If you’re investing in AI, building with it, or worried about being replaced by it, there’s one principle you need to understand: Verifier’s Law.
Last Friday, we hosted a session as part of the MANS International AI Strategy Series, where we unpacked Verifier’s Law — developed by Jason Wei, a leading AI researcher who recently left OpenAI to join Meta’s Superalignment team.
This under-the-radar principle explains why AI is transforming some industries faster than others — and how to anticipate what’s coming next.
What is Verifier’s Law?
“Any task that is possible to solve and easy to verify will be solved by AI.” — Jason Wei
In simpler terms: if a machine can be trained to judge whether an output is good or bad—quickly, accurately, and at scale—it will master that task faster than we expect.
What Is the “Asymmetry of Verification”?
The Asymmetry of Verification refers to tasks that are difficult to perform but easy to verify. In other words, while producing the solution requires significant time or expertise, checking whether the solution is correct is fast and straightforward.
Here are some everyday examples:
Sudoku: Solving a puzzle can take 20 minutes, but verifying a completed solution takes just 2 seconds.
Software development: Writing backend code may take weeks; running it to check if it works takes only seconds.
Research questions: Finding a reliable answer might take hours; verifying it can take just a few clicks.
This asymmetry is exactly what today’s most advanced AI systems — like ChatGPT, GitHub Copilot, and AlphaEvolve — are harnessing at scale. They don’t just solve problems — they thrive in domains where verification is easy, allowing them to learn faster, adapt quicker, and outperform expectations.
Five Criteria That Make a Task “AI-Solvable”
According to Verifier’s Law, developed by AI researcher Jason Wei, certain tasks are particularly well-suited for AI. They share five key characteristics:
Objective truth — There’s a clear standard for what’s correct, with little room for disagreement.
Fast verification — You can quickly determine whether a solution is right or wrong.
Scalable verification — It’s easy to verify many outputs at once, enabling rapid learning.
Low noise — The verification process reliably reflects the true quality of the solution.
Continuous reward — Solutions can be ranked along a spectrum from poor to excellent, not just right or wrong.
If a task meets three or more of these criteria, it’s a strong candidate for AI automation — and likely to be transformed sooner than expected.
Case Studies: How OpenAI and Google Brain Apply Verifier’s Law
AlphaEvolve (Google Brain)
Challenge: “What’s the smallest hexagon that can fit 11 unit hexagons?” Insight: Solving this geometric optimization problem is difficult — but verifying a solution is instant. Result: AI explored solutions at scale, uncovering designs that human experts hadn’t imagined.
These breakthroughs didn’t just rely on AI’s intelligence — they were engineered around verifiability, accelerating progress in ways traditional approaches couldn’t.
What This Means for You
For Investors
Use Verifier’s Law as a lens for evaluating AI startups. Ask yourself:
Can the AI’s output be tested and measured quickly?
Are feedback loops built into the product design?
Does the task satisfy three or more of the five AI criteria?
Invest where verification is fast and iteration is cheap. That’s where AI compounds value quickly.
For Founders
Want to build a high-impact, defensible AI product? Focus on:
Tasks that are time-consuming or expensive to solve manually
But cheap and fast to verify
Don’t just solve problems — engineer them to be verifiable. That’s where you unlock scalable feedback and competitive advantage.
For Professionals
Concerned about being replaced by AI? You’re not alone. Here’s the hard truth:
AI will dominate tasks where solutions are easy to verify.
Resilient jobs will involve:
Ambiguity, complexity, or nuance in defining a “good” result
High-context decision-making
Human trust, ethics, or emotional intelligence
The more difficult it is to verify your work, the longer it will take for AI to replace it.
Ready to Apply Verifier’s Law?
We’re opening 2 spots for investors who want expert support evaluating their AI portfolios — and 3 spots for founders seeking feedback on whether their project has real market potential.
Depression doesn’t just affect one person—it casts a shadow over everyone who loves them. If you’re walking alongside someone in this struggle, here are three often-overlooked truths that may help guide the way.
1. “Just Cheer Up” Won’t Help Them—And It’s Not Your Fault
When my friend first learned of her daughter’s illness, she thought sunshine and the ocean might help lift her daughter’s spirits. She carefully planned a trip to Hawaii. But when they arrived, her daughter locked herself in the hotel room and didn’t come out all day. My friend sat alone on the rocks by the sea, tears mixing with the saltwater as she quietly sobbed.
