Paid traffic to one offer
Facebook and Instagram ads point to a dedicated glutathione IV drip landing page, which adds enquiries to Zoho CRM.
Client work · Website AI chat
MB Aesthetics: the 92% who never filled in the form
The clinic's glutathione IV landing page converted about 8% of visitors into enquiries. The other 92% showed interest and left. We built the AI concierge that answers their questions, checks real availability in Zoho CRM and books them — before they close the tab.
Zoho CRM booking
TimelinesAI WhatsApp
Live availability
At a glance
- Client MB Aesthetics Clinic — London W2, IV drip and skin treatments.
- Surface AI chat widget on the Glutathione IV drip landing page.
- Stack Zoho CRM Events, n8n, TimelinesAI WhatsApp, OpenAI.
- Outcome Visitors can get answers and a confirmed booking without leaving the page.
The client
A specialist clinic with a trust-heavy offer
MB Aesthetics is a London clinic led by founder Marie, a Level 5 aesthetician with over a decade of experience, certified in phlebotomy and IV treatments, and specifically experienced with Black and dark skin, hyperpigmentation, melasma and dark spots.
A considerate purchase
IV treatments raise real questions about suitability, safety and process. People research before they enquire — and research quietly.
WhatsApp-native clients
Confirmations, changes and cancellations already ran through TimelinesAI on the clinic's business WhatsApp — a flow worth preserving.
The problem
The form was the only way to ask — and most people would not fill it in
The clinic's own figures framed the build.
- Only about 8% of landing-page visitors completed the enquiry form — roughly 120 enquiries in 30 days.
- The other 92% had shown interest in the offer and left with zero interaction — around 1,250 people in the same month.
- Simple questions about suitability, process or aftercare had nowhere to go except a form the visitor did not want to complete.
- Manual callback requests were slow, and contacting leads afterwards was difficult.
- Any booking flow had to respect the clinic's existing Zoho + TimelinesAI WhatsApp setup — not replace it.
Why it happens
A form is a commitment. Many visitors are not ready to commit — they are ready to ask. Without a way to ask, interest evaporates silently and the ad spend buys nothing.
Numbers in this section are the clinic's own reported figures from the brief.
What we built
A concierge that only offers times it can actually book
The system was rebuilt around one hard rule: the assistant may only offer or book slots returned by the live calendar, and it must always reply after a tool call — never end a turn on silence.
Availability from the real calendar
A sub-workflow reads the clinic's CRM meetings and computes free slots: Monday–Saturday 12:00–18:00, an 8-hour minimum lead time, and sensible caps on how many options are shown at once. Date and time constraints the visitor states stay sticky across the conversation.
Deterministic booking
Booking was moved out of the AI's hands into a fixed sub-workflow: upsert the contact, build the meeting payload, create the CRM event, and return a clear result. The model cannot skip a step or book on a guess.
CRM connected the way the clinic works
The audit found the first build wrote to a separate Zoho Calendar app while the clinic's real diary lives in Zoho CRM Events. The rebuilt flow targets the CRM module directly — the same place manual bookings and WhatsApp confirmations already live.
WhatsApp confirmations preserved
Meetings are linked to the contact so the clinic's existing CRM workflow rule fires TimelinesAI's WhatsApp confirmation and change messages — the client's system, untouched, now fed correctly.
Compliance-shaped conversation
Plain text only, replies capped at three short sentences, one question at a time, no medical claims outside the approved knowledge base, confirm-before-booking, and a strict no-home-visits policy stated positively.
Callback capture and recovery
If a visitor is not ready to book or a tool fails, the assistant captures a callback request with contact details instead of losing the lead — and never invents times while doing it.
Evidence
The numbers behind the build
Two kinds of numbers: the clinic's reported conversion figures that framed the problem, and the structural facts of the system that was delivered.
~8%
of landing-page visitors completed the enquiry form (client-reported)
~1,250
interested visitors in 30 days who left with zero interaction (client-reported)
6 days
of the week covered by live availability (Mon–Sat, 12:00–18:00, 8h lead)
2
fixed booking sub-workflows: availability and appointment creation — no improvised steps
Verified build facts
The 8% / 92% figures are the clinic's own reported numbers, labeled as such. The availability rules and workflow structure come from the delivered build. We do not claim post-launch conversion changes — those were not part of what we measured.
Zoho CRMn8nOpenAITimelinesAIWhatsApp Business
How a conversation runs
Ask, qualify, book, confirm
1
Visitor asks
The chat widget answers from the clinic's approved content, in plain text.
2
Constraints kept
Stated day and time preferences are remembered for every later offer.
3
Live slots
Real free times come from the CRM calendar — never from memory.
4
Confirmed booking
Read-back, yes, then the fixed booking flow creates the CRM meeting.
5
WhatsApp confirm
The clinic's existing TimelinesAI flow sends the confirmation message.
Honest limits
What this case study does not claim
No post-launch metrics
We show the problem numbers the clinic reported and the system we delivered. We do not claim the 92% became bookings — that depends on traffic, offer and follow-through.
Medical caution by design
The assistant answers only from the clinic's approved knowledge and never gives medical advice. Suitability decisions stay with the qualified practitioner.
Client systems respected
The clinic chose to keep its current WhatsApp provider rather than migrate to the API. The build works with that decision instead of forcing a platform change.
FAQ
Questions about this build
Couldn't the AI just make up available times?
That is exactly what the architecture prevents. The assistant may only offer slots returned by the availability workflow reading the real calendar, and booking runs through a fixed sub-workflow that creates the meeting and returns a result. If a tool fails, the assistant says so and captures a callback instead of inventing a time.
Why did the original calendar connection have to be rebuilt?
The first version wrote to a separate Zoho Calendar app, while the clinic's actual diary lives in Zoho CRM Events — the module its manual bookings and WhatsApp confirmations use. The two do not sync, so bookings were landing where nobody looks. The rebuild targets the CRM Events module directly, so the same workflows that always ran now fire on AI-made bookings too.
What happens if someone asks about a medical condition?
The assistant answers only from the clinic's approved knowledge base and does not give medical advice or dosing guidance. Anything outside that scope gets an honest "let me have someone confirm with you" and a callback capture. The practitioner makes every suitability decision.
Get your plan
How many visitors leave your landing page without a word?
Tell us where your traffic goes and what it asks. On a free strategy call we will map the fastest way to turn silent interest into booked appointments.
- A free 15-minute AI strategy call — no obligation.
- A scoped blueprint: the biggest leak, and the first automation to fix it.
- Honest fit check — we will tell you if automation is not your fastest win.
The fastest way in: answer six short questions and we will map your landing-page journey, find the biggest leak and show the fastest win — before we ever get on a call.
Free strategy call · No commitment · We reply by email
Keep going
Related pages
More client work
Relevant AI services
Industry pages
Ready to catch the visitors who leave without asking?
Book a free AI strategy call and we will walk your landing page against this build.