Overall, I’m very grateful for this assignment because I had been wondering for the past year how to find time to finally learn about chatbot development. Although it’s not directly related to my primary master’s thesis topic, it still connects to part of it. As I mentioned earlier, I plan to test the bot with private doctors or small businesses. If it works as I hope, I would love to integrate it into my design service offerings.
Tag: Chatbots
15. Setting up the chatbot: Final tweaks
After trying several tools and platforms for building a WhatsApp-based booking assistant, I decided to go with n8n, a powerful automation platform that gave me more flexibility and control over the workflow.
I started by uploading a simple website template from Framer.com, which allowed me to embed the chatbot directly into a webpage. This gave me a smooth front-end experience where I can test the interaction with the assistant as if it were part of a real business website.
Then, I configured ChatGPT within n8n to act like a helpful assistant. I trained it to ask and collect key pieces of information needed for scheduling an appointment:
- Preferred date and time
- Full name
- Email address
- Final confirmation
For now, the assistant is connected to my personal Google Calendar, which I’ve labeled with a mock clinic name. Eventually, I plan to replace it with a real calendar once I find a doctor or clinic to partner with.
One important technical note: this chatbot setup also works perfectly with Google Calendar, so appointments are created automatically without the need for manual input.
At this stage, everything is functioning smoothly. The next step is finding professionals who could benefit from this tool.
Initially, I was planning to pitch the idea to doctors, but I’ve realized the potential goes far beyond the healthcare field. This kind of chatbot could be incredibly useful for any entrepreneur or small business owner who books appointments manually, like:
- Lawyers
- Therapists
- Beauty salons
- Coaches
- Freelancers
- Small companies offering services
To test the waters, I recently showed the prototype to a beauty practitioner, and she’s already interested in buying it. She told me she spends a lot of time answering basic questions from clients, things like availability and pricing, which the chatbot could easily handle. She loved the idea of freeing up that time to focus on her actual work.
I’m excited about this pivot. Over the summer, I plan to stick with this idea, improve the product, and start selling it to service providers who need smarter scheduling tools.
Let’s see where this journey leads next.
14. From WhatsApp to Web Chatbot: Adapting Plans with n8n
Sometimes, the most interesting part of a project is when things don’t go as planned.
I originally set out to build a simple booking system via WhatsApp, using a chatbot to help users schedule appointments. I explored various platforms, including Twilio, Zapier, and specialized tools like Libromi, which offer multi-agent support, payment integrations, and booking systems via WhatsApp. My plan was to test a single hypothesis, so I decided to build the core logic calendar booking myself.
However, to use the WhatsApp Business API, I needed a Facebook Business account. That’s where things took an unexpected turn: my Facebook account was suddenly blocked, and with it, the API access was gone. I have written an email to the customer support but they didn’t reply(
Pivoting to a New Stack
So, I started exploring alternatives. That’s when I discovered n8n, an open-source automation platform that provides visual workflow creation, powerful integrations, and complete ownership over the process.
Instead of relying on WhatsApp, I decided to embed the chatbot directly inside a website. Since many doctors in Austria have websites, it’s still a working hypothesis. Here’s what my updated setup looks like:
- n8n: To control and visualize the conversation and booking flow.
- ChatGPT API: For natural, conversational responses.
- Google Calendar API: To check availability and schedule appointments.
- Embedded web widget: So users can interact with the chatbot on a website, without needing to go through WhatsApp or Facebook.
What’s Next?
This new approach is actually more flexible than what I originally planned. I don’t need third-party approval, and I have full control over the chatbot’s tone, interface, and features.
Next steps include programming the chatbot to respond to particular responses, testing a real-time booking, and possibly adding reminders or email confirmations. Even though the platform shifted, the goal remains the same: to build a smart, lightweight, and user-friendly appointment system
13. Exploring tools: Connecting WhatsApp and Google Calendar
It all started with a simple idea of booking appointments just by chatting on WhatsApp. I wanted to test this hypothesis without diving into complex paid tools right away. So I rolled up my sleeves and began researching how to connect WhatsApp with a calendar system in the most efficient, hands-on way possible.
As I combed through forums, websites, and YouTube videos, I came across platforms like Libromi. This company specializes in implementing chatbots for messaging apps. Their tool stood out because it offered something I wasn’t even looking for at first: a multi-agent chat system. This means multiple people can manage customer support and sales conversations using a single WhatsApp number, while keeping performance and access control in check. It was impressive.
Libromi also offered neat perks: integration with Google Sheets, payment gateways, automation for Facebook/Instagram lead ads, and most importantly, a ready-made booking system. Tempting? Very. But I decided not to use it. Since my focus was testing just one specific idea, I challenged myself to build the booking feature on my own instead. + Why pay 50$ per month for functions that might not be used 🙃
Before diving into development, I needed to make a key decision: Should I use the WhatsApp Business App or go for the WhatsApp Business API? Here’s what I learned:
The WhatsApp Business API is where the real power lies. With no device limits, full automation, integration support, and no broadcast caps, it’s made for businesses that want to scale. Naturally, this was the right choice for what I had in mind.
The WhatsApp Business App is great for small businesses. You can connect up to 5 devices, but it’s not designed for automation or integration. There’s also a limit on broadcasts (256 contacts), which can quickly become a bottleneck.
From Idea to Execution
Next, I needed to bridge WhatsApp and Google Calendar using the API. After evaluating several tools, I found that platforms like Zapier and Twilio could help make the connection smoother. They allow you to set up triggers and automate workflows, like creating calendar events based on user messages.
