Skip to content

Home / FAQ

FAQ · South Africa

Straight answers about WhatsApp automation, AI callers and CRM integration.

These are the questions South African business owners ask us before anything gets built: how quickly a WhatsApp chatbot can go live, what an AI caller actually handles on the phone, which CRMs and tools we integrate with, how POPIA is handled, what setup and retainers cover, and what support looks like once the system is running. No sales language, no invented numbers. We build these systems in Cape Town on the tools a business already runs.

Built around your workflowBased in South AfricaHuman oversight by design

Live conversation queue · todayExample view
Bergview Plumbing WhatsApp enquiry qualified, slot offered for 14:30Booking
Cape Coastal Removals missed call answered by AI caller, summary in CRMCall handled
Umhlanga Dental Rooms pricing question outside policy, sent to a humanEscalated
Ridgeway Auto Fitment job marked complete, review request queued for 17:00Reputation

How fast can a WhatsApp chatbot go live?

A WhatsApp chatbot goes live in days rather than weeks, because a WhatsApp chatbot is assembled from flows we have already proven: welcome, qualification, quotes, bookings and review requests. Only the copy, the tags and the routing get rebuilt for the brand.

Heavier integrations, payments, stock lookups and delivery updates, are added in phases so nothing blocks the first launch. We start with the one conversation that costs the team the most time, put it live on the business's own number, and watch real traffic before widening the scope. An existing business number can be onboarded to the WhatsApp Business Platform, or a new number provisioned, and we guide verification, templates and message quality along the way. Wording is drafted and approved before a single message sends. A shared inbox lets staff take over any thread at any point. We have built this way for 35+ companies from Cape Town, and the business owns the flows, the prompts and the data.

What does the AI caller actually handle?

The AI caller is an AI receptionist that answers the phone, greets the caller, categorises the reason for the call, answers common questions, takes orders, books appointments and hands off to a person when the call needs one.

Every call is summarised in plain language and written to the CRM against the right contact, so nobody rebuilds the story from memory. Where the caller agrees, the conversation continues on WhatsApp with the summary, the links and the next step already sitting in the thread, which is how a phone enquiry stops dying the moment the line drops. Edge cases route to a human by design: complaints, anything sensitive, anything the AI caller has no grounded answer for. The voice, the greeting, the working hours and the escalation rules are set by the business, reviewed before go-live, and adjusted once real calls start landing. Multiple languages are supported, either auto-detected or set as a default per line.

Which CRMs and tools do you integrate with?

CRM and tool integration covers the stack a business already runs, or ours where nothing is in place yet. Integration work reaches InOne CRM, HubSpot and GoHighLevel, Google Sheets and Google Workspace, Microsoft 365, Shopify and WordPress, and South African payment gateways.

Messaging runs over WhatsApp Business Cloud API or Twilio. Data that needs its own home lands in Supabase or PostgreSQL, workflows are assembled in n8n or Make.com, and language is handled by OpenAI, Anthropic Claude or Google Gemini. Webhooks and custom APIs keep records in sync end to end, so a booking taken on WhatsApp appears in the calendar and the CRM without anyone retyping it. The systems the team already trusts stay the source of truth. Existing contacts, tags and funnels can be migrated, re-mapped and rebuilt in a sandbox, and nothing goes live until it has been validated there. If a tool exposes an API, we can usually talk to it. If it does not, we say so before the build starts.

How do pricing, setup and retainers work?

Pricing has two parts. Pricing starts with a once-off setup fee for the initial build, covering the chatbot, the AI caller and the integrations, then continues as a monthly retainer for hosting, monitoring, iteration and support.

We do not publish a figure on this page, because scope depends on how many conversations, channels and systems are involved, and quoting before we have seen the process would be guesswork. New or larger projects, extra bots, a second brand, a complex automation, are quoted separately rather than absorbed quietly into the retainer. What the retainer buys is a system that keeps working: someone watching the queues, weekly improvements to copy and routing, and changes when the business changes. The build belongs to the client. Workflows, prompts and data can be exported, and nothing is locked behind a platform the team cannot leave. Scope, deliverables and the review cadence are written down before any build begins.

How is customer data protected under POPIA?

Customer data under POPIA is protected by design rather than by policy document. POPIA-aware building means consent is captured explicitly with a source and a time stamp, every automated message carries clear opt-out wording, and template usage is logged so an audit can show what was sent and when.

Each journey collects only the fields that journey needs. Retention windows delete records on time, access is role based, and change logs record who touched what. Transcripts and logs can be redacted where they hold sensitive detail. Payments run through PCI-compliant providers, and card details never sit inside a chat flow. Data is encrypted in transit and at rest, webhooks are signed, and least privilege applies across every connected system. Risky actions wait for a human sign-off. Compliance documentation is available under NDA, and the business remains the responsible party for the data we process on its behalf.

What happens after go-live, and how do we get support?

Support after go-live runs WhatsApp first. Support means a helpdesk on the channel the team already uses, with agreed response times, on-call escalation for priority incidents, and a shared roadmap board where feature ideas are tracked instead of lost in a thread.

The system keeps improving after launch. We review transcripts and analytics weekly, tighten copy, fix routing rules, sharpen prompts and add automations the data says are worth building. Dashboards cover conversations, conversion and response times, so the business can see what the system is doing without asking anyone for a report. Accuracy is guarded with retrieval from the company's own documents, strict intents, allowed-answer lists and escalation logic, so the assistant answers from approved material instead of guessing. Anything sensitive routes to a person by design. Guardrails stay fixed while prompts and content evolve.

Related capabilities. The same parts, your business.

Keep reading. Pages close to this one.

Did not find your answer? Ask us the real question.

Send one message describing where the business loses hours, whether that is missed calls, slow WhatsApp replies, quotes going cold or admin nobody has time for. We reply with an honest read on what automation can fix and what it will take.