The Survival Dilemma of Small and Medium Banks: Is Not Embracing AI Waiting to Die, and Embracing It Seeking Death?

In the dead of night, I always have the habit of brewing a strong cup of tea for myself, sitting alone in my study, leafing through those heavy financial audit reports carrying the scent of paper and ink. As a veteran who has navigated the financial circles for thirty years, it has been a long time since I felt a genuine shudder from a set of data. But just a few days ago, when I finished compiling the industry reshuffle data from the short dozen months between 2025 and 2026, the hand holding my cigarette actually paused in mid air for a long time.

My friends, we are in an extremely grand yet unusually cruel period of financial geographic reshaping. According to the latest report released by the Shanghai Institution for Finance and Development at the beginning of this year, in January 2026, a total of 494 small and medium-sized banking institutions across China were quietly deregistered due to mergers or dissolutions throughout the entire year of 2025. This included 310 rural bank branches and 160 rural commercial banks. What is even more suffocating is that in just the first half month of 2026, another 65 rural banks were approved to vanish, completely wiped off our financial map.

In the official grand narratives and press releases, this is called the reducing of quantity and improving of quality of financial institutions. It is a forward looking layout by regulators to prevent and defuse financial risks. This is, of course, a correct macroeconomic tone, representing the Balancing Force at work to maintain the order of the entire economic ecosystem. However, if we lower our perspective from the macroeconomic sky down to the micro level dirt, if we shine a spotlight on these over five hundred vanished individual banks and flip through their financial statements from the last two years before they were merged, you will discover an extremely bizarre anomaly that completely contradicts business common sense.

Driven by professional sensitivity, like a detective facing a complex crime scene, I meticulously extracted the balance sheets and income statements of over a dozen rural commercial banks and village banks that had deeply cultivated county level markets and originally had decent asset quality. I cross referenced their data from 2022, 2023, and the eve of their collapse in 2024.

At first, everything looked normal. Under the overall environment where the economic cycle faced the Contraction Force, the net interest margins of these small banks were narrowing, profits from traditional deposit and loan businesses were sliding, and non performing loan ratios were slowly climbing. These are all normal physiological reactions when the weather turns cold. However, when my eyes slid down to the information technology expenditure under the operating and administrative expenses category on their income statements, a glaring combination of numbers suddenly leaped into my sight.

In the one to two years before these dozen or so banks went bankrupt or were forcibly restructured, while their main business revenue had already stagnated or even shrunk, their information technology expenditure actually showed an explosive growth of three hundred or even five hundred percent. Some county level rural commercial banks, whose original annual net profit was a mere thirty to fifty million, continuously poured tens of millions of huge funds every year in 2023 and 2024 to purchase smart risk control systems and cutting edge large model technologies that claimed to offer instant approval and instant lending with natural language processing capabilities.

At this point, if you are observant, you will certainly notice the massive paradox here. Logically, armed to the teeth with the most advanced technological weapons, these small banks should have been invincible in the lower tier markets, like donning a super exoskeleton from a sci fi movie. Why did these banks, which were the most willing to spend money, the most eager to embrace artificial intelligence, and the most desperate to grasp the Evolutionary Force of the era, become the first cannon fodder to fall and die the most tragically in this massive financial reshuffle?

This leads to the life and death puzzle we are going to deeply dismantle today. In the current climate of the times, in 2026 when the tide of artificial intelligence is sweeping the globe with the momentum of a landslide, for small and medium sized banks, not embracing artificial intelligence seems like waiting to die like a frog in boiling water; but blindly and panickily embracing artificial intelligence is absolutely a suicidal acceleration of their own destruction.

To uncover the truth of this survival dilemma, we must temporarily push aside those dry financial numbers and, like a masterful psychologist, penetrate the inner world of the executives of those small and medium banks. We must peek into the Prism of Human Nature refracting in their eyes and understand that massive survival anxiety capable of making people lose their reason.

