王峰 AI × 经营 约一次诊断

总目 / 文章 / WF-0329

文章 · 存档

从元年科技的阵痛,看业财融合的下一程

——一份尊重当事人、面向行业的冷静分析与行动清单

本文依据网络传播的关于元年科技清偿方案通知、仲裁与社媒信息的截图,以及公开可得的公司历程信息来梳理与分析。

坦白的说,当年元年科技获得大笔融资的时候,我还有些嫉妒的,因为那个时候我也在创业,做了一个不懂财务也可以用的财务软件,结果,资金链断裂,死了。

我自恃对业财融合的理解深度没有几个人能够比拟,但可惜自己没有创业天赋,也没有营销能力,只能看着别人成功。

而元年科技的许多做法,跟我的想法是一致的,那个时候我想他们应该能成功,因为市场需求很大。这几年我自己的深度思考以后的结论,只有让全球会计人都行动起来各自实施轻量级AI业财一体的落地才能让整个行业聚变,而不是让一个大系统干掉会计人。

我写这篇文章目的不是轻视谁,而是尊重一家深耕二十年的企业、尊重每一位曾为业财融合投入心血的人,在事实可得的范围内,给这个行业一次必要的复盘与前行建议。

他们曾经做对了很多事

如果不了解元年科技,很容易忽略它在中国业财融合领域里起到的“破土”作用。

- 把管理会计方法论产品化:从作业成本、全面预算、合并报表到盈利分析,努力把“财务报表”翻译成“经营语言”。

- 技术路线并不保守:早期吸收国际先进的内存多维计算思路,后来走向国产化平台与PaaS,并在近年尝试将对话式智能引入财务分析。

- 长期扎在复杂场景:服务集团化、多层级组织与多口径数据,做的不是“演示级”项目,而是硬骨头。

这份“从0到1”的开荒精神值得尊重。也正因如此,今天的困难才更让人唏嘘。

问题与阻力:四条互相咬合的链条

1. 现金流的项目原生制张力。业财融合项目客单大、周期长、验收复杂,回款常常与里程碑绑定。只要宏观收缩、银行抽贷或甲方延期,现金流就会被瞬间掐住——工资、云资源、实施外包费都立刻承压。这不是一家公司的问题,是整个行业的共性结构性风险。

2. 高定制侵蚀毛利与人效。为每个大客户做“量身定制”,表面看单价高,实则超范围需求频发、主数据与口径治理滞后,导致“人拼人”的交付模式难以复制,毛利被蚕食,人效被拉低。

3. 从项目收入到订阅收入的跨越鸿沟。从一次性项目转向经常性订阅(ARR、NDR)本该是正确方向,但这条路需要两三年的阵痛期:要同时投入产品化、运维与续费能力建设。如果续费与扩张没形成规模、而研发与售后已经开足马力,现金消耗会加速。

4. 智能化的期待与现实落差。企业级智能分析、对话式BI、财务Agent等,短期更多带来体验与黏性,真正能够变现,取决于流程重构、数据治理与价值闭环。投入先行、回报滞后,是大多数厂商都会遇到的“时间差”。

放在一起看:项目制现金流 + 高定制交付 + 订阅化阵痛 + 智能化回报滞后,在周期下行与抽贷冲击时,会叠加成“多因素共振”。任何一家重咨询、重交付的厂商都需要敬畏这条底层逻辑。

我从元年的处境学到的七堂课

第一课:现金流不是KPI,是生命。把“13周滚动现金流 + 工资保障专户”提升为公司级红线;任何战略与研发都不能以拖欠工资为代价。对创始层而言,这既是法律底线,更是品牌底线与生态信用底线。

第二课:把“订阅化”做成收入地基。每个项目都必须带出可续费的经常性收入:基础订阅、模块包、席位/组织扩展、行业场景包。把一次性交付改造为“基础订阅 + 场景包渐进放量”,让续费与扩张成为增长主引擎。

第三课:交付工厂化,向“80%装配、20%定制”转型。建立行业蓝图库(指标口径、审批流、看板模板、常用集成),用低代码/参数化建模封装可复用能力;把主数据治理与指标血缘前置为开工条件,不满足就不立项。