Depression isn’t something that sunshine or sheer willpower can fix. Just as you wouldn’t tell someone with asthma to “just breathe,” you can’t expect someone with major depressive disorder to “just think happy thoughts.” Depression is a medical condition—a neurochemical imbalance—not simply a matter of mindset.
The first and most important step for family members isn’t to rush their loved one into “getting better,” but to first accept this simple truth: healing takes time—and professional help. If your efforts haven’t brought immediate results, it’s not because you haven’t done enough or haven’t tried hard enough.
2. You Don’t Need to Be a “Saviour” — Just Being There Is Enough
When I visited my friend’s home, I noticed a printed “Recovery Plan” taped to the fridge — an A4 sheet filled with tightly packed schedules and rules:
“7:30 AM — Must eat a breakfast rich in Omega-3s”
“3:00 PM — Minimum 30 minutes of sunlight”
“8:00 PM — Parent-child reading (no negative topics allowed)”
Pointing to the plan, she said, “I’m scared I’ll miss something. What if things get worse because I didn’t do enough?”
This reflects a common misconception among families: thinking they have to be the “rescuer,” carrying the burden of “fixing” their loved one.
But the truth is, being there for someone with depression is less about lighting the path and more about sitting with them in the dark.
You don’t need to carry a torch — just let them know they’re not alone. That’s enough.
As one doctor wisely put it: “A person with a broken leg needs crutches, but the crutches don’t blame themselves for not being the leg.” What you’re doing now is offering something to lean on, to take the pressure off the pain.
If they don’t want to talk, don’t force it.
If they repeat the same worries over and over, don’t rush to say, “Stop overthinking.”
If they eat one more bite than yesterday, or even just manage a small smile in your presence, that’s a flicker of light worth noticing.
Your presence alone is a quiet, powerful form of support!
3. Don’t Forget Yourself — Your Exhaustion Deserves To Be Seen Too
Another time, at the grocery store, my friend stood frozen, holding two yogurt: her favourite peach, and the sugar-free one her daughter preferred. She put back the peach. Then, she crouched down by the shelf and cried. “With my daughter like this, how could I possibly care about what I like? Even picking a yogurt feels selfish.”
Her words cut deep. Her phone’s album told the story: the last 100 photos, all of her daughter — medication schedules, meals, a profile staring out the window. It was as if she’d emptied her own life just to make room for her daughter’s every emotion.
But have you considered: if you’re withering inside, how can you truly water someone else?
Ten minutes a day to just zone out. One call a week to chat with friends. Even allowing yourself to cry behind a bathroom door. These small acts of “selfishness” are not betrayals. These are ways to sustain yourself, so you can keep standing beside your loved one for longer.
You deserve to be seen. You need care, too.
Final Words
Family members supporting someone with depression are like travellers on a road without streetlights. You have to gently help your loved one take each step, while also finding your own source of light. This path is incredibly difficult — there will be exhaustion, guilt, and moments when you feel like giving up. But please, always remember: you are not fighting alone.
The tears you wipe away in secret, the smiles you force when you feel broken, the late nights spent researching — these quiet acts of love carry both of you forward.
May everyone silently supporting someone in the darkness be gently embraced by this world.
If you know someone currently experiencing the shadows of depression, please share this message with them. And quietly, softly, whisper these words:
“Hold on a little longer — you’re not fighting this battle alone.”
In past articles, we’ve focused mainly on innovative treatments for mild depression. Today, we turn to a breakthrough in the treatment of major depressive disorder (MDD) — a condition that affects millions worldwide. For the first time, patients with major depression can access a prescription digital therapeutic via their smartphone: Rejoyn.
What is Rejoyn?
Rejoyn is currently the only mobile app approved by the U.S. FDA as an adjunctive treatment for adults (22+) diagnosed with MDD who are already on antidepressants. Available by prescription only, Rejoyn offers a 6-week course that combines proprietary brain training exercises with medication to reduce depressive symptoms, with no significant side effects reported.
Unlike typical wellness apps, Rejoyn is classified as a medical device. It earned FDA approval in 2024 after successful Phase 3 clinical trials, with strong scientific evidence supporting its safety and efficacy.
How Does Rejoyn Work?
In many patients with depression, communication between the amygdala (emotion centre) and prefrontal cortex (thinking centre) becomes impaired, making emotional regulation increasingly difficult over time.
