← All systems

Here’s what would solve your problem

AI Marketing Attribution and Content Engine

Cheap leads don’t always mean profitable ads.

An ad that costs more per enquiry can still make you more money if more of those people buy. ABC Sales AI helps you find which ads bring buyers, understand what persuaded them, and prepare more versions to test as you grow.

If you do not fix it

Look at what happened after the enquiry.

In this example, the ad with the cheaper enquiries brought no bookings or sales. The other ad brought fewer enquiries, but some of those people booked and bought.

Read the full explanation

You keep paying for ads that do not sell. You never put more into the ones that do. Next month's budget is a guess again. You leave money on the table every month, and you never see which month it was.

Sample data

Ad AAd B
Cost per enquiryRM15.88RM32.70
Chats94
Bookings02
Sales revenueRM0RM4,800

How we fix it

Find out why your best ads work. Then make more like them.

Your consultant and AI Manager study the ads that brought bookings and sales, alongside what those customers said. We turn the strongest messages into more versions for you to test. The aim is to attract more serious buyers as you grow, while watching enquiry costs and sales.

Read the full explanation

More sales from the same ad spend. We connect your ads to your chats and send your sales and appointments back to Meta, so Meta learns who buys and finds you more of them. Leads arrive marked with the ad, post or page they came from. Nobody types anything. Within a month you know which ads bring buyers and which waste money, and the wasted money moves to the ads that work. Every week the AI reads your new chats and drafts your next posts and ad angles from what buyers actually said, in your voice, for your approval.

Watch how it works

Watch how it works in about a minute.

How we fix it

Here’s what we plan to do for you

See what happened after someone clicked your ad

Follow the enquiry through to a booking or sale, so you can compare ads by the business they brought in.

  1. Ad
  2. Enquiry
  3. Sale
  4. Report

Make more ads based on what persuaded your buyers

We look for the promise, product benefit or question that caught buyers’ attention. AI Manager drafts new versions around those ideas, giving you more ads to approve, test and rotate as you grow.

  1. Winning ads
  2. Buyer chats
  3. New versions
  4. Your approval

Agree which outcomes to track

  1. Serious enquiry
  2. Confirmed booking
  3. Paid sale
Kong Ming holds a customer profile with a heart to represent a personal customer experience.

Workflow illustration

You agree what each stage means for your business.

01Step 1

Decide what counts as a serious enquiry

We agree with you what counts as an enquiry, a confirmed booking and a paid sale, so the reports reflect how your business actually sells.

See the details: Decide what counts as a serious enquiry

I already know my cost per lead.

Cost per lead and cost per buyer are different numbers pointing in different directions.

1. Agree what a buyer actually is

We map how a customer buys from you and agree what a real enquiry, an appointment and a purchase actually mean in your business, because Meta can only learn from a sale once we have agreed what a sale is. We also decide the source categories that matter to you, and write source questions for the channels that carry no technical source, so referrals and walk-ins stop reading as zero.

This is the business expert half, and it is why the report answers your question rather than a generic one.

Your numbers mean what you think they mean.

Follow the enquiry back to its source

  1. Available ad or website details
  2. Customer conversation
  3. Booking or sale
Kong Ming wears a headset and holds a notebook for a helpful customer conversation.

Workflow illustration

For referrals and walk-ins, ask how the customer heard about you.

02Step 2

See which ad brought the enquiry

We save the ad and website details that are available. For referrals or walk-ins, we add a question asking how the person heard about you.

Mobile mockup, adapted from the original example.

Mobile mockup of the sample attribution sheet: ten leads, seven bookings and two sales, with lead XXX1 expanded to show its source ad and RM4,800 payment.

Swipe or scroll sideways to see more

Attribution sheet matching each WhatsApp lead to its ad.
One row per WhatsApp lead: campaign, ad, booking, show-up and sale. Leads with no source are flagged to fix by hand. Sample data.
See the details: See which ad brought the enquiry

How can a chat possibly know which ad it came from?

Every click already tells you where it came from. Today that information is thrown away.

2. Attach the source to your leads

Leads are created automatically with where they came from, on every channel you use, and the tracking codes on your website, web form and chat links are saved with it.

Your AI Employee notes where each lead is in the journey as the conversation happens. Leads arrive carrying where they came from, without anyone typing anything.

Leads stop arriving anonymous.

