Websites, WhatsApp automation, and reminder systems — built as one system, not three separate purchases

Because they're solving one problem, not three. A customer's actual journey with a business doesn't respect the boundary between "website" and "messaging app" and "booking software" — it's one continuous conversation that happens to move across different screens. They find the business through a search result or an AI-generated answer, they land on a page that either loads in two seconds or loses them, they ask a question on WhatsApp because that's faster than filling out a form, and if they book something, they either show up because someone reminded them or they don't.

Treating those as three separate vendor relationships means three separate failure points, three separate invoices, and — the part that actually costs money — three systems that don't know anything about each other. A chatbot that can't see what the website already told a visitor asks them to repeat themselves. A reminder system bolted onto a generic SaaS platform doesn't share a phone number, a tone of voice, or a booking history with anything else the business runs. Building all three from one foundation means a single decision — how fast should this load, how should this sound, what happens when the system genuinely doesn't know the answer — gets made once and applied consistently everywhere the customer actually shows up.

A screenshot of two websites can look identical. Their load times rarely do. Most agency sites are assembled from a page builder plugin, a slider plugin, a gallery plugin, sometimes an SEO plugin on top — each one loading its own JavaScript before the page the visitor actually wants ever renders. That stack, not the photos or the content, is the real reason most UAE business sites take five or six seconds to become usable. A hexweb.dev site is written directly — PHP, HTML, CSS by hand — with every image compressed and loaded individually only when it's needed, which is the entire difference between a clinic's full before-and-after gallery loading in six seconds versus under two.

That speed isn't cosmetic. It's the foundation everything else sits on: a slow site doesn't just lose the human visitor who gives up waiting, it's also less likely to be the page an AI system actually opens and reads when deciding what to cite — a point worth its own section below. Full detail on cost, timeline, and what's included lives on the web design page, with the complete AED pricing breakdown on the cost & timeline guide.

It would be easy to plug in a large language model and call it an AI chatbot — plenty of agencies do exactly that. hexweb.dev builds the WhatsApp automation as a rule-based state machine instead: it recognizes a defined set of intents — pricing, hours, availability, "I want a person" — and responds with clear, deterministic logic rather than generating an open-ended reply and hoping it's accurate. The tradeoff is real and worth saying plainly: a rule-based system won't hold a freeform conversation about anything outside what it's built to handle. What it will do is answer correctly, every time, for the questions that actually make up most of a business's WhatsApp volume, and hand off to a real person the moment something falls outside that — no confidently wrong answer sent to a customer at 11pm because a language model filled in a gap with something plausible-sounding.

It also runs on the business's own WhatsApp number, connected directly through Meta's Cloud API in real time — not a shared line, not a third-party platform holding the conversation history hostage. Full architecture and the actual pricing tiers are on the chatbot page; the working interaction pattern itself is something you can try directly on the live demo.

The appointment reminder system exists to solve a specific, unglamorous problem: front-desk staff are busy, the manual reminder call the afternoon before an appointment is the first thing that gets skipped on a hectic day, and the client isn't being difficult — they simply forgot, in the one place they weren't reminded. An automated system removes the dependency on a person remembering to make that call at all, checking upcoming bookings on a schedule and sending a reminder automatically.

The deliberate choice here is that it runs as its own self-hosted stack on standard hosting — the same deployment model as the rest of the site — rather than a subscription to a third-party booking platform. That matters for the same reason it matters with the chatbot's channel independence: a rented platform's pricing and rules belong to the platform, not the business running on it. Where this genuinely stands today is worth saying honestly rather than dressing up: the core system is built and in active testing, not a finished, polished rollout yet. Exactly what's proven versus what's still being refined is laid out plainly on the reminder system page.

A few weeks back, this site's News section covered a lawyer sanctioned for hiding invisible instructions in a legal filing aimed at manipulating an AI system into agreeing with it — an old trick dressed up as a new one. The common thread with every version of that trick, whether it's hidden text or an overstated product claim, is the same: it works right up until someone checks. A "finished" reminder system that's actually still being refined is a smaller version of the same dishonesty, and it tends to get caught the same way — the first time it doesn't quite work as promised.

Saying plainly what's proven and what's still in progress isn't a weakness to manage around. It's the same standard applied consistently everywhere else on this site: schema that matches the visible content exactly, pricing that's checkable against Meta's own invoice, a chatbot described as rule-based instead of oversold as something it isn't. A business considering any of these three systems can trust the description of what's live today precisely because the description doesn't inflate it.

Google itself has said plainly there's nothing special a site needs to do specifically for AI Overviews or AI Mode beyond the same fundamentals that make a page good at ranking normally. hexweb.dev builds to that standard directly, on every page including this one:

This is also why the three systems above are built the way they are, not just described that way for marketing purposes. A rule-based chatbot gives a predictable, auditable answer — the same property that makes a page trustworthy to an AI system deciding whether to cite it. A self-hosted reminder system doesn't depend on a third party's uptime or pricing — the same independence this site has argued for repeatedly when covering platforms like WhatsApp changing their own rules unilaterally. The reasoning isn't three different philosophies bolted together; it's one philosophy — own what you can, be honest about what you can't yet promise, and build fast — applied to a website, a chatbot, and a reminder system in turn.

Want the detail behind this? See what SEO actually means in 2026 and what Google AI Mode actually is, or the ongoing News section tracking this as it develops.