Let us turn the clock back to 2023 and 2024, a period that made all county level bank presidents feel suffocated. At that time, national macroeconomic controls, in order to stimulate the real economy, guided capital interest rates continuously downward. During this process, those large state owned banks, possessing astronomical amounts of cheap capital and massive national data networks, began a cruel downward market penetration movement.

You can imagine those large state owned banks as ocean going super trawlers equipped with the most advanced sonar systems. They sailed into the coastal waters that originally belonged to small and medium banks and cast extremely precise digital fishing nets. Utilizing loan interest rates as low as three percent or even lower, coupled with the ultimate experience of instant one click approval on mobile banking apps, they mercilessly swept away the highest quality customers in the county economy, those local leading enterprises and premium small micro merchants with good tax records and healthy cash flows, picking them off like cherry picking.

Facing this dimensional strike, those rural commercial banks and village banks, accustomed to manual bookkeeping and relying on the two legs of loan officers running through streets and alleys to solicit deposits and issue loans, were like traditional fishermen rowing wooden boats with hand woven, torn nets. They watched helplessly as the high quality customers they had painstakingly cultivated for over a decade were snatched away by large banks with a few taps on a smartphone, while they were completely powerless to fight back.

This fear of helplessly watching their territory being eroded spread rapidly among the executives of small and medium banks, evolving into a deep, collective technological anxiety. Driven by the instinct to protect their turf and their life’s work, the Expansion Force deep within them, yearning to break out, was extremely distorted and magnified. They desperately craved to find a shot of technological adrenaline that would yield immediate results, fantasizing that as long as they bought the same smart systems as the big banks, they could fend off this terrifying invasion.

There is an extremely profound saying in Zhuangzi. Those who lose themselves in material things and lose their true nature in worldly trends are called inverted people. This means that if a person loses their true self in the pursuit of external material things and loses their inherent nature to cater to secular trends, they are a pathetic person who has put the cart before the horse. And in those two years, countless small and medium banks, precisely because of their fanatical worship of technology and extreme fear of death, became the inverted people in the business world.

Right at this moment, salesmen from financial technology companies, dressed in suits and speaking eloquently, knocked on the office doors of county level bank presidents with carefully crafted presentations. They exploited information asymmetry and precisely grasped the survival anxiety of these presidents. They packaged extremely expensive general large models or black box risk control systems, which were originally designed for massive internet platforms with tens of millions of users, into omnipotent life saving elixirs, selling them at high prices to these weak bottomed small banks.

I once personally witnessed such a sigh inducing scene. A few years ago, an old acquaintance of mine, the chairman of a rural commercial bank in a major agricultural county in the central region, mysteriously invited me to visit the smart risk control command center they had just spent a fortune building. It was a massive server room full of a sci fi vibe, taking up an entire floor. On a giant LED screen, various complex light dots and real time charts representing the flow of funds across the county were flashing. This veteran financial worker in his fifties, who started from a grassroots credit cooperative, stood in front of the screen. His eyes flashed with a fanatical yet slightly guilty light as he endlessly introduced to me the machine learning algorithms and big data knowledge graphs they had purchased.

But I knew that behind that dazzling screen was a massive money swallowing beast that was frantically devouring the bank’s profits. Blindly purchasing technology systems is never a one off deal; it is a bottomless pit. The high initial system integration fee is just the tip of the iceberg. Next, to keep this behemoth running, they had to pay exorbitant computing power leasing fees, cloud service fees, and various external data interface calling fees every year. More fatally, to serve these complex systems, they had to poach a group of so called IT elites from big cities at salaries far exceeding the local average wage.

For a county level bank whose original net profit was already meager, this blind technological leap forward, detached from its own asset scale and profitability, directly led to a severe loss of cost control. This is a highly typical Expansion Force overload. They intended to use technology to reduce costs and increase efficiency, but as a result, the technology itself became the heaviest stone that crushed their balance sheets.

But this was only a disaster at the financial level. If we apply the span dimension and nexus dimension from our five dimensional phase transition law, cutting our perspective into biology and ecology, you will discover a far more fatal organ rejection reaction that dismantled the bank’s core competitiveness from within.

