let me know if you are interested in the results of the "Surprise Me" experiment. ^This might be a hot take, but ChatGPT's Cambrian success is carried by first-mover advantage. Operator exists, but I believe there is an opportunity for this payments agents idea to be independently developed. OpenAI is stretched thin and is heavily invested in the AGI race. Experience and design seem like an afterthought. ^This sentence was inspired by my desire to have an AI evaluate my spending decisions. I was going to build and sell as a personal finance tool before realizing why this idea wouldn't work: (1) Most people don't try to budget or save let alone pay for a tool like that (see why Mint.com failed). (2) People aren't rational when it comes to money.  (3)  Revealing statistical insights about your spending habits is not enough. You need to give users a reason to return multiple times in a day. Numbers always have to be accompanied by context that is ideally proactive. ^This is a reasonable assumption. The analogue is passive investment vehicles (index funds, ETFs or algo-trading). Discuss

2025-08-27 23:00 GMT · 5 days ago www.lesswrong.com
2025-08-27 23:00 GMT · 5 days ago www.lesswrong.com

Published on August 27, 2025 9:58 PM GMTHere is the question that I have been losing sleep over: "What would it take for personal AI or agents to find its consumer moment?". About 95% of public startups are working on agents for B2B SaaS. So far, these agents are semi-reliable and perform the best when an outcome is clearly definable and measurable such as minutes or money saved. When building agents for consumers though, we are still very early. There are a lot of tensions to navigate when building the "perfect consumer agent". Output can quickly become indeterministic in a human-AI conversation. Consumers also have a high bar for what an AI presence should feel like in their lives. They expect the personality to be realistic and not sycophantic. Context about users such as their preferred brands, budget, and allergies should be documented implicitly. Finally, the agent should be able to inform the user on why it can or can't do something (ex: this website looks suspicious to buy from). Let's do a thought exercise. I'm going to make a series of statements on what I think would need to happen for agents to become ready for mainstream adoption by consumers. Hint: the revolution begins with payments & purchases. ChatGPT is uncool culturally in my generation (I am 18). The relationship is one-way and reactive.The future of e-commerce will be driven by agents. It is only a matter of how soon we get there.While there are hundreds of use cases that agents could potentially assist in, only a few of them are high-leverage. One such use cases is paying for what you need, which is currently fragmented. Need groceries? Instacart it. Need to buy protein powder or mascara? Go on Amazon. What if I had one agent to streamline all of these purchases?To ease agents into people's lifestyles, automating repeat purchases of the essentials is a good start. With health, fitness, & beauty, we rarely deviate from our favorite brands. If you can keep me stocked proactively, that would earn my goodwill.Money spent and saved is clear and measurable, providing agents the best conditions to thrive in. So step it up a notch by helping me save. Negotiate on my behalf for bulk orders, and find price drops so that you can get my money back. This was actually the core model of Paribus, founded by Ramp CEO, and they did this in 2014 with just NLP.The last but most difficult part is proactive discovery by the agent. Once you know my past purchase history (gleaned from Plaid/Email receipts) & taste profile, find things that surprise me and fall within my constraints. The play here is that people end up comparing agents and their most "underground finds" (shareable moments ). Who has the coolest agent? If you're not happy with how your "niche" agent is, keep interacting with it until it gets you[1].The agent can't just be designed to be useful. It should feel like a helpful person in your life who just "gets" your taste[2]. I say let the agent have a name of your choosing. Also, modality can create separation of roles. Ex: the agent only texts you, but as the final authority, you can text or send a voice message to the agent.The actual agent is invisible and operates in the background (by consequence, you don't need to run it 24/7). Users only interact with the outcomes & give the final approval. This is very different from computer-using agents, which actively walks you through each step.For the user, programming the personal agent should be fun. Again, this is mainly a design question, but I think a solid way to do this is to let humans curate through their words[3].Designing the agent's flow from suggestion to execution will be interesting. Tap into existing frameworks (Apple Pay) & form factors such as a mobile app from which you can text the agent. In this agent-driven economy, I've  seen some suggest that agents & humans will pay via stablecoin from a wallet[4]. I think phone & contactless pay will reign supreme for another decade.  Plus, ~90% of online retailers now accept Apple Pay, and cash on hand is not a norm with my generation.Soon, your agent goes from buying from self to buying for others. It gathers your social context by, for instance, integrating with your Google Calendar and seeing whose birthday is coming up.  