Why a Robotaxi Debate Matters to Your Business
You may not be planning to operate driverless cars, build artificial intelligence systems or compete with Tesla. Yet the disagreement between Waymo and Tesla contains a practical lesson for every Malaysian SME: when technology handles real-world decisions, speed alone is not enough. You also need reliable data, safeguards and a clear way to manage unexpected situations.
That matters whether you run a logistics company in Shah Alam, a service business in Johor Bahru, a restaurant with delivery operations or a professional firm managing customer records. The technology you adopt may not drive a vehicle, but it can still affect orders, schedules, customer communications and staff decisions.
Waymo is promoting a system that combines cameras, lidar and radar, while Tesla is pursuing an approach centred on cameras and artificial intelligence. The disagreement is about more than engineering. It is about whether a simpler system can perform safely at scale, or whether a more cautious system is needed for difficult conditions.
TL;DR
Waymo’s argument is simple: automation working in the real world needs multiple signals, tested safeguards and experience handling unusual situations. For your SME, do not automate a process merely because a tool is impressive. Start with reliable information, human oversight and measurable checks.
The best automation is not always the most advanced-looking option. It is the one that performs consistently when customers, staff and conditions do not behave as expected.
What This Means
Waymo says fully autonomous vehicles cannot safely depend on one type of input alone. Its driving system combines cameras, lidar and radar to create a more complete view of the road. The company says it has accumulated more than 200 million real-world miles, and operates around 4,000 robotaxis across 14 US cities, providing 500,000 paid trips each week. Source: TechCrunch
Tesla takes a different route. It is building its autonomy system mainly around cameras and AI, while developing a purpose-built Cybercab without a steering wheel or pedals. A recent filing reportedly targets annual production of more than 125,000 vehicles. Source: TechCrunch
In plain language, this is a choice between a focused system and a layered system. A focused system may be simpler to deploy and easier to scale if it works well. A layered system may be more complex, but it can provide additional checks when one source of information is incomplete or misleading.
Your business faces similar choices. You might use one spreadsheet as the master record, or connect your accounting, inventory, sales and customer service information. You might allow an AI assistant to approve every reply automatically, or require approval for sensitive messages. You might trust a single forecast, or compare it with actual sales and staff feedback.
Automation should not only handle the normal case. It should also help you notice, stop and resolve the abnormal case.
How This Applies to Malaysian SMEs
1. Delivery and field-service businesses need more than location tracking. If you manage riders, technicians or sales representatives, a GPS map alone does not tell you whether a job is truly progressing. A stronger workflow combines location, appointment status, customer confirmation and exception notes. For example, if a technician is marked as “completed” but the customer has not confirmed the visit, the system should flag the record rather than silently close it.
This layered approach is useful in Malaysia, where heavy rain, traffic disruptions, construction and changing customer availability can affect schedules. Your automation should allow staff to record reasons for delays and alert the right person when a delivery or appointment falls outside the expected pattern. The goal is not to monitor people unnecessarily; it is to give you an earlier view of problems that would otherwise appear only after a complaint.
2. Retailers and wholesalers should avoid relying on one sales signal. An inventory system that uses only past sales may recommend the wrong replenishment level when demand changes, a supplier is delayed or a promotion creates an unusual spike. Add stock-on-hand, open purchase orders, supplier lead times and recent order patterns. A simple dashboard can show which items need attention, but a staff member should review unusual recommendations before placing an important order.
This is especially practical if you sell through several channels, such as a physical shop, WhatsApp, a marketplace and your own website. When these records are separate, you may not know the actual stock position. Connecting them reduces duplicated work, but you still need checks for cancelled orders, returns and manual adjustments. Your system should make discrepancies visible instead of hiding them behind a neat-looking number.
3. Professional-service firms need safeguards around AI-generated work. An AI tool can draft quotations, meeting summaries, follow-up emails and social media posts. That does not mean every output should go directly to a customer. A draft may contain the wrong service scope, an unsupported claim or a tone that does not fit your relationship with the client.