We must deeply understand that in China’s counties and rural areas, what has formed over thousands of years is a business ecosystem based on an acquaintance society. In this ecosystem, credit is not a tax certificate or a credit report printed on A4 paper, but an extremely complex soft information full of human worldly wisdom.

When an experienced old loan officer judges whether the boss of a brick kiln in town can get a loan, he often doesn’t look at the financial statements provided, because in that environment, statements are often highly distorted. What does the old loan officer look at? He looks at the cash gifts given by village officials and fellow villagers when the boss’s family hosts weddings or funerals; he looks at whether the boss’s wife is frugal or extravagant when buying vegetables at the market; he looks at whether the smoke rising from the brick kiln’s chimney in the middle of the night is thick or thin. This trust network, rooted in rural human relations and interwoven from thousands of tiny details, is an extremely effective evolutionary stable strategy that humanity evolved during a long agricultural civilization.

This is the true, and only, moat for small and medium banks to fight against large state owned banks. Because no matter how advanced the big banks’ radar is, it cannot scan the gossip of the aunties chatting under the big tree at the edge of the village.

However, when these small and medium banks spent huge sums to forcefully stuff a cold AI brain, trained on standardized and formatted hard data, into the body of this acquaintance society full of the smell of dirt, disaster struck.

This was exactly like performing an organ transplant with completely mismatched blood types. This high end smart risk control system simply could not read or understand that soft information concerning human nature. It would only mechanically demand networked tax data and standard commercial cash flows from enterprises. The result was that those truly hardworking local manual workshops and large scale farmers, whose financial records were extremely non standardized, were coldly shut out by this algorithm; while those opportunists who were deeply versed in packaging themselves, spending money to hire intermediaries to create perfect data flows, could easily deceive the AI’s review and blatantly take away large loans.

Yanzi left a famous quote through the ages in the Spring and Autumn Annals of Master Yan. Oranges grown south of the Huai River are true oranges; once transplanted north of the river, they become bitter枳. The leaves are similar, but the taste of the fruit is different. Why is this so? The water and soil are different. Those top tier algorithms proven effective in the skyscrapers of Silicon Valley or Wall Street, once detached from the soil of massive standardized data they rely on and forcibly transplanted into the rough water and soil of China’s county economy, not only fail to bear sweet fruit but turn into bitter fruit full of toxins.

When the rural risk control network originally built by walking on two legs was completely abandoned and replaced by a mechanical algorithm that was completely out of touch with reality and unacclimated to the local environment, the bank essentially destroyed its own Great Wall. Under the cover of technological fanaticism, they seemingly put on the most advanced AI masks, while internally, their vital organs had long been hollowed out by this violent rejection reaction.

Telling the story up to here, we have only peeled back the first layer of clothing on these death specimens of small and medium banks. Cost loss of control and environmental unacclimatization would certainly severely damage the vitality of these small banks, but it wouldn’t be enough to cause hundreds of institutions to die suddenly and collectively within a short year or two.

What truly pushed them into the abyss of eternal doom was the cup of deadly poison they were forced to swallow in order to quickly dilute the massive technological costs after introducing these expensive systems. Hidden in that cup of poison was the most secretive and fatal logical trap of the financial industry in the era of artificial intelligence.

Please adjust your breathing slightly. In the following time, we will continue holding our magnifying glass, diving into the deepest part of this technological maze, to see how those once glorious county level banks unknowingly sold their souls to external algorithms, ultimately reducing themselves to empty shells whose bodies had been completely hijacked. Wang Anshi once wrote in a poem, We fear not the floating clouds obscuring our vision, for we ourselves are at the highest peak. Only by penetrating this layer of floating clouds woven by code and computing power can we clearly see the true hidden cards behind this life and death game.

Let us follow the detective perspective just now and continue digging deeper into this financial maze. After dismantling the cost loss of control and environmental unacclimatization caused by small and medium banks blindly purchasing technology systems, we have actually only seen their first step toward death. What truly sounded the death knell for these over five hundred institutions was a cup of deadly poison they were forced to drink under extreme anxiety and massive financial pressure.

