To determine whether a company's fate will be reshaped by AI, one single detail is enough.

Recently, I watched a speech delivered by Dr. Wang Jian during the Wuzhen Summit, which struck a deep chord in me.

In a special interview, he was asked a very interesting question: For the exact same AI model, if it is handed over to a company located in a small county, and to a world-class industry giant respectively, what will be the gap in their subsequent development?

Dr. Wang Jian replied: "For large enterprises, AI is often a revolution of tools. For small enterprises and players in remote areas, AI is a revolutionary tool."

The exact same model, with two different application methods, leads to two totally different fates.

One type of companies treats AI as a tool to cut headcount, reduce costs and save time. Their business remains exactly the same, only running a little faster than before.

The other type of companies treats AI as a revolution, using it to accomplish things that were impossible in the past, open up paths that were inaccessible before, and foster a brand-new business line.

We have seen numerous companies carrying out AI transformation, but if you think carefully, are you truly undergoing transformation? To judge whether a company is genuinely advancing AI transformation, one single detail is enough: When AI is introduced, do you treat it as a mere tool, or use it to open up new business?

In this article, we will talk about this topic.

01

Do Not Use AI in a "Show-off" Manner

Nowadays many companies are working on AI. When AI is mentioned in meetings, everyone nods in agreement; when writing year-end reports, everyone claims to have relevant application scenarios.

But if you ask them: What new things has your company created thanks to AI?

Nine out of ten cannot give a proper answer. What they can come up with are that they use AI to write copy, make PPTs, and their reports are now generated automatically.

This is exactly the first problem I want to talk about: using AI in a "show-off" manner.

What does "using AI in a show-off manner" mean?

Your business remains the same, your workflow remains the same, you just plug AI in as an extra tool.

You must have seen this scenario before.

Employees write weekly reports. Originally, 200 words are enough to make things clear: I met three clients this week, lost one deal, and the current bottleneck lies in the delivery cycle. It is concise and clear, and the boss can spot the bottleneck at a glance.

Now it is different. The company advocates using AI, so people first feed the 200-word content to AI, and ask it to expand it to 2000 words, adding background information, review and reflection, methodology, and "three key takeaways and future outlook". It looks substantial, as if a lot of work has been done, and the week is well spent.

As for the boss, it is too tiring to go through 2000 words, so he feeds the content back to AI and compresses it back to 200 words.

One person uses AI to add redundant content, the other uses AI to remove the redundancy. After the whole loop, the company gains nothing extra, except a false sense of security that "we are very AI-driven".

Why does this happen? Because everyone is just using AI to hit their KPIs.

This tendency is even more severe in large companies.

Dr. Wang Jian said that when large enterprises adopt AI, their first reaction is definitely to reduce costs, increase efficiency and optimize old businesses. Because all that is on their ledger is stock assets: factories, channels, clients, brands, all accumulated over decades. The moment AI is introduced, they will inevitably calculate "how much can we save".

Worse still, it is the mechanism problem. Each department carries its own targets, and when AI is introduced, it is first used to help fulfill old KPI, and after full rollout, AI is only used to speed up the old business. The failure to create new business is caused by the mechanism, not the tool itself.

Think about it, if the sales department's KPI is the sales volume of this year, what else can AI do for the sales department? It can only help salespeople sell old products faster. If the product department's KPI is the number of version iterations this year, AI can only help it develop old features faster.

Every department stays within its own boundary, using AI to make old things run a little faster. From the perspective of a single department, efficiency is improved; but from the perspective of the whole company, AI is split into countless fragments, all used to speed up old businesses.

That's why Dr. Wang Jian said: Large companies are not incapable of doing new things, they are just reluctant to let go of their old stakes.

This is the prisoner's dilemma for large companies: In terms of capability, they are fully capable of doing new things; but in terms of mechanism, they are simply not allowed to do so.

In many companies, AI has not created any new value at all, it just makes the old game a little more lively.

And liveliness never equals growth.

02

Genuine AI Transformation Must Be Rooted in Real Business Changes

Then what exactly is genuine AI transformation?