第四课:智能化要对“钱”负责。把智能问答、预测、Agent与明确的财务与经营KPI绑定,例如:预算周期缩短X%、关账时间缩短Y天、采购/差旅节省Z%、营运资金周转提升N天。可计量、能对账,才有可持续定价与续费依据。

第五课:合同与回款机制重写。将回款与业务里程碑挂钩,同时约定甲方的数据与资源投入义务,减少“系统已就绪但数据未就位”的空转。

第六课:透明沟通与合规刚性。对外发布交付SLA与里程碑达成率,对内做工时与毛利透明;困难时期更要守住发薪合规,这是组织韧性的最后防线。

第七课:不抵触事后诸葛亮。任正非经常问自己下一个倒下的是不是华为,这种模拟事后诸葛亮的态度能够让自己始终保持足够的清醒和客观。正如现在我做的这些总结一样,看起来是时候诸葛亮,其实是跟我自己说如果我自己做的话,这些我能做得到吗?

为什么我长期强调“轻量级AI业财一体”

不是因为“轻量”好听,也不是“AI”多酷,而是因为它更能活下去。

我真的发现,业财融合是一个大坑,无数结构想要跟我合作共同开发业财融合系统,我经过深思熟虑以后都婉言拒绝了。不是我对我自己没信心,而是我对业财融合这个创业的赛道没信心。

正如当年如果卢卡帕乔利(全世界第一个把借贷记账法写入书中的数学家)不是写了这本《数学大全》而是给借贷记账法申请了专利,来给各个公司提供记账服务,我猜借贷记账法的发展要推迟100年。

真是让这种记账方法符合经济规律的环境下自由发展,才得到了长足的增长,全世界都在使用。

而今天的业财融合,也应当是如此的状况,就是应当作为就基础理论与实践的结合,让每一个会计人如同考CPA考会计师一样的去考业财融合的AI执行方案,唯独这一条路才能让业财融合走向光明。

给甲方与同行的“可执行清单”

给甲方(采购/CFO/业务负责人)

1. 先签“主数据与口径治理条款”,不达标不立项。

2. 以3个月能落地的KPI签约:周期缩短、周转改善、节省比例。

3. 订阅化采购:基础订阅 + 明确扩展包路线图。

4. 每月一次价值对账会:用现金与效率说话。

5. 合同与回款与上线与效果绑定,防止“装了却用不起来”。

给厂商与从业者

1. 建行业蓝图库与低代码装配中心,缩短交付尾部。

2. 把AI应用与“钱”绑定,避免炫技。

3. 强化续费与NDR,构建健康的经常性收入结构。

4. 做到劳动合规与信息透明,在困难期更要守底线。

5. 为一线会计人提供可用的Prompt、看板模板与流程Agent,让他们“即学即用、即用即见效”。

下一个元年,这些做法或许有些帮助

1. 轻量的目标,把“能见到钱的结果”前置:

- 三张表:利润、现金、营运资金(应收、应付、存货);

- 三条线:DSO(回款天数)、项目毛利、人效。

任何AI功能与可视化,都要落到这六个指标的改善上。

2. 轻量的方法 - 四步闭环:

① 数据对齐(主数据、口径、指标血缘);

② 小场景切入(预算滚动、关账加速、费用控与采购控);

③ 流程Agent化(报销、PO匹配、三单对账、异常预警催办);

④ 价值对账(用节省/周转改善对齐费用与续费)。

“五个1原则”:1套主数据、1张经营看板、1个高频流程Agent、1次月度价值复盘会、1份续费与扩展清单。

3. 轻量的交付节奏(90天SOP)

- 0–30天:数据口径/主数据就位 + 首个看板上线;

- 31–60天:一个流程Agent落地(例如报销或采购控);

- 61–90天:首次价值对账会 + 明确续费/扩展包。

让甲方“看得见、摸得着、算得清”,而不是被一个庞大的系统“吞没”。

我反复强调:不是让一个大系统干掉会计人,而是让会计人利用轻量工具、把“业财一体”变成自己可驾驭的生产力。会计人应当成为系统的主人,而不是系统的操作员。

业财融合赛道真的能赚到钱吗?