This cup of poison has a very glamorous name in the financial circles: internet joint loans based on big data.

Imagine for a moment, if you were the president of a county level rural commercial bank. Fooled by technology companies, you threw in the entire bank’s net profit of two full years, tens of millions or even hundreds of millions of real money, to purchase an artificial intelligence large model risk control system that claimed to offer instant approval and instant lending. Massive sunk costs had already been incurred, and you still had to pay expensive computing power leasing fees and system maintenance fees every year. At this moment, what is your most urgent need?

It is scale. It is the absolute necessity to expand the asset scale in an extremely short period to dilute the massive technological costs.

However, the county you are located in has at most a few hundred thousand people, and the high quality customers have long been cherry picked by the large state owned banks with extremely low interest rates. The remaining small and micro merchants and farmers simply cannot digest the massive production capacity brought by your expensive system. This is like spending a fortune to buy the most high end lithography machine, only to use it to forge kitchen knives in the village blacksmith shop. This extreme mismatch of capacity would instantly crush your balance sheet.

Just when you were as anxious as an ant on a hot pan, the financial technology company that originally sold you the system, or the internet traffic giants behind them, opportunely appeared once again like saviors. Smiling, they offered you a seemingly flawless solution: President, you have a financial license and cheap deposit funds, while we have massive national internet traffic and countless young consumers eager to borrow money. More importantly, we have the most advanced artificial intelligence algorithms. How about this, we connect our interfaces and launch joint loans. We handle online customer acquisition and use our large models to risk score the customers. You just need to sit in the back office and issue the loans with your eyes closed. We split the profits, and you use the returns to cover the bad debt risks.

Under that extremely twisted mentality, desperately craving expansion and desperately craving to save their financial statements, hardly any helmsmen of small and medium banks could resist this temptation. In their eyes, this was simply a technological feast where they could reap the rewards without having to go down to the fields to work themselves.

Thus, an unbelievable magical realism unfolded. A township bank that was originally only capable of serving a few hundred local villages saw its asset scale balloon like a blown up balloon from five billion to twenty billion in just one year. Through the pipelines of the internet, they frantically loaned the hard earned money they had gathered penny by penny from local commoners, in the form of joint loans, to strangers thousands of miles away whose faces they had never even seen.

Sima Guang left a deafening warning in the Comprehensive Mirror in Aid of Governance. The clear sighted see danger before it takes shape; the wise see disaster before it sprouts. A truly wise person can instantly see through the destructive crisis deeply hidden in the invisible amidst this seemingly flourishing, wild surge. But the executives of those small and medium banks at the time had already been blinded by the surging financial numbers in front of them, completely unaware that what they had signed was not a cooperation agreement, but an indenture selling their souls.

In the first principles of finance, what exactly is the essence of a bank? A bank is not a simple money mover. The core business model of a bank is managing risk. If you absorb deposits, you must bear the risk of lending that money out. This requires you to possess absolutely independent risk pricing power.

However, in this so called technological empowerment and joint lending, small and medium banks actually committed an extremely fatal act, which was risk control outsourcing. They handed over the core approval power deciding whether a loan could be issued, how much to issue, and how high to set the interest rate entirely to the black box algorithms of external technology companies. They abandoned their brains and voluntarily degenerated into soulless shells whose only responsibility was to provide funds.

This was not only a mutation of the business model but an extremely unequal, even malicious, fatal game.

We must use an extremely cold gaze to examine both sides of this game. What were the external financial technology companies pursuing? It was the monetization of traffic, the service fees and commissions from single play games. The core optimization goal of their algorithm models was the approval rate and the lending volume, because only when the scale was large enough could they earn enough toll fees. They did not bear the ultimate credit default risk.

But as the funding party, the small and medium banks bore the real, solid credit risk. They were the ones who had to use real money to fill the bad debt losses.

When the macroeconomy was in an upward cycle, that is, when Expansion Force filled the entire market, everyone had jobs, borrowers were borrowing new money to pay off the old, and this game of passing the parcel seemed to run incredibly smoothly. The data generated by the AI algorithms in the background looked extremely beautiful, and the non performing loan ratio was astonishingly low. The tech companies made a fortune, and the banks’ statements glittered with gold.