Create approval rules based on risk. Routine appointment reminders may be sent automatically. Contract wording, tax-related explanations, complaints and sensitive customer issues should be reviewed by a responsible person. Keep the original request, the generated draft and the final approved version in one place so you can understand what happened if a customer questions the response later.
4. Restaurants and food businesses can use exception-based automation. An ordering workflow can confirm new orders, send kitchen notifications and update customers without staff retyping information. However, the system should pause when an item is unavailable, a delivery address appears incomplete or a customer requests an allergy-related change. These are situations where a human decision is safer than an automatic assumption.
Start by listing the conditions that require attention. Examples include a large group order, repeated payment failure, an unusual delivery request or a customer complaint containing urgent language. Automation can identify and route these cases while your team focuses on making the right decision.
Practical Takeaways for Your Business
- Map one process first. Choose a repetitive workflow such as enquiry follow-up, appointment booking, stock updates or invoice reminders.
- Identify the inputs. List the information the workflow needs, such as customer details, order status, payment confirmation and staff notes.
- Add a second check. Do not let one field or one AI recommendation determine a high-impact action.
- Define exceptions. Write down when the process must stop and go to a person.
- Keep an audit trail. Record who changed a status, approved a message or overrode a recommendation.
- Test difficult cases. Include incomplete forms, duplicate orders, cancelled appointments, bad addresses and customers who reply unexpectedly.
- Measure outcomes. Track response time, missed follow-ups, order errors and unresolved exceptions. Every data point should be reviewed against your own records rather than assumed to be accurate.
- Train for handovers. Staff should know what the automation does, what it does not do and how to take over.
A Simple Readiness Checklist
| Question | What to check | Action |
|---|---|---|
| Is the process repetitive? | Does the same task happen regularly? | Document the steps before automating. |
| Are the records reliable? | Are customer, order and stock details current? | Clean duplicate and incomplete records. |
| What can go wrong? | List unusual or high-risk situations. | Create alerts and approval rules. |
| Who is accountable? | Can one person review exceptions? | Assign an owner and backup. |
| Can you measure the result? | Do you know whether the workflow improved? | Compare a small set of before-and-after measures. |
The figures reported about Waymo’s operations illustrate why real-world testing matters: the company says it has driven more than 200 million miles, operates in 14 cities and provides 500,000 paid trips each week. Source: TechCrunch You do not need anything close to that scale to apply the lesson. Even a small workflow should be tested with enough real cases to reveal where customers, staff and data behave differently from your original plan.
The Bigger Picture
The long-term competition between Waymo and Tesla may be decided by more than which vehicle looks better or which AI model appears more capable. Each company must deal with weather, emergency situations, school zones and other unpredictable conditions. Source: TechCrunch That is the same operational challenge faced by your business: normal transactions are easy to automate, while exceptions determine whether the system earns trust.
For Malaysian SMEs, the practical direction is clear. Build automation in layers. Use software to collect information, reduce repeated entry, highlight risks and prepare recommendations. Keep people involved where judgement, context or customer sensitivity matters. Over time, you can automate more steps when your records show that the workflow is stable.
This approach also helps you avoid becoming dependent on one tool. If an AI service changes, a staff member leaves or a platform becomes unavailable, documented processes and clean records make it easier to switch. Your business remains in control of the workflow rather than simply following whatever the software assumes.
Waymo’s message to Tesla is that physical systems need redundancy and evidence. Your message to yourself should be similar: choose automation that can explain what it did, show when something went wrong and give your team a clear way to respond. That is how technology becomes useful in daily operations—not because it removes every human decision, but because it helps you make the important decisions earlier and with better information.
Ready to Streamline Your Operations?
Your business should run itself. AutoRunBiz deploys AI agents to automate your daily operations — WhatsApp orders, invoicing, customer follow-ups, and accounting. Book a free 15-min ops audit to see where automation fits your business →