Every week, another AI tool launches with the promise of helping businesses save time, reduce costs and grow faster.
Companies subscribe. Employees attend workshops. Leadership teams announce that the organisation is now “AI-enabled”.
Yet months later, many businesses are still facing the same problems.
Founders remain buried in operations. Teams continue chasing approvals through email. Customer enquiries are manually routed. Marketing work is duplicated across platforms. Decisions still depend on one or two key people.
The organisation may have more AI tools, but it does not necessarily have better operations. That distinction matters.
AI can help a business reduce costs and increase revenue, but only when it changes how work moves through the organisation. Simply giving employees access to a chatbot rarely creates meaningful transformation.
The technology is not the problem. The operating model is.
AI tools improve tasks, AI operations improve businesses
Most organisations begin their AI adoption journey at the task level.
Someone uses AI to write an email. A marketer generates social media ideas. A sales executive asks AI to improve a proposal. A manager uses it to summarise a meeting.
These are useful productivity gains. They may save several minutes or even a few hours. But they remain isolated activities.
Once the task is completed, the employee returns to the same workflow, the same approval structure and the same operational bottlenecks. The task became faster. The business did not necessarily become better.
The more important question is not: “How can AI help us complete this task?” It is: “How should this work move through the organisation if AI were part of the operating model from the beginning?”
That shift moves the conversation from productivity to organisational design.
Instead of asking AI to assist occasionally, leaders are beginning to look at how humans, automation, and specialised AI agents can work together across entire business functions. That is where the real gains begin.
From one assistant to an AI operating layer
Across several of my businesses, we stopped treating AI as software that someone opens whenever they need help. Instead, we began designing operations around specialised AI agents.
Different agents support different functions, including administration, operations, class delivery, marketing, retention and customer support.
At the centre is Seraphina, my AI chief of staff and digital twin. I often describe her as my AI co-founder because she does more than respond to a single instruction.
We discuss ideas. She helps structure the execution. She assigns tasks to specialised agents. She reviews their work. She identifies gaps. She analyses the outcome before the work reaches me.
Also Read: ASEAN doesn’t need to win the AI race, it needs to run it together
My role is to oversee Seraphina, who, in turn, oversees the broader AI crew.
This is very different from using a general AI tool to complete one task at a time. A tool waits for an instruction. An agentic operating model can coordinate, execute, review and improve a process across multiple stages.
That does not remove humans from the organisation. It changes where humans create the most value.
The biggest return is not speed, it is a thinking space
Many conversations about AI focus on how many hours it can save. That matters, but I do not believe it is the biggest return.
The bigger return is thinking space.
When founders and leaders spend less time on repetitive execution, they have more capacity to think at a macro level. They can focus on strategy, partnerships, product development, positioning, customer experience and long-term growth.
Without that space, leaders often become trapped at the micro level. They are answering messages, correcting documents, following up with teams, checking small details, and repeatedly solving the same operational issues. They may be working extremely hard, but they are not necessarily moving the business forward.
AI can help shift leaders from micro-execution to macro-direction. But that only happens when the work has been designed properly.
If a founder is still involved in every step, every approval and every exception, AI becomes another tool that the founder personally has to manage. The founder remains the bottleneck.
Founders need to understand both the macro and the micro
There are two levels to every business process.
The macro level describes the overall journey. For example, a customer discovers the business, makes an enquiry, receives information, makes a purchase, goes through onboarding, and eventually receives ongoing support.
The micro level includes every action within that journey.
Who responds to the enquiry? Where is the customer information recorded? What happens when the person does not reply? Who approves a discount? Which message is sent after payment? What happens if the payment fails? When should a human step in?
Leaders need to understand both levels. At the macro level, they need to see how the process supports the wider business objective. At the micro level, they need enough detail to delegate, automate and maintain quality.
This has always been important for scalability, even before the age of AI. A founder who cannot explain how work gets done will struggle to delegate it to a human team. The same is true with AI.
The difference is that AI makes poor process design much more visible.
Also Read: The new border: Why server farms are the battleground of AI sovereignty
AI does not fix chaos, it scales it
One of the biggest misconceptions about AI is that it automatically creates efficiency. It does not.
If a process is unclear, AI cannot magically make it clear. If different team members use different methods, AI may simply reproduce that inconsistency more quickly. If nobody knows who owns a decision, an automated workflow may move the problem around rather than solve it.
Many companies believe they have an AI adoption problem. What they actually have is an operational clarity problem.
AI does not only automate good processes. It can also automate confusion, duplication and unnecessary work.