There is only one core point: it must be reflected in real changes to the business. It is not that you have used AI, but that you have reconstructed business scenarios including R&D and innovation, production and manufacturing, sales and service, and enterprise management.

Here I give you three judgment criteria.

First, you must be able to answer "what exactly has changed in your business".

A real AI project can definitely clearly answer: thanks to this AI, what changes that never happened before have taken place in our business?

Is the delivery cycle shortened from 30 days to 7 days? Are we now able to serve clients that we could not serve in the past? Can we now run businesses that used to be unprofitable? Or have we developed a brand new product or service that never existed before?

If you cannot give a proper answer, this project is most likely not "transformation", but just "cosmetic renovation".

Second, it must precipitate into new capabilities.

A one-off efficiency improvement is not transformation. Genuine transformation means turning one-time success into a replicable new capability.

For example, if a company uses AI to build an intelligent quotation system, if it only makes quotation faster this time, it is just a tool; if it builds a new business capability of "AI-assisted pricing + client stratification + rapid response" based on this system, that is real transformation.

I often say: For a good result, its process must stand review and refinement, and only replicable results are good results.

This rule applies exactly the same to AI.

Third, and the most important point, the changes must be perceptible to clients.

If the clients cannot feel anything at all from your company's AI transformation, then this transformation is 90% self-entertainment.

Because the value of AI must ultimately land on client value: are clients more satisfied? Do they get faster, better and cheaper services? Do they get new things that never existed before?

Any AI transformation that does not land on client value means nothing.

Therefore, genuine AI transformation must be rooted in business. Change one link of the business, precipitate new capabilities, and create new value. If you cannot meet these three points, your AI transformation is fake.

03

Reconstruct Your Organization Starting from Business

When AI is truly rooted in business, you will find that the original organization can no longer accommodate the new business.

What should you do? You must reconstruct your organization starting from business.

Because AI never only changes "efficiency", it changes the very form of business. And once the business form changes, the organizational structure, division of labor and workflows are bound to change accordingly.

I will give you a few simple examples to make it clear: In the past, producing a report required multiple steps of writing, reviewing, revising and finalizing; now AI can finish it in one go, are those intermediate approval links still necessary?

In the past, customer service required dozens of people working in shifts; now AI handles 80% of common problems, where should the remaining staff be arranged? What more valuable things should they do?

As you can see, when AI is introduced, it does not only affect a single position, but the very foundation of the entire organization.

When business changes, the organizational structure must adjust. To reconstruct the organization starting from business, you need to focus on the following points.

First, sort out the new business first, then define the new organization.

The order must not be reversed. You first figure out where the business is heading, what new things you want to achieve in the AI era, then work backwards to figure out what kind of organization, roles and talents you need to fulfill this goal.

Second, allocate a brand-new set of mechanisms for new business.

Mechanisms have inertia. That's why Wang Jian said "the failure to achieve that is caused by mechanism problems".

If you want to develop new business, you cannot use old mechanisms. The team for new business must have independent goals, independent space, independent fault tolerance permission, and independent assessment mechanism.

To put it specifically, you cannot use the assessment criteria for mature business to require innovative business. If you assess innovative business with the KPI system for old business, no one will be willing to work on it, and it is doomed to fail, because it completely violates the development law of innovative business growing from 0 to 1.

Therefore, you need to find a mechanism that fits innovative business, so that people have the courage to innovate and try and make mistakes.

Remember, if the organization does not change, new business will never grow. Only when the organization is reformed can new things have the soil to take root.

Finally, I want to say that AI is not a tool to help the company lay off a few people, but a mirror that forces the company to examine itself. Where can the business run faster, where can it perform better, where can it create more client value, AI will magnify all these problems.

Therefore, do not ask "what AI tool should we adopt", first ask "what new thing do we want to achieve with AI". The answer to this question is the watershed that determines whether a company can be reshaped by AI.

I share this with everyone who wants to truly do something new in the AI era.

This article is from WeChat Official Account "Zhang Lijun", author: Zhang Lijun, published with authorization from 36Kr.