But, my friends, the Contraction Force of economic laws, just like the alternation of the four seasons in nature, is an objective force that can never be resisted. When the economic winter of 2024 and 2025 truly descended, when the macro Contraction Force ruthlessly squeezed every corner of society, when those young people borrowing money on the internet faced unemployment and sharp drops in income, that seemingly perfect algorithmic myth collapsed in an instant.

The large models of those external tech companies were often based on borrowers’ behavioral data on the internet, such as online shopping frequency, food delivery preferences, and ride hailing records, to assess risk. On sunny days, this data might serve as a reference for willingness to repay. But in the freezing winter, when a borrower couldn’t even afford next month’s rent, the behavioral data of how many branded sneakers he had bought in the past could not translate into even a single cent of repayment ability.

Bad debts do not appear one by one; they come surging in like a tsunami. When disaster struck, those external tech platforms immediately initiated their escape protocols. They modified agreements, cut off data interfaces, and even directly announced bankruptcy liquidation, dumping all the massive messes and astronomical bad debts onto the small and medium banks that acted as the funding parties.

At the end of 2025, I participated in the asset liquidation work of a rural commercial bank on the verge of bankruptcy. When I opened their core credit system, I felt a suffocating despair. Up to eighty percent of the loans were issued to strange customers several provinces away. There was no collateral, no real business scenarios, and even the effective contact information of the customers was forged or desensitized by the tech companies. The loan officers of this bank sat at their desks facing thousands of names that had turned red due to overdue payments on their screens, unable to get through on a single collection call.

Han Feizi contains an extremely ruthless principle of imperial statecraft. The sword of Tai’e must not be lent to others. Tai’e was an ancient sword symbolizing supreme power. Han Feizi warned the world that core power must never be lent to others. Once lent out, not only will you become a puppet, but it will also invite your own assassination. For a bank, independent risk control capability is its life saving sword of Tai’e. Those small and medium banks, in pursuit of short term scale and an illusory technological halo, handed this sword over to external algorithmic platforms, and were ultimately slashed across their own jugulars by this very sword.

To more profoundly understand the devastation brought by this loss of core control, we need to use the nexus dimension in our five dimensional phase transition law to conduct an interdisciplinary topological folding, turning our gaze to the vastness of astrophysics.

In the universe, black holes possess extremely terrifying gravity. At the edge of a black hole exists an invisible boundary known as the event horizon. Once any matter, even light, the fastest thing in the universe, crosses this event horizon, it can never escape being swallowed by the black hole. All its information and physical structure will be completely destroyed at the singularity of the black hole.

The moment small and medium banks outsourced their core risk control to black box algorithms, it was equivalent to voluntarily crossing the event horizon of system collapse in this cruel commercial competition.

Before crossing this horizon, although this small bank was weak and not smart enough, it possessed its own social capital. It was deeply rooted in the locality. Its president knew every entrepreneur in town, and its loan officers knew which farmer’s orchard had a good harvest this year. This trust network based on geography and human relations is an extremely resilient anti risk structure.

However, when it connected to joint loans, when it relied entirely on external algorithms to determine the flow of funds, the soft information and social capital it had originally accumulated over decades were instantly torn apart by the massive gravity of external internet platforms, just like a star falling into a black hole. In astrophysics, this is called spaghettification. During the process of being swallowed by a black hole, a star is stretched by extremely unbalanced gravity into long, thin strands of spaghetti until it shatters.

The capital and credit systems of small and medium banks were thoroughly spaghettified under the pulling of this algorithmic black hole. They abandoned the familiar land beneath their feet and extended their tentacles into the completely uncontrollable abyss of the internet. When risks broke out collectively, they found they had lost all their leverage. They could not conduct on site collections from borrowers far away on the horizon, much less mobilize the moral pressure of a local acquaintance society to force borrowers to repay.

