AI Automation Results: Real Case Studies from South African Businesses
Four anonymised case studies from South African businesses we shipped in the past 6 months. Each example describes the operational problem, the automation approach we built, and the workflow improvement that resulted. Names withheld where clients prefer privacy.
Results vary by industry, lead volume, implementation quality, and follow-up speed. The figures shown reflect each client's specific context and engagement scope — your numbers will differ. Pricing depends on workflows, channels, integrations, call or message volume, and setup complexity.
Sea Point Dental Clinic — After-hours call recovery for a multi-dentist practice
The problem
A 4-dentist Sea Point practice was leaving voicemail to handle calls between 16:00 Friday and 09:00 Monday. Roughly 30-50% of patient calls in that window went unanswered, and based on the practice owner's estimates, 60-75% of those callers never called back — they booked the next clinic on Google. At an average consultation fee of R850, the practice was bleeding R12,000-R20,000 per week in walked-away revenue.
The solution we built
An AI virtual receptionist for medical practices on a forwarded line. The voice agent ran on Twilio for inbound voice and a Railway-hosted LLM agent. Patient records and booking state lived in Supabase (eu-west-1 region for POPIA). Daily Telegram summary at 07:00 SAST showed call volume, booking rate, and any patients the AI flagged for human follow-up.
Workflow improvements
- After-hours call coverage — voice agent answered patient calls during the previously voicemail-only Friday-evening to Monday-morning window
- Reduced front-desk follow-up pressure — bookings flowed straight to the calendar without manual reception entry
- Patient WhatsApp confirmations — supported a more consistent booking and confirmation process
- Daily Telegram visibility — practice owner saw call volume, booking patterns, and any patient flagged for human follow-up
- Growth System engagement (multi-system retainer scope)
“Our after-hours calls used to go to voicemail and most of those callers never came back. With the AI line, the practice now has a way to answer those calls and route bookings without me being on the phone.”
Read the deeper write-up: how a Cape Town medical clinic improved after-hours call coverage with AI automation.
Cape Town Real Estate Agency — Faster enquiry response across WhatsApp
The problem
A 3-broker boutique agency focused on the V&A Waterfront and Sea Point markets was averaging a 4-hour 18-minute reply time on Property24 enquiries. By that point, qualified buyers had moved on to competing agencies. The brokers were losing roughly 60-70% of leads simply because nobody picked up the phone or replied to WhatsApp fast enough during viewings, family time, or evenings.
The solution we built
A WhatsApp lead-qualification bot running on the WhatsApp Business API (Meta) with an n8n cloud workflow handling intent classification and routing. Hot leads pinged the broker's phone within 30 seconds; cold leads entered a 7-touch nurture sequence. Calendar integration meant viewings booked themselves into the broker's diary without back-and-forth.
Workflow improvements
- Faster enquiry response — WhatsApp bot handled inbound enquiries during viewings, family time, and evenings
- Improved enquiry routing — hot leads pinged the broker's phone within seconds; cold leads entered a structured nurture sequence
- Reduced missed-enquiry risk — every Property24 enquiry routed into a structured workflow rather than depending on whoever was free at the time
- Calendar-integrated viewings — viewings booked themselves into the broker's diary without back-and-forth
- Focused Automation engagement (single-channel retainer scope)
“The response gap on Property24 enquiries used to be the biggest leak we had. The bot replies on Sunday evenings while we are at dinner — we just see the qualified leads when we get back to our phones.”
Johannesburg E-commerce Store — Multi-channel lead-scoring engine
The problem
A Sandton-based founder running an e-commerce store doing R450K/month in revenue was working 65-hour weeks chasing wholesale enquiries and B2B leads. Inbound came in scattered across email, WhatsApp, Instagram DMs, and trade-show contacts. Most leads went stale because the founder could not chase fast enough, and no qualification step existed before each lead landed on his desk.
The solution we built
A multi-channel AI lead-generation engine for Johannesburg businesses. Make.com workflows pulled new leads from every channel into Supabase. A Groq-powered LLM agent scored each lead on a 100-point scale (intent, fit, budget, geography). Hot leads got an automated 7-step Hormozi-style email sequence via Hostinger SMTP. Daily Telegram digest at 06:30 SAST showed the top 10 prospects with one-click profiles.
Workflow improvements
- Centralised inbound channels — scattered enquiries across email, WhatsApp, Instagram DMs, and trade-show contacts pulled into one structured workflow
- LLM-scored lead prioritisation — each lead scored on intent, fit, budget, and geography before landing on the founder's desk
- Improved follow-up consistency — top prospects entered an automated email sequence rather than waiting for the founder to chase
- Daily Telegram digest — top prospects delivered each morning with one-click profile views
- Multi-Channel Automation engagement (full-stack retainer scope)
“The AI lead-gen engine pulls everything into one place. I went from chasing leads on weekends to having a structured Monday-morning queue with the highest-fit prospects already prioritised.”
Cape Town Accounting Firm — Self-service onboarding + SARS deadline tracking
The problem
A 6-partner accounting firm in Claremont spent roughly 4 hours of partner time on every new client onboarding — collecting documents, populating Xero, scheduling kickoff meetings, sending engagement letters. SARS deadline reminders were tracked manually, and the firm had missed two filing deadlines in the prior quarter (R8,500 in penalties). New client throughput had plateaued at 12-15 onboards per quarter despite demand for more.
The solution we built
A Cape Town AI automation agency retainer with three connected systems: (1) a self-service onboarding flow on Vercel that collects client docs, runs an AI verification pass, and populates Xero via API; (2) a Make.com SARS-deadline tracker that scans client filings monthly and sends WhatsApp + email reminders 14, 7, and 3 days before deadline; (3) a monthly Telegram digest summarising client status, missed deadlines, and partner workload.
Workflow improvements
- Self-service onboarding flow — clients submitted documents through a Vercel-hosted intake; AI verification + Xero auto-populate replaced manual partner-led onboarding
- SARS deadline visibility — Make.com tracker scanned client filings monthly and sent WhatsApp + email reminders 14, 7, and 3 days before deadline
- Reduced manual partner workload — partner time previously spent on document collection and reminder phone calls redirected to client-facing work
- Monthly Telegram digest — partners saw client status, missed-deadline risks, and partnership workload patterns
- Growth System engagement (multi-system retainer scope)
“Client onboarding used to be a back-and-forth that ate partner time. The self-service flow plus the SARS deadline tracker means our team now spends that time on client-facing work rather than chasing documents.”
Want a workflow review for your business?
Take the free 5-minute AI assessment and we will return a personalised view of where automation could fit your workflows, channels, and current process. No pitch unless you ask. See the consultation-led pricing scoping page for engagement-tier descriptions. Anonymised case studies are presented as workflow examples, not universal promises — results depend on industry, lead volume, implementation quality, and follow-up speed.