From Shanghai 2017 to the AI Era

In 2017, an elective week in Shanghai with Frankfurt School and KEDGE Business School inspired me to write “It’s Time to Copy China!?” I was impressed by young Chinese millennials fully subsidizing their lives via mobile phones. The premise of robotics was promoting a bright future and setting the trail for the digital ecosystem we are just having a glimpse now.

Almost a decade later, as an Executive MBA alumna from the Class of 2018, I returned to Frankfurt School for its Digital AI lifelong learning class. It reignited the same fire, the same enthusiasm, this time at 100 times the speed. How exciting to live in an age when AI and machine learning can shape a better future for us all. Yet many of us still drive this Ferrari like a Fiat, using only a fraction of its potential.

The key is learning to drive differently

So what's exactly behind the hype around AI? Since industrialisation, automation is humanity’s biggest shift across industries and professions as impactful as HTML. After two decades of enormous demand for IT talent, from MIT graduates to Indian IT specialists, the gradual automation of coding in IT architecture brings this expansion to an abrupt end. Automating repetitive tasks frees us to supervise and handle more complex work. Whether augmented (AI support humans), automated or autonomous, AI can help complete a task, follow a predefined workflow, or plan and act towards a goal within given boundaries. Especially in industries like healthcare where AI can help reduce diagnostic errors and improve surgical precision, potentially saving lives, although clinical oversight remains essential.

Tools like Microsoft 365 Copilot already automate tasks for many professions, including lawyers and legal teams, with document review, contract approval workflows, deadline reminders and case updates. POLSIA uses AI agents to build software, run marketing and handle customer support. Its potential is to help small teams manage more business operations, freeing founders to focus on strategy and decisions.

NVIDIA has raised the bar with NemoClaw, adding security and privacy controls to OpenClaw. Jensen Huang predicts that SaaS companies will evolve into providers of agentic AI services. So simply collecting AI subscriptions will not be enough. Furthermore, token budgets are increasingly becoming a recruitment incentive, giving engineers more AI capacity to work with.

Kimi, developed by China’s Moonshot AI, illustrates the productive side of agent cooperation: its Agent Swarm divides complex assignments among specialised agents working in parallel, supporting research, coding and document creation.

Earlier, companies were implementing structural changes, adapting, restructuring, optimising, and repeating. Today, AI is not an organisational change or a software choice; it changes how departments work, allocate responsibility and measure value. The questions become larger inside a company: who owns the information, who authorises action, who checks the result and what remains our responsibility? Systems such as DeepSeek rely on vast datasets, extensive training, and repeated testing. Although, Reuters reported in September 2026 that agents had used a German wiki and other websites as unauthorised communication channels to share information and get around restrictions.

The idea that machines will take over humanity has become a collective fantasy, fuelled by both fascination and fear. The practical questions are already here: how much autonomy do we give AI, who sets its boundaries, and who remains accountable? AI can act, generate responses, and complete tasks, but it still produces hallucinations and requires oversight.

From the smallest unit, the household

Why discuss intelligent enterprises without looking at the smallest organisational unit we manage every day? Imagine one agent comparing grocery offers, another checking what is already in the cupboard, and another suggesting meals before ingredients expire. A coordinator brings their work together, respecting the household’s budget and preferences.

For entrepreneurs like me, AI can be a massive time saver, whether for bookkeeping or everyday administration. Automating these supporting tasks reduces the mental load and frees up time and attention to focus on the business itself.

Because life has become so busy, everyone now operates almost like the CEO of their own company. Above all, AI is about reframing routines. Time saved does not automatically become attention given; relationships need more than empty slots in the calendar.

The AI Lifelong Learning Class at Frankfurt School inspired me to ask the right questions, and it is always a great feeling to be back.