Give Meta confirmed outcomes

  1. An agreed outcome happens
  2. The confirmed event is sent
  3. Meta can learn from the result
Kong Ming teaches a small group beside a gold star.

Workflow illustration

Only the enquiries, bookings and sales you agree to track are reported.

03Step 3

Tell Meta when someone books or buys

We report the enquiries, bookings and sales you agree to track to Meta, the company behind Facebook and Instagram. This helps its ad system learn who is more likely to become a customer.

Mobile mockup, adapted from the original example.

Mobile mockup of the original blue booking confirmation and successful AppointmentBooked event sent to Meta. Private meeting link remains hidden.
Three-step flow: booking confirmed in chat, Meta conversion event sent successfully, Meta learns who buys.
Real screens: the customer books in WhatsApp, the AI fires the CAPI booking event, and Meta learns who buys. Personal details blurred.
See the details: Tell Meta when someone books or buys

Why does Meta keep sending me people who never buy?

Because Meta only knows who clicked. It never learns who bought, unless you tell it.

3. Teach Meta what a buyer looks like

We connect your Meta ad account to your chats and tell Meta the three moments that matter: a real enquiry, a booking, and a sale with its value.

From then on Meta hears about sales, not just messages, and starts hunting for more people like your buyers instead of more people who message and go quiet.

Your ad spend starts optimising toward buyers.

Use the result to guide the budget

  1. Compare enquiries and sales
  2. Review costs by source
  3. Choose what to increase or stop
Kong Ming points to a report chart on a tablet.

Workflow illustration

Your consultant reviews the first month with you. You decide the next move.

04Step 4

Decide which ads deserve more of your budget

AI Manager reports show enquiries, bookings and sales from each source. We review the first month with you and discuss which ads to spend more on, reduce or stop.

Mobile mockup, adapted from the original example.

Mobile mockup of selected AI Manager report rows: 36 chats and four bookings marked SCALE; 23 chats and zero bookings marked CUT. Ad names and costs are hidden.
AI Manager scorecard highlighting the ad to scale and the ad to cut.
Real AI Manager report on our own account, last 30 days. Ad names and spend blurred.
See the details: Decide which ads deserve more of your budget

Will I actually use this, or just stare at another dashboard?

A chart is not a decision. The last step turns one into the other.

4. The monthly spend decision

AI Manager sends you a brief every week and every month: leads, real enquiries, appointments and sales, by source. It also warns you the moment a source that used to bring buyers stops bringing them.

Then we sit with you on the first month of real data and turn it into a spend decision, not just a chart: which source to feed, which to starve.

Next month's budget is a decision instead of a habit.

Build on what buyers respond to

  1. Effective ad message
  2. Buyer questions and concerns
  3. Ideas for the next draft
Kong Ming uses a magnifier to examine an advertising card and its results.

Workflow illustration

Your consultant helps choose which selling points are worth testing.

05Step 5

Find out why your best ads brought buyers

AI Manager compares the messages in your winning ads with the questions and concerns in buyer chats. Your consultant helps choose which selling points are worth building on.

Mobile mockup, adapted from the original example.

Mobile mockup of an AI Manager response: a winning ad brought five of eighteen bookings, buyer questions appeared eight, six and five times, and a new hook is drafted for approval.
AI Manager content brief with the winning angle, buyer questions and a new ad script.
Real AI Manager output: it learns from the winning ad and what buyers ask, then writes the next ads with 15-second scripts.
See the details: Find out why your best ads brought buyers

I never know what to post.

Your customers already wrote the brief. It is sitting in your chat history.

5. Mine the chat history

Your customers have told you exactly what stops them buying. It is in your chat history, thousands of messages deep, and nobody has read it. We have AI Manager read that whole history to rank the five questions people ask before buying and the three objections that kill sales, with counts.

Then a consultant reads the findings and picks the angles that are actually worth spending money on. That part is judgement, not a report.

You know exactly which worry your next ad should answer.

Your approval comes first

  1. Your brief
  2. Draft and image
  3. Approve, edit or reject
Kong Ming holds a pen and a checklist ready for review.

Workflow illustration

You or your agency decides what gets tested and published.

06Step 6

Create more versions of the ideas that work

AI Manager prepares different ad drafts and images in your voice. You approve each one before you or your agency tests and rotates it, while watching enquiry costs and sales.

See the details: Create more versions of the ideas that work

I do not want a robot posting for me.