Rejoyn’s exercises are based on neuroplasticity, the brain’s natural ability to rewire itself. By simultaneously activating emotional and cognitive regions, Rejoyn strengthens these connections, helping the brain better manage emotional responses.
EFMT: The Emotional Faces Memory Task
One of Rejoyn’s key technologies is the Emotional Face Memory Task (EFMT), a scientifically validated exercise. Users are asked to recognize facial expressions (e.g., happy, sad, angry) and recall matching ones from earlier. This trains both working memory and emotional processing, gradually increasing in difficulty to challenge the brain to stay focused even in the face of emotional distraction.
In addition to EFMT, Rejoyn offers short skill-building lessons that teach users how to manage strong emotions, combat negative thinking, and take positive actions. Each lesson is around 5 minutes, and it is recommended to do it three times a week. Motivational reminders are sent to help users stay on track.
Importantly, Rejoyn is not a replacement for medication, but a complementary tool prescribed alongside existing antidepressant treatments. For patients seeking accessible, structured support in their daily lives, Rejoyn is a promising new option.
Advances in medical science have led to FDA-approved devices like Rejoyn, opening new pathways for treating depression. Still, even the most advanced technologies must be supported by compassionate, human-centred care to have a meaningful impact.
In our next issue, I want to speak directly to the families and loved ones standing by someone with depression. There are important truths you may not yet recognize, but understanding them can make all the difference in your journey together.
China is rapidly emerging as a global front-runner in the race to integrate AI with consumer hardware.
On June 26, 2025, Xiaomi officially launched its first pair of AI glasses, positioned as “the personal intelligent device of the next era and your portable AI gateway,” directly challenging the Ray-Ban Meta AI glasses.
Image source: Xiaomi, Xiaomi AI glass
Core Feature Comparison: Xiaomi AI Glasses vs. Meta AI Glasses
Here’s the comparison table with prices converted from CNY to USD using the latest exchange rate (¥1 = $0.1396 as of early July 2025):
Real-time multilingual translation (10 languages), LLM Q&A, smart home control
Meta AI Q&A only
Xiaomi (richer AI capabilities)
Battery Life
8.6 hours (typical use)
8 hours
Xiaomi (longer life)
Charging
45 min (USB-C)
75 min
Xiaomi (faster)
Ecosystem
Deep integration with HyperOS, Mi AI assistant, and smart home devices
Meta ecosystem
Xiaomi (stronger interconnectivity)
Style
3 colors (black, tortoise brown, parrot green)
20+ Ray-Ban styles
Meta (more fashion-forward)
While Xiaomi lags slightly behind Meta in terms of fashion appeal and brand recognition, it demonstrates clear advantages in hardware performance, AI interaction, ecosystem integration, and pricing, especially within the Chinese market.
With more Chinese companies entering the space, leveraging lightweight design, intelligent features, high cost efficiency, and tightly integrated ecosystems, China is poised to lead global consumer AI hardware-software innovation.
China and the U.S. in Foundational AI Research: Each With Its Own Edge
According to Trends-Artificial Intelligence, the BOND AI trends report released by “Internet Queen” Mary Meeker, the U.S. and China are now in an intense phase of competition in AI. In foundational research, each country holds distinct strategic advantages.
🇨🇳 China’s Strength: Dual Engines of Open-Source Ecosystems and Industrial Intelligence
China is rewriting the global playbook for AI through the explosive growth of its open-source ecosystem and rapid AI industrialization:
Scale and Quality of Open Models: As of Q2 2025, China has released several benchmark open-source models, including DeepSeek-R1 (trained at just 1/10 the cost of OpenAI’s), Alibaba’s Qwen-32B, and Baidu’s Ernie 4.5, covering a wide range of use cases from language to multimodal and code generation.
World’s Largest Open Model Hub: Alibaba Cloud’s ModelScope now hosts over 70,000 open models and has a developer base of 16 million, making it one of the largest open-source AI communities globally.
Affordable, High-Performance AI: Chinese models like DeepSeek V3 are optimized for cost-efficiency, offering industry-grade performance at a fraction of traditional training costs (as low as 1/10). This affordability accelerates the democratization of AI in SMEs and developing nations, enabling broader global access to advanced technology.
Image source: Mary Meeker BOND Trends-Artificial Intelligence Page 265
2. China’s Lead in Industrial Robotics: A Result of Policy Support and Supply Chain Strength
Global #1 in Installations: In 2023, China accounted for over 50% of the world’s industrial robot installations. International manufacturers such as BMW and Tesla are already incorporating Chinese robotic solutions into their production lines.
