Availy Case Study: AI-Powered Dental Product Design

CASE STUDY

Availy

The native booking platform that ended my chapter at Dental21

An honest account of product thinking, tradeoffs, failures & impact.

Daniele Dalia

Team of 8: PM, UX/UI, AI ENG, DEVS, QA

1≠1

6

PREMISE

1 ≠ 1

What happens if two non-contradictory truths exist simultaneously?

2

CONCEPT

1 ≠ 1

Structural Tension

In any company, two non-contradictory truths can exist simultaneously.

TRUTH A

our product should be like this

≠

TRUTH B

our product should be like this

3

REALITY CHECK

1 ≠ 1

This formula shows the normal reality in any company.

4

ROLE

The Mature Designer

has the specific responsibility to navigate this tension and remove the obstacles from production.

misalignment & ambiguity

structural tension

conflicting ownership

unclear strategy

5

AGENDA

What's inside.

01

Context

What is Dental21 · What is Availy · Why it matters

02

The Problem

65% drop-off · Internal conflict · Why the obvious fix was wrong

03

The Strategy

Governance first · Discovery · The insight that changed everything

04

The Solution

Symptom-based AI booking · Team structure · How we built it

05

The Outcome

+24% conversion · −70% ops load · What I'd do differently

7

01 — CONTEXT

What is Dental21

A full-stack dental platform combining patient experience, clinic operations, and AI tools.

PATIENT EXPERIENCE

Websites · Online booking · Reminder messages · Digital anamnesis · Health Assistant · Recommendations · Appointment rating · PROMs

CLINIC EXPERIENCE

Claire PMS · Calmaster · ToDo list · AI caries detection · AI cephalometry · AI documentation

OPERATION EXPERIENCE

Shift Planner · Salary Calculator · Employee mobile app · Time tracking · Access Control · System config

8

01 — CONTEXT

What is Availy

A native booking software with key qualities that differentiate it from generic booking tools.

Patient data stays in-house

No third-party data sharing

Higher revenue per appointment

Optimized for bundled treatments

Upfront Data Sharing

Patient info collected before visit

Bundled treatments

Cross-sell and upsell at booking

9

02 — PROBLEM

Challenge 1

65% of visitors abandoned the booking flow

Treatment type

City

Clinic

Date & Time

📞 CALL THE CLINIC

65%

drop-off rate

10

01 — CONTEXT

Availy Structure: Cognitive Issues

The same clinics presented in 4 different orders — creating confusion and decision paralysis.

ORDER 1

Trending Order

Lacks transparent criteria. Users struggle to understand who defines "trending" and why these clinics are pushed first.

ORDER 2

Alphabetical Order

Irrelevant for exploration. Helpful only if the user already knows the exact name of the specific clinic they wish to book.

ORDER 3

Distance Order

Overrides availability. Proximity is useful but forces users to manually filter out nearby clinics that have no open slots.

ORDER 4

Time Order

Prioritizes immediate empty slots, completely ignoring practitioner quality, user preference, and travel distance overhead.

No consistent logic = user confusion = abandoned booking

11

02 — PROBLEM

Two Non-Contradictory Truths

At this stage everyone wanted a solution. These two sides were pulling the product in opposite directions.

CEO

"One-click booking. We need revenue now."

Focus: speed, conversion, empty chairs → revenue

✓ Locally correct

MEDICAL DIRECTOR

"Add 10 questions. We need safety."

Focus: compliance, classification, clinical accuracy

✓ Locally correct

≠

15

03 — STRATEGY

Strategic Call

I negotiated a 20% roadmap reduction to fund a 2-week 'High-Velocity Discovery' phase.

WITHOUT PAUSE

Ship fast

6 months of failure

Rebuild

WITH PAUSE

2 weeks discovery

Right direction

Lasting outcome

2 WEEKS

16

03 — STRATEGY

Discovery Methods

01

Funnel Analytics

Told us WHERE the drop-off was happening.

