文章 · 存档
在2031年的时间镜子里,企业能不能撑到适应AI的那一天
如果有一面来自未来的镜子,能让我们提前看到2031年企业的生存样子,大概没有哪位管理者会拒绝看一眼。
这份由Gartner发表的《Over 100 Data, Analytics and AI Predictions Through 2031》(2031年数据、分析与 AI 百余项预测)就是这样一面镜子,只不过它没有耸人听闻的标题,也没有颠覆革命这些花哨的词,反而用了一种冷静甚至有点残酷的方式在提醒:接下来这几年企业真正要面对的不是要不要用AI,而是还能不能撑到适应AI的那一天。
这份报告一开篇就在打破一个很多中国企业仍然习惯的思维——五年规划。
作者直接说现在的AI和数据分析投资已经进入狂热期,但规划周期正在塌缩:五年的宏大蓝图很可能永远落不了地,真正可执行的只有一到两年,而且就算是18到24个月的计划也随时可能被突发的技术突破和外部变化打乱。
对于习惯写厚厚规划、层层汇报、缓慢落地的组织来说,这句话的含义其实很扎心:过去我们喜欢先弄清楚十年后要成为什么样子再往回推;而现在更现实的是你得先想清楚明后年不死再谈远期梦想。
这并不意味着不要远见,而是意味着远见不能再停留在PPT上。
报告提到数据与分析D&A职能在2025年前后会迎来一个关键拐点:首席数据与分析官CDAO会成为AI战略成功与否的关键角色,真正有话语权的将不再只是CFO、CTO,而是那个既懂数据又懂业务、能把技术和经营串起来的人。
但紧接着它又泼了一盆冷水,到2028年那些没有接受过专门培训、没有经验传承的第一次上任的CDAO里,将有75%在三年内离开岗位,重新审视自己是不是适合干这件事。
这句话背后其实藏着一个很现实的判断:AI和数据分析从来就不是给一个头衔、搭一个部门、招几个算法工程师这么简单。它考验的是一个人能不能在技术语言和经营语言之间来回切换,既理解AI的能力边界,又看得懂利润表、现金流和业务模式的细微变化。过去我们把会看财务报表的人送上决策桌,现在则要把能让数据真正驱动决策的人推到台前。而这种人目前在全世界范围内都是稀缺物种。
报告里有一个很打动我的细节:他们发现那些采用持续性、基于产品形态的资金模式来支持D&A和AI能力建设的组织,比那些只依赖年度一次性预算审批的组织多拿到了15%的资金。
换句话说在AI这件事上谁还在拿它当项目谁就已经慢了半拍。年度预算那套一年批一次、按项目给钱的玩法面对AI这种需要反复试错、持续迭代、随时更新的技术时,会变成一种严重的束缚。真正聪明的企业开始用产品思维看待内部的分析平台和AI能力,把它们当成企业长期资产,而不是一个个短期项目。
这点对中国的业财管理其实是个很大的启发。
我们过去在预算和资金使用上强调的是控制和可预见,而AI时代要求的是试验和快速重构。
你要的不是一次性砸钱搞个看上去很厉害的平台,而是允许团队围绕AI做一系列小而快的实验。今天在信用风控上试一试,明天在定价算法上做一做,后天在智能报表上再叠一层能力。不断踩坑、不断修正,把真正有效的能力养出来。
更有意思的是报告预测到2027年会有40%的组织重新重启知识管理,原因不是怀旧,而是被生成式AI倒逼。如果不把内部的知识、流程、经验系统化、结构化,AI根本喂不饱。 过去我们做知识管理很多是为了通过内审、拿认证,或者做给上级看。现在则是很现实,如果知识还散落在人的脑子、聊天记录和各种不规范文档里,AI就只能胡说八道。你希望AI少瞎编就得先把企业自己的知识和数据管好。
这里就绕回了一个很多人本能抗拒的话题:治理。
报告说得很直接:没有AI-ready data,AI的承诺就完全无法兑现,数据管理和治理不是可有可无的官僚流程,而是通往AI成功的必经之路,而且这些治理工作本身也可以在AI的帮助下提速。这听起来一点都不性感,但却是最诚实的现实,真正把AI用到经营里去的企业搞的不是炫技,而是打地基。
如果你把这份报告当作一封从未来寄来的信,会发现它没有告诉你哪一个模型会赢,哪家厂商一定胜出,它反复围绕的是几个极其朴素又极其要命的问题:规划不能再按老路来做,角色不能再按老逻辑去设,预算不能再按老节奏去批,数据不能再按老习惯去放任。用一句更直白的话说,过去我们用老制度驯服新技术,而从现在开始是新技术在倒逼我们重写制度。
这件事回到中国企业其实有一个非常现实的判断标准:如果你的AI项目讲出来很好听,但财务报表、管理驾驶舱、现金流预测、定价决策这些最关键的经营场景里看不到AI的存在,那就说明你离把AI当技术玩具还挺近,离把AI当经营基础设施还挺远。真正的突破不是多做了一份智能分析报告,而是在某个原本靠经验拍板的场景里敢让AI先给一个答案,再由人来做最后的判断和承担。
报告写到最后那段话,我觉得非常适合送给任何一个在纠结要不要上AI项目的管理者。
他们说现在的急迫感和曝光度使得所有人都在问一个问题:数据、分析和AI到底能不能“deliver the goods”,能不能真的交付结果。 这句话特别朴素,别再讲概念、别再谈远景,你到底能替企业多赚一块钱、少亏一块钱、提前躲过一次风险吗?能,就继续;不能,就别再烧钱造样子。
未来不是突然有一天砸在我们头上的,而是从今天开始每一次对数据的重视、对制度的微调、对AI的一点点尝试慢慢堆出来的。等到2031年回头看,你会发现活下来的企业未必是那些喊得最响的,但一定是最早认真对待这些提醒的那一批。
你现在看到的只是一份报告里的文字,真正重要的是你在看完之后愿不愿意在自己的企业里哪怕先改动一小步。那一步迈出去的时候,其实你已经不再只是预测未来的人,而是悄悄开始塑造未来的人了。
Can a Company Live Until It Adapts to AI
In a Time Mirror of 2031, Can a Company Live Until It Adapts to AI
If there is a mirror from the future that can let us see how companies live in 2031, almost no manager will say no to look at it.