Challenges Remain: Despite rapid progress, China still relies heavily on imported high-end sensors and real-time control systems. However, domestic substitution is accelerating as local innovation and investment ramp up.
Image source: Mary Meeker BOND Trends-Artificial Intelligence Page 288
🇺🇸 U.S. Strength: Dual Advantage in Foundation Models and Chip Dominance
The U.S. continues to hold a clear edge in the development of core AI foundation models:
Closed-Source Superiority: Proprietary models like OpenAI’s GPT‑4.5 and Anthropic’s Claude consistently lead on benchmarks such as MMLU and HumanEval, particularly in complex reasoning and long-context understanding.
Open-Source Closing In: While Chinese models like DeepSeek-R1 have rapidly narrowed the performance gap (from 15.9% in 2023 to just 1.7% by 2025), closed-source U.S. models still dominate high-end commercial ecosystems, especially in enterprise and advanced research use cases.
Image source: Mary Meeker BOND Trends-Artificial Intelligence Page 142
2. Chip and Compute Supremacy: From Hardware Monopoly to Global Ecosystem Control
The U.S. has secured a dominant position in global AI infrastructure through its leadership in chip technology:
NVIDIA’s Absolute Dominance: NVIDIA GPUs power over 90% of AI model training worldwide. Its hardware, particularly the H100 and H200 series, remains the gold standard for large model training.
Infrastructure Leverage: NVIDIA-related compute accounts for an estimated 25% of global data centre capital expenditure (CapEx), underscoring its strategic role in the AI economy, from cloud providers to model developers.
Image source: Mary Meeker BOND Trends-Artificial Intelligence Page 109
Export Controls Spark a New Paradigm of “Accelerated Evolution” in China’s Chip Industry
Amid the growing U.S.–China tech decoupling, particularly around semiconductors, U.S. export restrictions on AI chips to China have triggered a powerful “crisis-response mechanism” across China’s chip sector:
Technological Pressure: Real-world deployment demands are driving architectural innovation and improved chip yields.
Market Pull: Rising substitution demand for domestic chips is fueling greater R&D investment.
Strategic Loop: A dual-cycle strategy is emerging — import substitution at home, technology export abroad.
Beyond Huawei, Xiaomi has also entered the self-developed chip race — primarily to reduce material costs (mirroring how Apple’s M1 chip boosted Mac profit margins) and to mitigate geopolitical risk, such as potential sanctions or supply disruptions similar to Huawei’s.
The launch of Xiaomi’s Surge O1 chip marks the company’s first fully self-developed System-on-Chip (SoC) for smartphones, giving it core control from architecture design to functional integration. This shift reduces dependency on Qualcomm, lowers licensing fees, and improves gross margins, similar to how Apple used in-house chips to reduce reliance on Intel and unify its mobile and PC ecosystems.
In 2024, even NVIDIA CEO Jensen Huang acknowledged the shift, stating: “Export controls on China have failed. Our market share in China has dropped from 95% to 50% in just four years.”
Strategic Battleground
According to a study by MacroPolo, the think tank under the Paulson Institute, nearly half of the world’s top AI researchers received their undergraduate degrees from Chinese universities, compared to just 18% from U.S. universities. In 2019, that figure for China was only 29%, highlighting the country’s remarkable progress in cultivating world-class AI talent.
Take DeepSeek, for example — a team composed largely of graduates from top Chinese universities such as Tsinghua and Peking University. The team exemplifies a new wave of AI talent: highly educated, younger, open-source driven, and innovation-focused.
Similarly, Chinese talent is playing a critical role in Elon Musk’s Tesla Robotaxi project. Notably:
Pengfei Duan, Tesla’s Chief Software Engineer for AI, earned his undergraduate degree at Wuhan University of Technology.
Charles Qi, the machine learning engineer behind Tesla’s Full Self-Driving (FSD) system, graduated from Tsinghua University.
These cases clearly demonstrate that China’s homegrown AI talent is now shaping global innovation at the highest levels.
Summary
In the global AI race, China is swiftly narrowing the lead, driven by two key strengths: seamless hardware–software integration and a rising wave of homegrown AI talent. The launch of Xiaomi’s AI glasses marks not only a milestone for China’s consumer AI hardware but also highlights its unique strengths in ecosystem integration and cost-effectiveness. Meanwhile, China’s momentum in open-source models and industrial robotics is actively reshaping the global AI landscape.
