Services
Paid Media SEO / AiEO / GEO Analytics & Technical CRM & Automation Social & Content Web & App Development AI Chatbots Strategy & Consulting
About Case Studies Contact Work With Us

AI Chatbots and Conversational AI: Built for Lead Capture, Qualification and Customer Service

Hero imageai-chatbot-lead-capture-south-africa-duly-noted.jpg
TL;DR — AI Summary
We build rule-based and AI-powered chatbots for web, WhatsApp and other channels. Every build starts with a use case audit. Every deployment is integrated with CRM and analytics. Every chatbot is tested and validated before it handles live conversations.

We build and deploy AI chatbots and rule-based conversational systems for businesses across South Africa and internationally. Chatbots built without expert direction and proper validation produce confident-sounding responses that are wrong, escalation paths that dead-end, and lead flows that disconnect from the CRM the moment a real prospect arrives. We scope the use case before selecting the technology, build to a defined specification, integrate with the systems that need to receive the output, and validate before deployment. The result is a chatbot that does what it was built to do — reliably, at scale, without constant manual intervention.

Rule-based or AI-powered — the right choice depends on the use case.

The chatbot market is full of vendors selling AI-powered solutions as the default. In many cases, a rule-based chatbot is the more appropriate and more reliable choice — and understanding the difference matters before any build begins.

Rule-based chatbots

Follows a predefined conversation flow. Entirely predictable and controllable within its programmed logic. Cannot hallucinate — it only outputs what it was explicitly programmed to output. Right for: lead qualification with specific fields, appointment booking, FAQ handling for a limited set of questions. Faster to build, cheaper to maintain, easier to audit.

AI-powered chatbots

Uses a language model to understand and respond to open-ended inputs. Appropriate where conversation flexibility matters — customer service across a broad product range, complex FAQ handling, conversational engagement. Requires careful implementation: a well-structured knowledge base, defined escalation logic, and thorough testing before deployment.

We do not deploy AI chatbots without validating their behaviour across a representative range of real-world inputs. The technology is powerful. The implementation determines whether that power is an asset or a liability.

What chatbots are built for in a marketing and sales context.

Lead capture and qualification

A chatbot deployed on a landing page, a service page, or a paid media destination can capture lead intent at the exact moment it is highest — when the visitor is actively engaged with your content. Rather than a static form, a conversational flow asks qualifying questions, collects contact details, and routes the lead based on answers — passing high-priority prospects directly to sales and lower-priority prospects into an automated nurture sequence. Lead qualification chatbots reduce unqualified leads reaching the sales team, improve response time for qualified prospects, and increase the conversion rate from site visit to captured lead.

Customer service automation

For businesses with high inbound enquiry volumes — common questions about products, services, pricing, availability, or account status — a chatbot can handle the majority of routine queries without human involvement, reducing call and email volume while maintaining 24-hour availability. Customer service chatbots are most effective when designed with clear scope — the specific questions they will handle and the specific escalation path for everything outside that scope. A chatbot that attempts to answer everything and handles escalation poorly is worse than no chatbot at all.

WhatsApp engagement and nurturing

In South Africa, WhatsApp is the primary digital communication channel for a large proportion of the population. A WhatsApp chatbot allows businesses to engage prospects and customers in the channel they actually use — with automated responses, lead qualification, appointment reminders, post-purchase follow-up, and re-engagement sequences. WhatsApp chatbot deployment requires WhatsApp Business API access through an approved Business Solution Provider, compliance with Meta's messaging policies, and careful conversation flow design within WhatsApp's specific constraints. We manage the full process.

Internal tools and process automation

Beyond customer-facing applications, conversational AI can be applied to internal processes — employee onboarding FAQs, internal knowledge base access, support ticket triage, and other workflows where a conversational interface improves the speed and consistency of information access.

Where chatbots are deployed and what they connect to.