The false prosperity forcibly manufactured by algorithms was incredibly fragile in the face of true human games and the Contraction Force of macro cycles. The core capital of the banks was rapidly pierced through, and they ultimately could only head down the dead end of being merged or declaring bankruptcy.

The disappearance of over five hundred banks is by no means just a string of cold statistics. Behind it are the shattered careers of countless financial practitioners and the tragic price of local financial ecosystems being uprooted. The death specimens beneath this technological carnival prove to us an unquestionable truth at an extremely tragic cost. In the age of artificial intelligence, no matter how advanced a tool is, it is still just a tool. If an enterprise, especially a financial institution managing risk, attempts to use the purchase of tools to replace its own thinking about business essence, and attempts to use cold code to take over complex moral and human games, what it welcomes will absolutely not be evolution, but thorough destruction after its soul has been stripped away.

With this detective style dismantling reaching this point, the truth of the case is now completely exposed to the world. Small and medium banks died from fanatical worship of the Expansion Force, died from lightly abandoning core risk control, and died from arrogance and ignorance in front of the algorithmic black hole.

However, as a financial veteran always seeking the optimal solution amidst uncertainty, our deduction must absolutely not simply stop at the ruins of despair. After analyzing the deep logic of why embracing AI means seeking death, we must immediately turn the wheel and search for that path of evolution, low in cost and high in efficiency, that can truly cross this survival dilemma.

Because no matter how tragic this reshuffle is, the small and micro enterprises and farmers serving China’s vast lower tier markets still need someone to serve them. After the great waves wash away the sand, those small and medium banks lucky enough to survive, or those about to be reborn through restructuring, must undergo a profound cognitive phase transition to rediscover their own survival coordinates in this new world enveloped by large models. In the following third part, we will completely clear away the fog and provide that breakthrough solution capable of perfectly stitching together hardcore technology and rural human warmth. Please take a short break, and we will immediately enter the most core territory of this ideological experiment.

When this brutal financial storm swept past, and the wreckage of over five hundred small and medium banks lay quietly in the historical footnotes of 2025 to 2026, what posture should we survivors adopt to face that already irreversible era of artificial intelligence?

Through this detective style dismantling, we have already seen the fatal causes of death clearly. Those fallen banks died from extreme greed for Expansion Force, died from having their blood drained by expensive technological blind boxes, and died even more from the dimensional strike encountered after outsourcing core risk control to black box algorithms. But if we merely stop at autopsying the dead, we fail the cruel revelation brought by this massive reshuffle.

As a veteran who has kept watch on the risk control frontline for thirty years, I deeply know that when an ecosystem is on the verge of collapse due to the disorderly spread of Expansion Force, the Balancing Force representing order and correction will inevitably descend with thunderous momentum. The over five hundred mergers and dissolutions of small and medium banks orchestrated by regulators over the past dozen months are precisely the concrete manifestation of this macro Balancing Force. Using an iron administrative fist, it forcibly severed those wildly growing internet joint loan chains, brutally yanking back down to earth those small and medium financial institutions attempting to probe madly at the edge of the algorithmic black hole.

For those rural commercial banks and village banks that luckily survived this great purge or just gained a new life through restructuring, an extremely severe life or death choice lies before them. The cherry picking plundering by large banks has not stopped, and the evolutionary frenzy of artificial intelligence will absolutely not press the pause button because of your fear. Retreating to the primitive era of manual bookkeeping is waiting to die; blindly buying expensive general large models is seeking death. Then, in this seemingly completely blocked dead end, where is the real lifeline?

To answer this question, we must activate the core dimension in the five dimensional phase transition law and complete a thorough cognitive leap. We must redefine what truly penetrating Evolutionary Force is in county economies and lower tier markets.

The opening of the Great Learning states, Things have their roots and branches. Affairs have their ends and their beginnings. To know what is first and what is last will lead near to what is taught in the Great Learning. This sentence pierces the underlying logic of the operation of all things in the world. For a commercial bank, especially a small or medium bank deeply cultivating local areas, what is the root? The core risk pricing capability, based on rural roots and insight into human nature, is the root. What are the branches? Those dazzling system interfaces, massive computing clusters, and complex code models are the branches.