That is why leaders should not begin by asking which tasks they can automate. They should begin by understanding the process itself.
A practical way to document your processes with AI
Many founders know their business well but struggle to document it. The process exists in their heads.
They know what to do because they have handled the same situation hundreds of times, but the steps, exceptions and decision points may never have been written down.
AI can help extract that knowledge. You do not need to begin with a blank document. You do not even need to type everything. You can speak to the AI and ask it to interview you through the process.
For example, you might say: “Ask me questions one at a time about how we handle a new customer enquiry. Start with the overall process, then go deeper into the detailed steps, decisions, tools, people involved and exceptions. Keep asking until you have enough information to create a complete workflow.”
Then answer naturally, as though you were explaining the process to a new team member.
Start with the macro view. Explain where the process begins, what the intended outcome is and which major stages are involved.
Then move into the micro view. Describe the specific actions, approvals, tools, timelines, handovers and possible problems.
At this stage, do not worry about speaking in perfect order. The goal is to capture the raw data.
Once the AI understands the process, ask it to organise the information into a structured flow. It can help turn your explanation into:
- A standard operating procedure
- A step-by-step checklist
- A workflow diagram
- A responsibility matrix
- An automation plan
- A list of decision points
- A quality-control framework
You can then review the process and identify which parts should remain human-led, which can be automated and which can be handled by AI agents. This is often a much easier starting point than writing an SOP manually.
A simple sequence is: Extract first. Structure second. Review third. Automate last. The order matters. Automation should not begin until the process is understood.
Delegation is becoming a core AI skill
For years, founders have been told that they need to learn how to delegate. That principle has not changed.
What has changed is who, or what, can be delegated. Today, a business process may be divided between:
- A human team member
- A specialised AI agent
- An automated system
- A founder or manager making the final decision
The leader’s job is to decide how these parts work together. That requires more than prompting skills. It requires operational judgement.
Leaders need to know what good output looks like, where the risks are, which decisions require context and where human oversight is essential. AI may perform the execution, but leadership still defines the direction, quality and boundaries.
In that sense, AI does not reduce the need for good management. It raises the standard.
Also Read: Why seniority is repricing in AI-augmented teams, and what gets valued instead
Reducing cost does not simply mean reducing headcount
When organisations discuss AI and cost reduction, the conversation often jumps immediately to replacing jobs. That is too narrow.
Cost is also created through delays, duplication, poor handovers, unnecessary meetings, manual data entry, missed follow-ups and founders spending time on low-value execution. AI can reduce these costs without removing the human contribution.
It can make customer support faster. It can shorten campaign production cycles. It can improve retention follow-ups. It can help sales teams respond more consistently. It can reduce the amount of time managers spend gathering information before making a decision.
These improvements can lead to both lower operating costs and higher revenue. A faster response may improve conversion. Better onboarding may reduce refunds. More consistent follow-up may improve retention. Better use of customer data may create more relevant offers.
The revenue impact does not come from AI alone. It comes from improving the process around the customer.
AI Crew is an operating model, not a collection of bots
This thinking has shaped what we are building through AI Crew.
The idea is not simply to give every department another chatbot. It is to create specialised AI agents with clearly defined responsibilities, working together within a coordinated operating structure and under human oversight.
An AI marketing agent should understand the marketing workflow. An AI customer support agent should understand escalation rules. An AI operations agent should understand how tasks move across the organisation. And there should be a coordinating layer that ensures these agents are not operating as disconnected tools.
That is the role Seraphina plays across my businesses. She acts as the chief of staff, connecting the different functions, while I remain responsible for the direction, decisions and final oversight.
This is still evolving, but I believe the broader model will become increasingly common.
The future may not be one employee with ten AI tools open in separate tabs. It may be a coordinated team of humans and AI agents working within a single, clear operating system.
The real AI advantage is organisational design
The companies that gain the greatest advantage from AI will not necessarily be those with access to the most advanced models. Most organisations will eventually have access to similar technology.
The competitive advantage will come from how well that technology is embedded into the business.
Can the organisation clearly explain how work moves? Can leaders separate macro strategy from micro execution? Can processes be delegated without losing quality? Can AI agents operate within defined roles and boundaries? Can humans focus on judgement, creativity, relationships and direction?
These are not primarily technology questions. They are leadership and operations questions.
AI is already here. The next phase is not simply about adopting more tools. It is about building organisations that know how to work with them.
AI will not cut costs or grow revenue simply because a company purchased it. Those outcomes happen when the business itself is redesigned.
—
Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.
The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.
Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.
The post AI will not cut costs or grow revenue until you redesign how work gets done appeared first on e27.