At the same time, the United States maintains a firm grip on the foundations of AI — core algorithms and semiconductor supremacy, particularly in high-end compute infrastructure and closed-source model ecosystems. The core of this technological rivalry has shifted from pure performance benchmarks to a full-scale competition of ecosystems and innovation models.
Looking ahead, as China continues to export top AI talent and deepen its industrial chain integration, the global AI landscape may be on the brink of a major realignment. As NVIDIA CEO Jensen Huang put it, “Technology blockades only accelerate innovation.” In this silent war of systems and scale, China is advancing its own AI era, driven by openness and grounded in real-world applications.
除了华为,小米也推出了自研芯片,主要目的是降低物料成本(参考苹果M1芯片对Mac产品线利润率的提升),并规避美国制裁风险(如华为遭遇的断供危机)。玄戒O1的推出意味着小米首次实现手机核心系统级芯片(System on Chip,简称 SoC)自主化,掌握芯片核心技术,实现从芯片架构设计到功能集成等关键环节的自主可控。未来可减少向高通支付专利费,提升毛利率。类似苹果通过自研芯片摆脱Intel依赖,实现从移动端到PC端的全生态整合。
根据美国保尔森基金会旗下麦克罗波洛智库(MacroPolo)的研究,从出身的本科院校来看,中国高校几乎培养了全球一半的顶尖 AI 研究人员,相比之下,仅有约 18% 研究人员来自美国大学。在2019 年,本科毕业于中国高校的顶尖 AI 研究人员占全球的比例还只有 29%,这一数据的大幅跃升体现了中国在 AI 人才培养方面的卓越成效。
Have you got your ticket to the AI-native world — the “Core AI Subscription”?
This concept was introduced by Sam Altman, CEO of OpenAI, at the 2025 Sequoia AI Summit. More than just a new term, it could become the central theme of AI commercialization in the years to come.
For investors, this signals a shift: the future winners won’t be the ones merely selling models, but those building long-term user relationships and platform-level capabilities.
For entrepreneurs, it offers clear direction: those who can create truly personalized AI assistants, and a sustainable business model around them, will be the ones to rise above the rest.
The “Core AI Subscription” is more than just a new business model; it may signal that humanity is accelerating toward a new era: the AI-native world.
What Is the AI-Native World?
The AI-native world refers to a future paradigm where artificial intelligence is not just a tool, but the foundational driving force deeply embedded in every aspect of society. It reshapes how we live, work, produce, and interact, much like electricity and the internet redefined previous eras.
In this world, AI becomes core infrastructure, pervasive and indispensable. It transforms human cognition, economic models, technological ecosystems, and social structures.
Imagine life in an AI-native world:
You no longer use AI just to complete isolated tasks.
AI understands your intentions, anticipates your needs, and acts before you even ask.
It’s seamlessly integrated into your life, work, and decision-making processes, becoming an essential part of your digital existence.
Core AI Subscription: The Gateway to the AI-Native World
Sam Altman’s concept of the “Core AI Subscription” is not just visionary — it represents the key pathway to realizing the AI-native future. It refers to a highly personalized, continuously evolving AI assistant service that is deeply embedded in users’ daily lives, much like an operating system that runs across every aspect of work and life.
This service is far more than just a voice assistant or chatbot. It functions as an intelligent agent with the following capabilities:
Personalized customization based on your habits and behaviours;
Seamless integration with other applications and services;
Constant learning and self-improvement over time, becoming smarter and more efficient the more you use it.
In other words, whoever owns the user’s Core AI Subscription essentially controls the “operating system entry point” to the AI-native world.
What AI Capabilities Are Required for Core AI Subscription Services?
To bring “Core AI Subscription” to life, there’s a key question we might want to explore: What kind of AI is capable of supporting such a service?
OpenAI has proposed a tiered framework for Artificial General Intelligence (AGI) — AI systems with the ability to learn efficiently, generalize across tasks, and act autonomously in complex, dynamic environments. True AGI would possess a blend of perception, cognition, decision-making, learning, execution, and social collaboration, all while aligning with human emotions, ethics, and moral standards.
Here’s a breakdown of the AGI capability tiers:
Level 1: Chatbot — Basic conversational ability, like current GPT models.
Level 2: Reasoner — Can solve human-level problems — mathematics, logic, coding, and debugging.