Those dead banks committed precisely the fatal error of putting the branches before the roots. They delusionally tried to use the branches they bought at a high price to replace the roots they relied on for survival. They thought that spending a hundred million to buy computing power could buy the ticket to evolution, not knowing that true Evolutionary Force is not spending a hundred million to buy machine computing power, but spending one million to arm the cognition of employees and reconstruct an exclusive risk control system deeply integrating human and machine.

Let us cut our perspective into specific execution paths and see how a smart, awakened small and medium bank should win this turnaround battle in 2026.

First, please immediately stop the superstition regarding those general large models costing tens of millions and external black box risk controls. In this era where open source technology is highly developed, computing power and foundational models are rapidly depreciating, becoming basic infrastructure as cheap as tap water. There is absolutely no need for small and medium banks to bear such suffocating tech R&D costs. What you truly need to do is to embrace those low cost, even open source, Small Language Models.

These SLMs do not require massive server rooms and can even be deployed on local lightweight servers. Their task is not to predict global macroeconomic trends, nor is it to write macro analysis reports filled with flowery rhetoric. Their task is extremely singular and vertical: to serve as the enhanced exoskeleton for loan officers, processing standard hard data in county economies with high concurrency and low cost.

Imagine such a real future work scenario. In a vast agricultural county, when a loan officer faces a large scale farmer with thousands of acres of contracted land, he no longer needs to spend days manually checking dozens of pages of bank statements and invoices as in the past. This lightweight AI assistant deployed on a tablet can, within seconds, through legal and compliant interfaces, retrieve satellite remote sensing multispectral images of this land, precisely calculating crop growth and estimated yield; it can instantly cross reference data from tax systems and agricultural supply purchasing platforms, telling you how much fertilizer this customer bought and how much electricity he paid for this year.

With the assistance of this highly cost effective technological exoskeleton, the machine tirelessly completes ninety nine percent of the data cleaning and logical cross verification work, and the cost is so low it can almost be ignored.

But, my friends, please note that this is absolutely not the end of risk control. This is exactly the starting point for small and medium banks to showcase their true core moats.

When the machine pushes a data report with zero logical flaws in front of the loan officer, the real test has just begun. Because in this county level business ecosystem composed of an acquaintance society, what decides whether a loan can ultimately be recovered is often not those few invoices, but the Prism of Human Nature hidden behind the invoices.

At this moment, that old loan officer who grew up in this county, drinks the local tea, and speaks with an authentic local accent, has his value irreplaceably highlighted. He doesn’t need to sit in the office staring blankly at a screen. He will close his laptop, walk to the fields of that large scale farmer, offer a cigarette, and chat about everyday family matters.

He will use tiny details undetectable by machines to gain insight into the soul of this customer. He will see whether this boss’s attitude toward hired farm workers is harsh or generous; he will ask around at the town’s convenience store whether this boss, in years of natural disasters and poor harvests, would rather borrow money everywhere himself than delay paying the land transfer fees to his fellow villagers.

If the default probability calculated by the machine is ten percent, but this loan officer, through a deep scan of the customer’s Prism of Human Nature, confirms that this is a person who treats credit as life and possesses extremely high moral prestige in rural society, then he has the confidence to sign his name on the approval form, using his personal professional judgment to override the machine’s cold conclusion. Conversely, if the machine considers this a golden customer with perfect qualifications, but the loan officer keenly senses at a wedding or funeral banquet table that this customer has acquired the bad habit of underground gambling, he will unhesitatingly cast a veto.

This is the ultimate weapon, and the only one, that small and medium banks can use to fight against big banks and internet giants: perfectly stitching together hardcore data penetration with human warmth and rural insight. Machines handle complex calculations and prevent low level fraud, while those risk experts full of human touch judge the moral bottom lines and survival resilience of business owners. This is a wisdom based on a long term evolutionary stable strategy that transcends single play games.