元年科技的阵痛,不是用来嘲笑的素材,而是一面镜子。

它告诉我方向是对的,节奏与底线更重要。

如果行业能从这次事件里汲取教训——把现金与订阅当作地基,把数据治理与交付工厂当作方法,把智能化与价值对账捆在一起,把员工合规当作红线——那么,业财融合就会走上一条更稳的路。

也向每一位会计人说一句:别等一个“大系统”来替代你。从今天起,用轻量级AI与数据化的方式,把“数字变价值”的主动权握在自己手里。

当越来越多的会计人成为业务与决策的推动者,这个行业才会真正强壮起来。

如果你觉得这篇文章对你和团队有帮助,欢迎转发。

也许,你的一次分享,能帮同行少走一些弯路,让更多人守住底线,也走向更好的业财融合。

From YuanNian Tech's Pain, See the Next Step of Biz-Finance Fusion

— A calm analysis and action list, with respect to the people and the industry

This article is based on screenshots of YuanNian Tech's debt repayment notice, labor arbitration and social media posts, and public company history info.

To be honest, when YuanNian Tech got big funding years ago, I was a little jealous. At that time I was also starting a business, building an easy finance tool for people who did not know accounting. But my money chain broke, and it died.

I always felt that my deep thinking about biz-finance fusion was hard for others to match. But sadly I had no talent in business start-up and no skill in marketing, so I could only watch others succeed.

And many things YuanNian Tech did were the same as my own ideas. At that time I thought they should succeed, because the market demand was big. My later deep thinking gave me a clear view: only if accountants around the world all act and build their own light AI biz-finance integration can the whole industry change, not by letting one big system kill accountants.

The purpose of this article is not to look down on anyone, but to respect a company that worked for twenty years, and respect each person who gave heart to biz-finance fusion. Within the range of known facts, I want to give this industry one needed review and some advice for the way forward.

They once did many things right

If you don't know YuanNian Tech, it is easy to forget the “pioneer” role it played in China's biz-finance fusion.

They turned management accounting methods into products: from activity-based costing, full budget, consolidation, to profit analysis. They tried to translate “finance reports” into “business language.”

Their tech path was not conservative: early on they used advanced in-memory multi-dimension ideas, later moved to local-made platforms and PaaS, and in recent years tried to add dialogue-based smart analysis into finance.

They worked long in complex scenes: serving big groups, multi-level organizations and messy data. They did not do “demo projects,” they did the hard ones.

This spirit of “from 0 to 1” deserves respect. And that is why today's trouble feels more sad.

Problems and barriers: four chains locked together

Cash flow stress from project model. Biz-finance fusion projects are big ticket, long cycle, hard to check and accept. Payment often tied to milestones. Once macro slows, banks pull loans, or clients delay, cash flow can be cut at once—salary, cloud fees, outsource fees all face stress. This is not one company's problem, it is a shared risk of the whole industry.

High custom eats profit and efficiency. Each big client wants tailor-made. On surface, high price; in fact, many extra needs, late master data and rules. Delivery relies on more people, hard to copy. Profit drops, efficiency falls.

Gap from project income to subscription income. Moving from one-time projects to regular subscription (ARR, NDR) is right, but it takes 2–3 years of pain: need to invest in product, service, and renewal ability at the same time. If renewals and expansion don't scale, while R&D and service already spend much, cash burn will speed up.

Smart tech hope vs. real return. Enterprise smart analysis, dialogue BI, finance Agents bring short-term wow and stickiness. But to make real money, need process rebuild, data cleaning, and value loop. Input first, return late—most vendors will face this “time gap.”

Seen together: project cash flow + heavy custom + subscription pain + smart return delay, when cycle goes down and banks pull loan, all join to form “multi-factor shock.” Any heavy consulting vendor must respect this base rule.

Seven lessons I learned from YuanNian's case

Lesson 1: Cash flow is not KPI, it is life. Build “13-week rolling cash flow + salary safe account” as company red line. Any strategy or R&D cannot be at cost of unpaid wages. For founders, this is law bottom line, brand bottom line, and trust bottom line.

Lesson 2: Make subscription the base of income. Every project must bring repeat income: base subscription, module pack, more seats/units, industry pack. Change one-time to “base sub + step add,” so renew and expand drive growth.