02

Competitor Benchmark

Doctolib, Jameda — limited signal from comparison.

03

Passive Listening

Shadowing reception calls. This is where it clicked.

04

User Interviews

Patients described confusion in their own words.

Passive listening

revealed the real insight.

17

03 — STRATEGY · KEY FINDING

9

Pain, not endodontics.

The interface was built in the language of dentists, presented to patients.

When confused, they called. Every time. The phone was easier.

9 calls per confused patient

19

03 — STRATEGY

Treatment → Symptom

A Human-Centric Shift

BEFORE

Treatment standard booking

Users must know their diagnosis before booking

→

AFTER

Symptom-based guided booking

+ AI agent for text

Users describe how they feel, AI maps to treatment

By reframing around symptoms, we satisfied both the CEO and the Medical Director.

20

03 — STRATEGY

Fix the structure before touching wireframes.

We moved from 'everyone is responsible' to explicit, documented accountability — ARCI Workshop (Auftragsklärung).

21

Area

Accountable

Responsible

Consulted

Informed

Foundation Research

Design Team

Research

Medical, Ops

Leadership

Medical Classification

Medical

Research

Design

Product

Tech Feasibility

Tech + Design

Engineering

Product

Leadership

03 — THE SOLUTION

From static form → symptom-guided, AI-assisted booking

OPTION A

Symptom-based guided booking

Higher conversion, human-friendly

Needs medical validation layer

OPTION B

Pure AI chat agent

Most flexible, most natural

Hardest to build, highest risk

OPTION C

Hybrid: symptom entry + AI routing

Balanced. Satisfies CEO + Medical Director.

More complex to spec

CHOSEN

22

RESOLUTION

1 = 1

By reframing the interface around symptoms instead of treatment, we could satisfy both the CEO (higher conversion) and the Doctors (accurate medical intent).

23

03 — THE SOLUTION

Early-Engage

Engineers and designer define standards before wireframes.

01

Logic Pod

Adaptive logic, symptom-to-treatment mapping, AI scoring parameters

02

Trust Pod

Patient-facing tone, transparency, consent flows

03

Validation Pod

Medical accuracy, edge cases, error states, clinical QA

24

03 — THE SOLUTION

How the AI booking agent works

PATIENT INPUT

my tooth hurts when I drink cold things

01

Symptom identified

Cold sensitivity

02

Patient profile

Existing patient · Adult · Public insurance

03

Urgency flag

Non-urgent — Factor 3

04

Treatment mapped

Filling / cavity check

05

Appointment suggested

22 Jan 15:30 · Dr. Muster · Clinic Wedding

26

04 — THE OUTCOME

Metrics

+24%

increase in completed digital bookings

from Doctolib Data

−70%

reduction in manual 'cleaning' calls by reception

Ops Efficiency

+31%

increase in classification accuracy at point of booking

Medical Safety

Design repositioned

from 'service desk' to strategic partner that solves business-critical conflicts

Strategic Value

Integration on a bigger platform: Health Assistant

28

04 — OUTCOME

Takeaways

Problem framing before solution

The real issue was vocabulary and trust — not flow length.

Structure before screens

RACI before wireframes.

Cross-functional as a design skill

The Early-Engage sessions were design work, not prep for design work.

Honest tradeoff-making

I traded speed for quality. And paid a real cost. That's product thinking.

Design as strategic partner

We moved Availy from a web form to a Human-AI collaboration platform.

29

04 — THE DRAMA

Why did I lose the job

AI Module → Phone Adaptation

5 languages rollout accelerated the scope beyond original team capacity

Website 2.0 → New Standard

Platform was set as new benchmark requiring full redesign

New Market Opportunity

Company shifted strategic direction mid-project

I traded speed for quality. And paid a real cost. That's product thinking.

30

END

Thank you.

Daniele Dalia · Product Design

Case Study: Availy · 2026

1≠1

31