In our previous article, we looked at one of the most practical uses of AI for a service business: the AI receptionist. When AI understands your services, prices, staff, opening hours, booking rules and the way your business operates, it can do much more than generate generic answers. It can answer questions, recommend services and help turn enquiries into bookings. We explored that idea in detail in Why AI Needs Business Context.
But customer service is only one part of the customer journey. Before someone can talk to your receptionist, they first have to discover your business. After they become a customer, you then need to give them a reason to come back. This is where another major use of AI becomes interesting: AI marketing.
Just like the AI receptionist, AI marketing only becomes useful when it has the right context. The difference is that marketing context can increasingly come from the data your business creates every day — customer conversations, bookings, CRM records, sales and previous campaigns.
Step 1: Bring New Customers In
For many clinics, salons, spas and other local service businesses, Facebook and Instagram are an obvious way to find new customers. The problem is that advertising through Meta can become complicated very quickly. Meta Ads Manager separates advertising into campaign, ad set and ad levels, with different decisions around objectives, audiences, placements, budgets, schedules and creative. You can see this structure in Meta's official Ads Manager documentation.
For a professional marketer, that flexibility makes sense. For someone running a clinic or salon, it can feel like learning an entirely different job just to advertise the business. I have seen this personally: one business owner I know was paying around ฿8,000 per month for another company to manage Meta advertising.
There is nothing wrong with using an agency. A good agency can provide real expertise, creative direction and strategy. But many smaller businesses are asking for something much simpler: “I have a service. I have some budget. Help me get more enquiries.”
That is one of the problems we are trying to simplify inside WelaOS. Instead of recreating Meta Ads Manager, WelaOS provides a simpler layer where the business can focus on the decisions it already understands:
- What do I want to promote?
- Where are my customers?
- Who am I trying to reach?
- How much do I want to spend?
AI can then help suggest the campaign angle, write the post and structure the offer. At the beginning, some of this still needs to be guided manually because a new business may not yet have much data in the system. The owner might tell the AI that they want to promote a particular treatment, increase weekday bookings or create awareness around a new service.
Over time, however, this changes. Because WelaOS is also where customer conversations, CRM records, bookings and sales are happening, the marketing module gradually gains more context. It can start seeing which services are growing, which are slowing down, what customers regularly ask about, which campaigns generated enquiries, and which of those enquiries actually became bookings and sales.
The marketing question can therefore evolve from “What should we promote?” into something much more useful: “Based on what is actually happening in this business, what should we promote now?”
For example, the system might identify that sales of a particular service have fallen by 20% compared with the previous period and suggest putting additional marketing behind it. But if the same service was already promoted last month and the campaign produced many messages but very few sales, simply repeating it may not make sense. The AI could instead recommend changing the offer, changing the audience, testing a different message or focusing the budget elsewhere.
Previous campaigns therefore become part of the business context. The system does not only know that an advertisement was published; because marketing, conversations, bookings, CRM and sales exist inside the same environment, it can increasingly connect that campaign with what happened afterwards.
This direction also reflects how smaller businesses are already applying AI. In Salesforce's Small & Medium Business Trends, 6th Edition, marketing campaign optimisation ranked as the leading AI use case among surveyed SMBs using or evaluating AI, ahead of content generation and automated customer recommendations.
Step 2: Convert the Enquiry
Once somebody responds to the campaign and messages the business, marketing has done its first job. This is where the AI receptionist takes over — answering questions, explaining services, checking availability and helping turn the conversation into a booking.
We already covered this part of the journey in detail in our article about why AI needs business context, so there is no need to repeat the full process here. The important connection to marketing is that the conversation itself becomes another source of data. What customers ask, what they are interested in and whether they eventually book all help the business understand whether the campaign generated useful customers rather than simply clicks or messages.
A connected system can therefore start asking a much more important question: did those enquiries actually become customers?
Step 3: Bring Existing Customers Back
Finding a new customer is only one part of marketing. The other is increasing the value of the customers you already have. Businesses can already send LINE broadcasts, WhatsApp messages and email campaigns, so the ability to send a message is not the difficult part. The difficult part is deciding who should receive it, what should they receive, and when should they receive it?
The easiest approach is to send the same promotion to everyone. A clinic launches a treatment and sends it to every LINE contact. A salon has a discount and sends the same message to its entire customer database. But customers are not all in the same situation. One may have visited last week, another may not have returned for six months, one may regularly purchase a specific service, while another may still have unused sessions in a package.
Treating all of them identically ignores information the business already has. McKinsey found that 71% of consumers expect companies to deliver personalised interactions, while 76% become frustrated when this does not happen. The original research is available in McKinsey's personalization research.
Personalisation is not simply putting someone's first name into a message. The useful part is understanding what is actually relevant to that customer at that moment.
Because WelaOS already contains CRM information, customer conversations, bookings and sales history, the marketing module can use this data to identify more meaningful customer groups. These might include customers who have not returned recently, customers who regularly buy a particular service, customers with unused sessions, people who asked about a service but never booked, high-value repeat customers, or customers whose previous purchases make another service relevant.
Instead of asking “Who can we send this promotion to?”, the system can ask “Which customers actually have a reason to receive this?”
That changes the type of campaign a business can create. Someone who has not returned for six months may need a reactivation offer. Someone who regularly buys one service might be suitable for a complementary service. Someone with unused package sessions probably should not receive another promotion trying to sell the same package again.
The Advantage Is the Connected Context
WelaOS can simplify creating Meta campaigns and help businesses run marketing through LINE, WhatsApp and email. But none of those capabilities are completely new on their own. Businesses can already use Meta Ads Manager, send LINE broadcasts and use standalone CRM or email marketing tools.
The advantage comes from having those functions connected to the same operating system. The marketing module knows about the CRM. The CRM contains the customer history. The booking system records what customers booked. The AI receptionist contains the conversations that happened before the booking. The POS records what was actually purchased. Previous campaigns become part of that same business history.
That allows AI to make better-informed decisions about what to sell, who to sell it to, when to contact them, how to present the offer and whether something similar worked before.
On day one, the AI might only know what services the business offers and what the owner manually tells it to promote. After months of use, the system may also have sales trends, customer behaviour, conversion information and campaign results. The AI has not simply received a better prompt — it has a better picture of the business.
This Is Not About Replacing Marketing Teams
A medium or large company with an experienced marketing team is solving much more complicated problems around positioning, branding, attribution, creative strategy, experimentation and customer acquisition. AI should help those teams, not pretend that their expertise no longer matters.
WelaOS is aimed at a different type of business: the clinic where the owner is also thinking about marketing, the salon where the receptionist also helps with sales, or the service business that does not want to become an expert in Meta Ads Manager or pay someone simply to operate basic marketing tools.
For these businesses, AI can reduce the technical knowledge required to get started and then use the data created by the business to make future marketing more informed.
From Finding a Customer to Bringing Them Back
The customer journey we are building around WelaOS can be summarised in three stages:
- Bring them in: simplify marketing and help the business generate new enquiries.
- Convert them: use the AI receptionist to turn enquiries into bookings.
- Bring them back: use CRM, conversation, booking and sales data to create more relevant marketing for existing customers.
The important part is not any single feature. It is that the three stages are connected, and each interaction creates information that can make the next marketing decision more informed.
That is where we believe AI becomes much more useful inside a small service business: not simply as a tool that can write another Facebook post, but as a system that increasingly understands what the business should sell, who it should sell it to, and what happened the last time it tried.