This report from Gartner called “Over 100 Data, Analytics and AI Predictions Through 2031” is like that mirror.
But it has no scary titles and no loud words like “big change.”
It uses a calm and even a little bit “hard to hear” way to tell us:
In the next years, the real question for companies is not “Should we use AI or not,” but “Can we live long enough to get used to AI?”
This report breaks a way of thinking that many Chinese companies still like — the five-year plan.
The writer says AI and data work are already in a hot time, but planning time is getting shorter.
A big five-year plan may never happen at all.
Only one to two years can really work.
Even a plan for 18 to 24 months may be broken by fast new tech or sudden changes.
For companies that love making big plans, long reports, and slow actions, this is a painful message:
Before, we liked to first think what we will look like in ten years, then plan backward.
But now we must first think:
“How can we not die in the next one or two years?”
This does not mean we should not dream far.
It means far dreams cannot just stay on slides.
The report says: around 2025, D&A (Data & Analytics) will meet a key turning point.
The CDAO (Chief Data and Analytics Officer) will become the key person for AI success.
Not only CFO or CTO will have power.
The one who knows both data and business, and can connect tech and money work, will be the real leader.
But then, a cold truth:
By 2028, 75% of new CDAOs who never had training or help before will leave their job in three years and ask if they are right for this work.
This shows:
You cannot just give a title, build a department, and hire a few engineers to make AI work.
A good CDAO must speak the language of tech and also the language of money.
He must know what AI can and can't do, and must understand profit, cash flow, and business changes.
This kind of person is rare in the whole world.
There is a detail I like very much:
They found that companies that give money to AI in a long-term and product-style way can get 15% more money than companies that give money only once a year like a normal budget.
This means:
If you still treat AI like a short project, you are already slow.
AI needs trying again and again, fast updates, and new ideas anytime.
Smart companies treat AI as a long-term asset, not a one-time task.
This gives China's business-finance management a big lesson:
Before, we cared a lot about control and sure results.
But AI time needs trying and fast change.
We don't need to spend a lot of money once to make a big platform that looks good.
We need to let teams make fast small tests:
Today test credit risk, tomorrow test price math, the day after tomorrow add smart reports.
Step by step, fix mistakes, and grow real power.
The report says:
By 2027, 40% of companies will restart knowledge management.
Not because of old memories,
but because Gen-AI forces them to do it.
If a company does not make its knowledge clear, clean, and in a system, AI cannot eat it.
Before, we did knowledge management for checks, awards, and to show someone.
Now, it is real:
If knowledge is only in people's brains, chats, and messy files,
AI will make up stories.
If we want AI to speak right, we must first manage our own data and knowledge well.
Now comes a word many people do not like: governance.
The report says:
If the data is not ready for AI, AI cannot help at all.
Data rules and management are not useless extra work.
They are the only road to AI success.
And AI can help make governance faster too.
It does not sound cool,
but it is real:
Companies that truly use AI do not show off tricks — They build strong ground first.
If you think this report is a letter from the future,
you will find it does not say which model will win or which company will be king.
It keeps saying simple but deadly true things:
Plans cannot follow old ways.
Roles cannot follow old logic.
Budgets cannot follow old timing.
Data cannot follow old habits.
In short:
Before, we used old systems to control new tech.
Now new tech forces us to rewrite the system.
For Chinese companies, here is a simple test:
If your AI project sounds cool,
but in profit reports, cash work, price decisions, and daily running of the business,
there is no real AI helping —
Then you are still treating AI like a toy.
You are not yet treating AI as the basic power of business.
A real breakthrough is not “we made a smart report.”
It is:
In a place where we always guessed before,
we now let AI give a first answer and humans make the final call and take the risk.
The report ends with a message I think is perfect for every manager who is unsure about AI.
It says:
Now everyone is asking:
Can data, analysis, and AI really “deliver the goods”?
Can they really bring results?
This is very simple:
Can AI help the company make more money?
Lose less money?
Stay away from danger earlier?
If yes, go on.
If no, stop wasting money.
The future does not fall from the sky in one day.
It comes from every small serious change we make today:
Caring about data,
fixing rules a little,
and trying AI step by step.
In 2031, when we look back,
the winners may not be the loudest ones —
But they will be the ones who start early and take these warnings seriously.
Right now, you only see words from a report.
The real question is:
After reading this, are you willing to make even one small change in your company?
Once you make that step,
you are no longer only predicting the future — You are already building it.