Ok. What a list. Personally, I  am excited for this future, and I'm confident that a version of what I have presented above will happen. Think about it. Your greatest strength and point of pride is knowing what you like and don't like. And if you are extra good, others trust your opinions and finds. So let the agent do the boring stuff (reorders, negotiations, returns, research). Your preferences like the type of food you eat, music you listen to, and style you buy will remain yours. Only now, these preferences are further amplified by your agent. Feel free to let me know though which parts of the list you agree or disagree with. If you want to help build around this question and push the category forward, please feel free to reach out. I truly believe in using AI to create something novel and delightful for me and my loved ones. I'm on X: @kavyavenkat4. Send hatemail to: kavyav500@gmail.com. Your grievances will be acknowledged. Two more bonus sections for you all. Downstream Implications: Your agent becomes a "status symbol". At the end of the day, you still call the shots. So the agent is just a reflection of  the values and desires you chose to encode.Can an AI emulate the taste of a human? This is a cool underlying question. And how much unpredictability (similar to 'temperature') should be programmed into the agent? I personally would love to give my personal agent pocket money ($20-$30) and see what it can surprise me with[5].Your agent will act like a paywall for your time and resources[4]. Products will have to compete for the attention of your agent(s). AI SEO and llm.txt files are examples of our digital infrastructure evolving to be agent-friendly.Agents can trade notes and learn from each other. With every user session, the agents indexes the Internet in a way that is increasingly relevant to human culture. Imagine a future where someone asks "How many people have actually bought this?", and you have real data coming from other agents to verify the purchase.  For the counterarguments I anticipate, I've put together a brief FAQ: Q: What about Perplexity Comet? A: Comet seems awesome, and if you prompt it on a workflow to execute, it will do a decent job. I do think that on a mobile phone, you don't want to see the agent go through every step in front of you (search for e-commerence should happen in the background on its own time). Perplexity's distribution is also weak. It feels like a great power user tool, but even then, the average user still chooses ChatGPT to compare items and get  recommendations. Q: Couldn't any of the foundational AI companies make their own version of this purchasing agent? A: Absolutely, and they probably will. These companies don't know how to distribute and design for cultural fit though[6]. The bigger issue is that no matter what these companies promise you, they will use your data against you (selling to advertisers). I think we are ready for a different model. Once you know the preferences of enough people, you have all the leverage. Intelligence about collective demand will matter more in a future when agents do the research, negotiation, and ordering on your behalf. Q: What is the monetization model? A: This is a good question. You can easily set up tiers where each agent has a different set of capabilities. Maybe you have some agents subscribe to each other if you want to buy the same things as your favorite human influencer or curator. Affiliate revenue and commission on transactions is also another method of monetization. Finally, I think collective intelligence is something future business would pay for. Here is a sample insight: "50 people want this product to be in the color blue". The key here is that the demand being extracted directly from what consumers actually want, not manufactured by advertisers trying to drum up interest. Q: Wouldn't this idea end up as an ad business once it matures? A: See the second part of question 2. Q: Why require the 'approval' by the user for each purchase? A: Money is deeply personal and sensitive. Most want control over where every dollar goes[7]. Some may argue that this kills the "autonomous" nature of this agent. Discovery does happen autonomously. It is the final "yes" (or Apple Pay Click) that requires the human input, and through this purposeful friction, you build trust anyway. Note that early adopters might be willing to allocate a part of their budget for serendipitous discovery and purchase by the agent [8].This post is cross-posted from Substack. ^Scott Belsky writes about how "personalization effects" are the new network effects. Valuable AI products shouldn't just collect context about you. They should use it to improve every turn of the conversation. ^The word 'taste' is thrown around a lot. For this post, I define taste as this higher-dimension algorithm that helps you choose which of the options you like. It doesn't matter if your taste is better or worse than someone else's. The agent is designed to capture YOUR taste. ^Humans are natural curators. Just look at how culturally relevant sites like Pinterest or Letterboxd and feautres like "Saved Reels" are. ^Great essay by Daisy Alioto titled "The Future of Media is Bank". She argues that agents will soon decide for us which media we consume. The media is accessed by paying for it with a model that is more flexible and spontaneous than subscriptions. ^Let me know if you are interested in the results of the "Surprise Me" experiment. ^This might be a hot take, but ChatGPT's Cambrian success is carried by first-mover advantage. Operator exists, but I believe there is an opportunity for this payments agents idea to be independently developed. OpenAI is stretched thin and is heavily invested in the AGI race. Experience and design seem like an afterthought. ^This sentence was inspired by my desire to have an AI evaluate my spending decisions. I was going to build and sell as a personal finance tool before realizing why this idea wouldn't work: (1) Most people don't try to budget or save let alone pay for a tool like that (see why Mint.com failed). (2) People aren't rational when it comes to money.  (3)  Revealing statistical insights about your spending habits is not enough. You need to give users a reason to return multiple times in a day. Numbers always have to be accompanied by context that is ideally proactive. ^This is a reasonable assumption. The analogue is passive investment vehicles (index funds, ETFs or algo-trading). Discuss

Original: https://www.lesswrong.com/posts/sDA3jEt4wAD3shKum/the-future-of-ai-agents