Why a Quantum Computing Experiment Matters to Your SME

Quantum computing may sound far removed from your daily business concerns. You are probably more focused on responding to customers, tracking orders, managing staff, preparing quotations, or keeping operations moving. Yet a recent experiment involving GPT-5.6 Sol offers a highly practical lesson: AI becomes most useful when it is connected directly to the software and processes that run your business.

Researchers at MIT’s Engineering Quantum Systems Group connected an AI agent to laboratory software so it could conduct routine measurements, analyse results, and decide what to do next. According to OpenAI’s case study, this allowed the researcher to spend more time on experiment design, analysis, and planning rather than monitoring every individual step (OpenAI).

You do not need a quantum computer to apply the same principle. A Malaysian SME can use AI to handle repeatable workflows across sales, customer service, finance administration, inventory, and operations—provided the process is clearly defined and suitable safeguards are in place.

What Happened

Superconducting quantum chips require careful calibration. Researchers send microwave signals to qubits, measure the returning signals, analyse the data, and adjust the next measurement. The measurements are interdependent, meaning one result influences the next action (OpenAI).

Yankelevich tested GPT-5.6 Sol on an uncalibrated six-qubit chip. The AI agent selected measurement parameters, operated laboratory hardware through software, evaluated the results, refined the measurements when necessary, and saved useful results for later stages (OpenAI). When the signals were clear, it completed a standard sequence with limited researcher intervention.

However, the system was less reliable when signals were weak or noisy. It took longer to find appropriate settings and sometimes required guidance from an experienced researcher. The case study therefore presents a balanced result: AI agents can manage clearly defined workflows, but ambiguous situations still require human judgement (OpenAI).

“The practical lesson is not that AI replaces the expert. It is that AI can carry out structured work continuously while the expert focuses on decisions that require context.”

The research group now uses agents for routine measurements. The researcher can allow measurements to run for hours, check progress remotely, and intervene when a result needs correction or a different direction is required (OpenAI).

Why This Matters for Malaysian SMEs

Many Malaysian SMEs already have the basic ingredients for this type of automation: cloud accounting software, customer relationship management systems, e-commerce platforms, messaging tools, spreadsheets, point-of-sale systems, and shared business documents. The missing link is often not another application. It is the ability to connect these systems into a workflow that can observe information, take an approved action, and request human help when something unusual happens.

Consider a wholesaler in Selangor. A new order arrives through WhatsApp or an online marketplace. An AI-assisted workflow could extract the order details, check stock records, prepare a draft invoice, flag items that are unavailable, and send the order to a staff member for confirmation. The employee is still responsible for approving exceptions, but no longer needs to copy information manually between several systems.

A service company in Johor could use a similar process for job scheduling. When a customer reports a maintenance issue, the system could classify the request, check technician availability, identify the required service area, prepare a suggested appointment, and notify the customer. If the request involves a safety concern, unclear description, or unusual equipment, it can stop and escalate the case.

A small manufacturer in Penang could apply the idea to quality checks. An AI system might compare inspection readings with approved specifications, identify normal results, and highlight measurements that need review. It should not independently approve every questionable batch. Its role is to reduce routine checking and direct attention to the cases where experience matters most.

SME workflow	AI can assist with	Human control should remain over
Sales enquiries	Classifying requests and drafting replies	Special terms, complaints, and final commitments
Order processing	Extracting details and checking records	Unusual orders, stock substitutions, and approval
Service scheduling	Matching jobs with staff availability	Urgent cases and customer exceptions
Finance administration	Organising documents and identifying missing information	Payments, reconciliations, and compliance decisions
Inventory monitoring	Flagging low stock and unusual demand	Purchasing commitments and supplier negotiations

Start with a Controlled Workflow

The strongest lesson from the quantum experiment is the importance of a defined workflow. The researchers gave the AI measurement-specific skills, design targets, and software access. They did not simply ask it to “run the laboratory”. They explained what each step meant, how to assess results, and what should happen next (OpenAI).

You should take the same approach. Choose one process that is frequent, repetitive, and easy to measure. Examples include preparing daily sales summaries, sorting customer enquiries, checking whether purchase orders contain required details, or reminding customers about appointments.

Document the process before automating it. Write down the starting information, the normal steps, the acceptable outcomes, the situations that require escalation, and the person responsible for approval. This exposes hidden assumptions that are often known only by one experienced employee.

• Choose a narrow process: begin with one workflow instead of attempting to automate the whole business.
• Define the source data: identify which system or document the AI may use.
• Set approval points: require human confirmation for commitments, sensitive records, and unusual cases.
• Create an exception path: specify what the system should do when information is missing or contradictory.
• Review performance: track completion quality, escalations, corrections, and response times.
• Improve gradually: update instructions and workflow rules based on real examples.

The Bigger Picture

The broader shift is from AI as a standalone chatbot to AI as a controlled operator inside business systems. In the MIT example, the value came from connecting the model to laboratory software and allowing it to act across a sequence of tasks (OpenAI). For an SME, that could mean connecting AI to your inbox, customer database, inventory records, scheduling tools, and internal procedures.

This does not mean handing over unrestricted control. The case study shows why boundaries matter: clear signals produced better results, while noisy or ambiguous signals required expert guidance (OpenAI). Business processes have the same pattern. Routine enquiries may be easy to automate, while a dissatisfied customer, disputed invoice, or confidential employee matter needs a person.

Your competitive advantage may come from building a business that combines automated routine execution with strong human supervision. Staff can spend less time copying data, checking predictable updates, and monitoring tasks that run overnight. You can then focus their attention on customer relationships, operational improvements, product decisions, and difficult exceptions.

For Malaysian SMEs, the sensible next step is not to chase every new AI announcement. Identify where work repeatedly moves between people and systems, then test whether AI can manage the handoffs safely. Start small, keep approval controls visible, and treat unusual results as a signal for human review. That is how an advanced laboratory experiment becomes a practical operating lesson for your business.

Ready to Streamline Your Operations?

Technology moves fast. Your operations should keep up. AutoRunBiz builds AI systems that run your daily workflows — from WhatsApp order capture to accounting. Book a free 15-min ops audit →