Lesson 3: Factory delivery, shift to “80% template, 20% custom.” Build industry blue library (metrics, flows, dashboards, links). Use low-code/param tools to pack repeat parts. Put master data and rule check before start, not good = no project.

Lesson 4: Smart must answer to money. Link smart Q&A, forecast, Agent to clear KPI: budget cycle shorter by X, close faster by Y days, save Z on travel/buy, working cash faster by N days. Must measure, must check, then price and renew is real.

Lesson 5: Rewrite contract and payment rules. Tie payment to business milestones. Also ask client to commit data and resources, reduce “system ready but data not.”

Lesson 6: Clear talk and hard compliance. Outside: share SLA and milestone done. Inside: open work hours and margin. In hard times, salary compliance is last defense of team health.

Lesson 7: Don't fear hindsight review. Ren Zhengfei often asked if Huawei is next to fall. This “hindsight sim” keeps one clear and cool. Same as my summary now: looks like hindsight, but also asks myself, “If I do it, can I hold these lines?”

Why I always stress “light AI biz-finance fusion”

Not because “light” sounds good, not because “AI” is cool, but because it helps survive.

I really found: biz-finance fusion is a big pit. Many partners wanted to co-build heavy systems with me. After deep thought, I said no. Not because I had no faith in me, but no faith in this start-up track.

It is like: if Luca Pacioli (the first man to write double-entry in a book) had filed a patent instead of writing Summa, and charged each company for bookkeeping service, I guess double-entry would be delayed 100 years.

It was because the method grew free in the right economy that it became strong and used worldwide.

Today's biz-finance fusion should be the same: it should be a base theory + practice mix. Each accountant should, like CPA exams, test on AI biz-finance fusion action plans. Only this road can lead to light.

Executable list for clients and vendors

For clients (CFO, buyers, biz heads):

First sign “master data and rule governance clause,” if not good, no project.

Sign KPIs that can land in 3 months: shorter cycle, faster turn, save percent.

Buy as subscription: base sub + clear add-on roadmap.

Monthly value check: talk cash and efficiency.

Tie contract and payment to go-live and effect, stop “installed but unused.”

For vendors and workers:

Build industry blue library and low-code center, cut tail of delivery.

Link AI use to “money,” stop show-off.

Push renew and NDR, build healthy regular income.

Keep labor law and info clear, in hard times guard bottom line.

Give frontline accountants useful Prompts, dashboard templates, flow Agents—let them learn fast, use fast, see fast.

The next YuanNian: these may help

Light goals: put “cash results” first:

Three sheets: profit, cash, working funds (AR, AP, stock).

Three lines: DSO, project margin, human efficiency.

Any AI or chart must land on these six.

Light method: four steps loop:

① Data align (master, rules, metrics).

② Small scene entry (rolling budget, fast close, expense/buy control).

③ Flow Agent (reimburse, PO match, 3-way check, alert).

④ Value check (save/turn vs. renew).

Plus “five ones”: 1 master set, 1 board, 1 high-use Agent, 1 monthly check, 1 renew list.

Light rhythm (90-day SOP):

0–30d: master/rule set, first board online.

–60d: one Agent live (reimburse or buy).

–90d: first value check, next renew.

Let clients “see, touch, count,” not be “eaten by a big system.”

I keep saying: not let a big system kill accountants, but let accountants use light tools to turn biz-finance into their power. Accountants should be masters, not operators.

Can biz-finance fusion really make money?

YuanNian's pain is not a joke, but a mirror.

It tells me: direction is right, but rhythm and bottom line matter more.

If the industry learns: put cash and subs as base, put data and delivery factory as method, bind smart to value check, guard labor as red line—then biz-finance fusion will walk a steadier road.

And to each accountant: don't wait for a big system to replace you. From today, use light AI and data tools, hold the power to turn “numbers into value” in your own hand.

When more accountants become drivers of business and decision, this field will be strong.

If you feel this article helps you and your team, share it.

Maybe, your one share can help peers avoid wrong turns, guard the line, and walk to better biz-finance fusion.

Would you like me to also polish this into a ready-to-post blog style (shorter sentences, catchy flow, with clear sub-headings), so it feels natural for a global audience on LinkedIn or Medium?