Deployment channels

We deploy chatbots across web (embedded widgets on any page of your site), WhatsApp (via the WhatsApp Business API), and other messaging channels depending on where your audience is most active. Multi-channel deployments use a shared knowledge base and conversation logic where possible, ensuring consistency across channels.

CRM integration

Every lead-generating chatbot we build is integrated with the CRM from the outset. When a chatbot captures a lead, that lead flows directly into the pipeline with all collected data — contact details, qualification answers, conversation transcript, lead source, and channel. Without this integration, chatbot leads require manual processing, which introduces delay, human error, and gaps in attribution.

Analytics and tracking

Chatbot performance needs to be measured — conversation volume, completion rates, drop-off points, escalation rates, and lead quality. We implement analytics tracking for every chatbot deployment, connecting chatbot data to GA4 and the paid media platforms so that chatbot-generated leads are correctly attributed in campaign reporting.

A chatbot is only as good as its implementation.

AI chatbot tools are increasingly accessible. The barrier to deploying something that looks like a chatbot has never been lower. The barrier to deploying something that performs correctly, integrates properly, handles edge cases sensibly, and improves over time has not changed.

A non-expert deploying an AI chatbot without a defined knowledge base, without tested escalation logic, without CRM integration, and without validation against real conversation data will produce a chatbot that gives confident wrong answers, loses leads at the handover point, and cannot be measured. We have been brought in to fix chatbot implementations that were generating negative outcomes — prospects receiving incorrect information about products, leads disappearing into a system with no CRM connection, AI responses that contradicted the business's actual policies.

The implementation questions that determine whether a chatbot succeeds are not technology questions. They are: what exactly should this chatbot do, what should it never do, how should it handle what it cannot answer, and how does it connect to the systems downstream. Answering those questions correctly before the build begins is the work.

Scope a chatbot brief

The use case determines everything — technology choice, conversation design, integration requirements. We scope it before we build anything.

Scope a chatbot
Deployment channels
  • Web (all pages)
  • WhatsApp Business API
  • Multi-channel with shared logic

AI chatbots — common questions

A rule-based chatbot follows a predefined decision tree — predictable, controllable, appropriate for structured conversations. An AI chatbot uses a language model to handle open-ended inputs — flexible, natural, but requiring careful implementation to prevent incorrect or off-brand responses. The right choice depends on the use case and the required level of conversation flexibility.
Lead capture and qualification, customer service automation, WhatsApp engagement and nurturing, appointment booking, FAQ handling, and post-conversion onboarding. In a marketing context, chatbots are most valuable where there is a high volume of repetitive interactions that currently require human time, or where a conversational interface improves conversion rates over a static form.
Yes — and it should. CRM integration is a standard requirement for any lead-generating chatbot. Without it, chatbot leads require manual processing, which introduces delay and attribution gaps.
Via the WhatsApp Business API through a Meta-approved Business Solution Provider. The process involves BSP selection, business verification, phone number registration, compliance configuration, and chatbot setup within WhatsApp's messaging framework. We manage the full process.
Rule-based chatbot for a defined use case: two to four weeks. AI-powered chatbot with custom training, CRM integration and multi-channel deployment: four to eight weeks. The most important variable is clarity of brief before the build begins.
Every chatbot we build includes a defined escalation path — seamless handover to a human agent with full conversation context. Escalation logic is one of the most important design decisions in any chatbot build. A chatbot that dead-ends unresolvable conversations loses leads and damages the user experience.
Neil Duly — Digital Marketing Consultant, Duly Noted

Neil has scoped and delivered AI chatbot and conversational AI implementations for clients across multiple sectors, including lead qualification systems integrated with CRM platforms and WhatsApp-based engagement tools for South African consumer audiences. Every implementation at Duly Noted is built to a defined specification, validated before deployment, and monitored in the initial live period to ensure it performs as designed. Full background →

Tell us what you need the chatbot to do.

The use case determines everything — the technology choice, the conversation design, the integration requirements, and the success criteria. We scope it before we build anything.

Scope a chatbot brief