As the story progresses to this point, some bank executives in front of the screen might fall into deep distress. I understand the reasoning. Abandon black boxes, embrace open source, human machine collaboration, return to rural roots. These strategic directions sound absolutely correct. However, in actual operation, the IT staff in our bank do not understand the complex logic of the credit business at all, while our experienced credit veterans get dizzy looking at code and algorithms. Between the two lies a massive, uncrossable chasm. How on earth can we land this ideal system?

This is precisely the most painful realization for countless small and medium financial institutions during the transition period. What they lack is never the most cutting edge code, because code can be bought; what they lack is a Cognitive Architect who can perfectly stitch together the physical attributes of technological tools with the human game underlying banking business.

This is also why, over the past two years, I and the Finsages team have been dedicated to an extremely arduous yet highly valuable consulting endeavor. We go deep inside those confused small and medium banks, playing exactly this role of the Cognitive Architect.

We do not sell expensive servers, nor do we peddle bottomless black box large models. What we do is help these county level banks, at a fraction of the industry’s average cost, translate the unspeakable soft information of the acquaintance society they have accumulated over decades into risk parameters recognizable by lightweight algorithms.

We take risk control veterans by the hand, teaching them how to train that exclusive AI assistant; we take the IT team deep into unfinished buildings and breeding farms, letting them smell the real scent of the earth, letting them understand that at the end of the code are living, breathing humans. We help these institutions, in the extremely involuted lower tier markets, build an exclusive risk control system that possesses both the sharpness of modern technology and the warmth of traditional human relations.

This is not just a technological upgrade; this is a profound organizational genetic recombination. When you liberate loan officers from heavy form filling work, endow them with a powerful data exoskeleton, while extremely respecting and amplifying their insight into human nature, this bank truly possesses negative entropy assets to resist any macro Contraction Force. When facing the frantic downward penetration of large banks, you are no longer defenseless wooden boats, but light frigates equipped with precision guided radar and extremely familiar with local hidden reefs and channels, truly achieving overtaking on a curve in a desperate situation.

This is a long and arduous trek. In this fanatical era that readily talks about disruption and large models dominating the world, doing small and micro finance and serving lower tier markets is inherently a tough job that requires bending down and getting covered in mud.

County level banks and village banks are the terminal capillaries of China’s real economy. You connect with countless farmers toiling under the scorching sun, with early morning food stall owners who open doors and light fires before dawn, and with small factory bosses whose hair has turned white worrying about paying their workers. These real lives carrying sweat, tears, and hope are things no supercomputer can precisely calculate, and no expensive black box system can replace.

Precisely because machines are too smart, too cold, and too obsessed with extreme efficiency, your hard work and sweat rooted in the soil, your deep affection for the villagers and this land, appear so irreplaceable and so brilliantly radiant.

In the twenty fourth chapter of the Tao Te Ching, there is a proverb that awakens the world. He who stands on tiptoe is not steady; he who strides cannot maintain the pace. Standing on tiptoe, wanting to follow the technological frenzy of large banks, is destined to be unsteady; taking large strides beyond your own capabilities, wanting to blindly develop beyond your own data foundation, is destined not to go far.

In this life and death game that washed away five hundred banks, I hope every surviving financial practitioner can engrave these words in their heart. In the fanatical algorithmic torrent, maintain a rare sobriety; in the cruel lower tier game, hold fast to that warm territory. Do not look up at the floating clouds that do not belong to you; lower your head and deeply cultivate the land beneath your feet. Arm your hands with low cost technology, and nourish your soul with profound cultural heritage. Only in this way can you stand as steady as a rock and thrive endlessly amidst the stormy waves of cyclical changes.

Thank you for accompanying me late into the night to dismantle this thrilling financial life and death game. Between cold data and fiery human nature, there is always a path full of wisdom waiting for us to explore. If you are also experiencing the growing pains of transformation, yearning to find that guiding beacon in the fog, and yearning to reconstruct your core moats at an extremely low cost, you are welcome to follow Finsages.org. I am financial veteran . Let us walk side by side in the sea of stars of cognition. See you next time.