It does not post. It drafts, and it drafts like you.

6. Your voice, your approval

We write the standing generation prompt with you: your voice, your formats, your products, what you never say, so what comes out is publishable rather than generic. Then we set the approval step, so drafts arrive for you to approve rather than posting themselves.

The images are generated on the ABC platform as part of each draft, so a post never waits for a designer.

Nothing goes out without your yes, and nothing waits on Friday.

The package

We set it up around the way you sell.

We agree what to record, save where enquiries come from, and set up reports and ad drafts for your review. The work included and what we need from you are set out below.

What you get

  • More buyers from the same ad spend, because Meta now learns from your buyers, not your clicks
  • Your next ad written from what buyers actually said
  • Leads marked with where they came from, on its own
  • Sales by ad, not leads by ad
  • After one month, a clear answer: which ad to double, which to switch off

What we need from you

access to your Meta Business Manager and ad account, your website and chat links, and a 30-minute kickoff.

Timeline:

1 to 2 weeks from kickoff. One month of real data gives you the answer.

Not included:

the ads themselves, ad spend, and media buying (we make the numbers honest, you or your agency act on them). Not included on any system: WhatsApp, LINE or other messaging fees and credits, ad spend and ad management, and your other apps' subscriptions or API fees.

What is included
  • A working session to agree what a real buyer is in your business: enquired, showed real interest, booked, paid
  • Your campaigns checked with your marketer so every click carries the right code, and the source on your leads matched back to the ad, the angle and the hook
  • Leads marked with where they came from, on every chat channel, with the tracking codes on your website, forms and chat links saved too
  • Your Meta ad account connected to your sales, so Meta learns from buyers, not clicks
  • Source questions for the channels with no technical source, so walk-ins and referrals stop reading as zero
  • Weekly drafts of your next posts and ad angles, in your voice, for your approval: the AI Manager is taught each ad by name, its angle and its script, so it sees which angles bring buyers, tells you which to do more of, and writes the next ones from what the people who actually buy or book ask before they do
  • The standing content prompt and schedule, yours to keep
  • The source report and a first-month review with your consultant: your best and worst source, by buyers

By yourself

Learn how Meta's sales tracking works and set it up yourself. Mark every lead by hand as it comes in. Then read your chats every week and write next month's posts and ads yourself. Most owners start, then stop by week two.

With us

Access to your Meta Business Manager and ad account, your website and chat links, and a 30-minute kickoff.

How we deliver it

We build it, check it and show your team how to use it.

We learn how your business works, set up the agreed connections, check the records and train your team before handing it over.

Every install is the deployment of an AI system into your business, delivered as a service. In the categories IRAS uses for qualifying AI business services, the work is:

Consultancy and strategy
The business expert work above: the mapping, the definitions, the copy and the working sessions.
System development
The AI and system configuration above: the AI Employee, AI Manager and platform capability switched on and wired to your business.
Data and analytics
The data mapping and checks, plus any reports, briefs or alerts included in the agreed scope.
System-related training
Handover and training so your team runs it without us.
Already on ABC Sales AI?

Guided Launch set up one workflow with you, a first version built to get your first win, so what you already have depends on the one you chose. Typically:

Onboarding gave you: the AI Employee trained on your materials. Usually nothing on attribution: leads arrive without their source.

This adds: Better leads from the same ad spend: your campaigns checked with your marketer so every click carries the right code; your sales and appointments fed back to Meta so it learns who buys; the source on your leads matched back to the ad, the angle and the hook, in a report your marketer can act on; and better content: the AI Manager learns each ad's angle and script, sees which angles bring buyers, tells you which to do more of, and drafts the next ones from what the people who actually buy or book ask before they do.

Investment

Agree on the work before you decide to buy.

Your consultant agrees the outcomes, scope and delivery checks with you, explains the investment, and provides a quote for your business.

Book a free strategy call

Runs on your ABC Sales AI platform subscription. If you came through Guided Launch, you already have it.

We review your booking tools, POS, payment system and customer records during discovery. Your proposal identifies the connections we can build, the access they need and how we will test the agreed workflow.

We are a Singapore company. Can we claim this under the EIS 400% AI tax deduction?

Singapore company? Install work like this is generally the kind of AI spend the EIS 400% deduction covers for YA2027 and YA2028: up to about 51 cents of extra tax saving per dollar over a normal deduction, at the 17% headline rate. No pre-approval: when the work is completed and accepted we send the claim documents, and your accountant adds it to your normal return. Forward our accountant page, and confirm your own eligibility with your tax advisor.