Level 3: Agent — Acts on behalf of the user — booking travel, managing calendars, and automating task chains.
Level 4: Innovator — Capable of invention and creativity — designing new products, writing screenplays, composing music.
Level 5: Organizer — Manages teams, coordinates resources, sets strategies, and even runs companies.
What AGI Level Is Needed to Enable Core AI Subscription Services?
To bring Core AI Subscription services to life, the AI must reach at least Level 3 — Agent on the AGI scale. At this level, AI isn’t just passively responding to user commands — it must actively understand user needs, take initiative, trigger tools, execute task chains, and switch contexts fluidly across various scenarios.
Since 2023, Baidu founder Robin Li has echoed a similar vision, stating that “large models will usher in a flourishing ecosystem of AI-native applications.” He emphasized that AI-native applications are not simple replicas of mobile apps or desktop software — they are meant to “solve problems that were previously unsolvable or poorly solved.”
This vision aligns closely with the concept of Core AI Subscription: true AI-native products are those in which AI agents are deeply embedded in users’ lives and workflows as a systemic, always-on digital partner.
Open Evidence: Core AI Subscription in Action in Healthcare
Are there early pioneers building AI-native applications? Yes—and a standout is Open Evidence, a medical AI company founded in 2021. By February 2025, it had raised $75 million from Sequoia Capital and achieved unicorn status with a valuation surpassing $1 billion.
At the 2025 Sequoia AI Summit, co-founder Zach shared a real-world case showing how their Core AI Subscription model supports physicians:
Emergency In-Flight Medical Case
Dr. Susan Wilberg faced a medical emergency mid-flight: a 63-year-old male cancer patient on immunosuppressive therapy developed a severe rash. Suspecting shingles, she had to make a critical call—should the plane turn back? What immediate actions were needed onboard?
She turned to ClinicalKey AI, Open Evidence’s subscription-based platform built for medical professionals. It delivered instant, personalized guidance by combining:
CDC Yellow Book protocols,
The latest research on cancer immunotherapy, and
Patient-specific recommendations (based on age, history, treatment, etc.).
The platform:
Assessed the patient’s risk level given his immunosuppressed condition,
Offered specific and timely treatment guidance,
Helped avoid an unnecessary emergency landing while ensuring proper care upon arrival.
What Makes ClinicalKey AI a True Core AI Subscription?
Open Evidence’s AI assistant is more than a diagnostic aid—it functions like an intelligent agent that continuously learns, personalizes its output, and proactively supports users:
Hyper-personalization: Tailors suggestions based on user preferences and patient context.
Seamless integration: Connects effortlessly with existing medical systems and workflows.
Continuous evolution: Becomes smarter and more efficient through real-world interactions.
Business Model & User Growth
Over 25% of U.S. practicing physicians now rely on Open Evidence daily. The system handles more than 10 real-time clinical questions per second. While the service is free for doctors, revenue comes from medical device and pharmaceutical advertising, mirroring consumer internet models, but adapted for healthcare.
To deepen value and retention, Open Evidence is embedding top physicians’ expertise, starting with gastroenterology, into its AI, creating a collective intelligence layer. This not only strengthens its data advantage but enables constant answer refinement.
The Future: AI as an Indispensable Partner
Looking ahead, Open Evidence plans to integrate broader medical reasoning, research capabilities, and workflow tools to build a fully-fledged Core AI Subscription platform, ultimately becoming a mission-critical partner to doctors worldwide.
If you’re a business owner aiming to integrate an AI subscription model into your operations, here are three essential principles to keep in mind:
1. Shift from “function thinking” to “companionship thinking.” Don’t just ask, “What can AI do for my business?” Instead, consider, “What do my users need AI to become?” A CFO doesn’t simply need a reporting tool—they need a proactive financial partner that can anticipate risks and guide decisions.
2. Capture high-frequency “scene entry points.” Identify must-have, recurring scenarios—such as in healthcare, legal services, or vertical workflows—and embed AI deeply into those daily user moments. Your goal is to make AI a seamless, indispensable part of how users work.
3. Build a “subscription-based emotional account.” Offer consistent, meaningful value—like weekly personalized insights—to create a sense of FOMO (fear of missing out). When users feel your AI is essential to staying ahead, loyalty follows naturally.
By applying these strategies, you can turn AI from a one-off tool into a trusted, subscription-powered companion that users depend on.