Built and tested before we hand it over

See it work with your own customer records before we hand it over.

We build it, connect it, and show it working on your real leads before we hand it over.

We test the setup we agreed with you, teach your team how to use it and hand over the settings and reports we have built.

What you keep

This is an install. Tracking either fires or it does not, and you watch it fire on real leads before we hand over. What we promise is the build: your Meta ad account connected to your sales, the source on your leads, the weekly report on sales by source, all tested on your live account, and your team trained. If your ads are not the click-to-chat kind that tracking can see, we say so during scoping, before you commit. The decisions that are yours stay yours: the offer, the prices, the rules. We bring the expertise, surface the options and build. You decide. Everything we build stays yours:

  • The source tagging, live on every channel
  • Your Meta ad account connected to your sales, and working
  • Every lead's source history, growing from day one
  • The weekly report on sales by source, sent on its own

You watch it work on your real leads before we hand over, and everything on this list is yours from that day.

Honest fit check

Do your ads ask people to message you?

This service is for Meta ads that bring people into WhatsApp, Messenger or Instagram chats, where they enquire, book or buy.

Perfect for:

  • A business spending on Meta ads that point at WhatsApp, Messenger or Instagram chat
  • An owner who knows cost per enquiry but cannot name which channel produces buyers
  • Multiple sources feeding one inbox: ads, posts, website, referrals, walk-ins

Not built for:

  • A business that spends nothing on advertising and gets everything by word of mouth: there is nothing to optimise yet
  • Ads that send people straight to a website checkout with no chat in the journey: Meta's normal website tracking already covers you

Before you ask

Quick answers

My ads do not point at chat.

Attribution here needs click-to-WhatsApp, Messenger or Instagram ads, because without that click there is nothing to attribute against. We check this at discovery, and if the fix is as simple as changing where the ad points, we show you how before anything is built.

I already see results in Ads Manager.

Ads Manager knows clicks and messages. It never knows which conversation became a sale, so a channel full of cheap enquiries that never buy can look like your best channel. This install reports by buyers, and feeds the sale back to Meta so the ads learn too.

How fast do I see it?

Sources attach from the day it goes live. One month of real data is enough to name your best and worst source by buyers, and we sit with you on that first month to turn it into a spend decision.

Do referrals and walk-ins show up too?

Yes. For channels that carry no technical source, we write source questions into the conversation, so referrals and walk-ins stop reading as zero.

Is this a content agency?

No. An agency sells you batches of posts forever. You own a generator fed by your own chat history: the AI Manager keeps drafting from your live conversations on a schedule after we leave. The first month of drafts is the proof, not the product.

Will it sound like AI?

The standing prompt carries your voice, your formats, your products and what you never say, and its memory of your business grows every month. You approve every draft, so nothing generic goes out with your name on it.

We are a Singapore company. Can we claim this under the EIS 400% AI tax deduction?

Often, yes. For YA2027 and YA2028 the EIS gives a 400% tax deduction on up to S$50,000 of qualifying AI spend per year of assessment. IRAS's examples of qualifying AI business services include system development, consultancy and strategy, data and analytics, and system-related training. At the 17% headline rate that is up to about 51 cents of extra tax saving per dollar compared with a normal deduction; your actual saving depends on your tax position. Grant-funded amounts and non-AI items such as messaging fees do not qualify, and mixed packages are split. There is no application and no pre-approval: you claim it in the tax return you already file. Our part is to make it effortless. When the install is completed and accepted, we send the necessary documentation, itemised under this system's title, and we work with your accountant on anything else the claim needs from our side. Whether your company qualifies depends on your own situation, so confirm with your tax advisor.

ABC Sales AI · abcsales.ai · Every install is scoped to your business and quoted by your consultant. Outcomes depend on your market, offer and execution; what we commit to is the work, not the result. You are responsible for the content you send; campaigns are vetted against your industry's rules before launch. ABC Sales AI is not a tax advisor; any Singapore EIS treatment is for your tax advisor and IRAS to confirm.

Build your next ads on what already brings buyers.

Book a free strategy call. We’ll discuss how to identify your winning ads, learn from buyer chats and prepare more versions for you to test as you grow.

Book a free strategy callOr talk to a consultant on WhatsApp