But here’s where I hit a roadblock: I couldn’t connect a chatbot to just any WhatsApp account. It has to be a verified business account through Facebook. So I created a Facebook Business Account, which is a must-have if you want to access the WhatsApp API and integrate it with any external tools.
What’s Next?
At this point, I’ve laid the groundwork: I understand the tools, I’ve set up the business account, and I’ve mapped out the data flow. The next step is to actually build the interaction where a user can chat with the bot, choose a time slot, and get a confirmation automatically logged into a calendar.
12. “Sprechen Sie Deutsch?” The Hidden Barrier to Booking Appointments
When I arrived in Austria as an international student, I expected that most of my challenges would come from my still-developing German skills (aber ich lerne es!). And ofc as it always happens after a few months of my arrival, I needed to see a doctor.
But to my surprise, the real problem wasn’t finding a doctor.It was booking the appointment.
Most clinics required a phone call. And most receptionists spoke only German. After four failed calls, I finally found a local doctor who could help and book an appointment. Lucky me!
But when I spoke to other international students, I realized this wasn’t just my problem:
“After a few calls, I gave up and asked my Austrian friend to call.” – Orlaith, 22, Ireland
“I accidentally booked a vaccination instead of a check-up.” – Younes, 23, Algeria
“Yes, I booked the appointment… but it took 40 minutes!” – Elske, 30, Netherlands
This process is frustrating for everyone involved: patients feel confused and helpless, and clinics lose time (and possibly clients) and money
So I started wondering:
- What if booking an appointment didn’t require fluent German?
- How could this process be simpler and less stressful?
- Could we implement ideas from other countries’ systems?
Most websites in my country offer virtual assistance or a chatbot on messengers. So, what if I develop a chatbot that works in both German and English? It would guide international patients step by step, collect necessary details, and ease the burden on receptionists without requiring extra staff or app downloads.
More in the next post…
1.9 The Emotional Intelligence of AI: Can Chatbots Truly Understand Us?
As AI technology advances, chatbots are evolving to recognize emotional cues, providing support in mental health, companionship, and conversational interfaces. By integrating techniques such as natural language processing (NLP), sentiment analysis, and machine learning, these systems aim to simulate empathy and create meaningful interactions. However, the development of empathetic AI comes with challenges, including technological limitations, ethical concerns, and potential risks of over-dependence.
Advancements in Empathetic Algorithms
Empathetic algorithms are designed to detect, interpret, and respond to human emotions using methods such as NLP, voice tone recognition, and facial expression analysis. For example: Woebot employs cognitive-behavioral therapy (CBT) techniques to guide users through stress and anxiety management, leveraging emotional cues from conversations. Wysa uses sentiment analysis to provide customized mindfulness exercises and mood tracking tools for emotional resilience.
Beyond mental health, empathetic algorithms are being integrated into other sectors like education and customer service, tailoring interactions based on emotional cues to improve engagement and satisfaction.
Chatbots as Relationship Simulators
LLMs such as GPT power chatbots like Replika AI and Character AI, which simulate human-like relationships. Replika AI enables users to design virtual companions for friendship, mentorship, or even romantic connections, raising questions about emotional reliance and blurred boundaries between humans and machines. Character AI allows users to interact with AI representations of fictional or historical figures, blending entertainment with relationship simulation.
These developments reflect themes from the movie Her, where an AI operating system becomes a deeply personal companion. While such systems offer emotional support, they highlight risks like over-dependence, which could potentially hinder real-life emotional interactions.
The Role of Empathy in AI
Empathetic AI is transforming human-AI interactions by making them more intuitive and emotionally aligned. However, achieving true emotional intelligence in machines remains a significant challenge:
- Complex Emotions: Emotions are shaped by individual, cultural, and situational factors, making them difficult for AI to interpret consistently.
- Simulated Empathy: Current AI systems simulate empathy by mimicking human responses rather than genuinely understanding emotions.
- Ethical Concerns: Privacy risks arise from AI’s reliance on sensitive emotional data, making transparency and data security essential.
Applications and Insights from Research
Recent studies emphasize how empathetic algorithms can enhance human emotional intelligence by fostering emotional awareness and resilience. For instance:
- Educational AI systems: Tailor learning environments to students’ emotional states, adapting content based on signs of frustration or confusion.
- Healthcare applications: Use empathetic AI to assess patients’ emotional needs and deliver personalized support, improving outcomes for individuals with anxiety or depression.
Despite these advancements, challenges such as cultural biases in emotion recognition and the need for interdisciplinary collaboration remain key areas for growth.
Sources
- “Character.ai: Young people turning to AI therapist bots.” BBC. Accessed: Jan. 24, 2025. [Online.] Available: https://www.bbc.com/news/technology-67872693?utm_source=chatgpt.com
- ” ‘Maybe we can role-play something fun’: When an AI companion wants something more.” BBC. Accessed: Jan. 24, 2025. [Online.] Available: https://www.bbc.com/future/article/20241008-the-troubling-future-of-ai-relationships?utm_source=chatgpt.com
- “Replika CEO Eugenia Kuyda says it’s okay if we end up marrying AI chatbots.” The Verge. Accessed: Jan. 24, 2025. [Online.] Available: https://www.theverge.com/24216748/replika-ceo-eugenia-kuyda-ai-companion-chatbots-dating-friendship-decoder-podcast-interview?utm_source=chatgpt.com
- Velagaleit, S. B., Choukaier, D., Nuthakki, R., Lamba, V., Sharma, V., & Rahul, S. (2024). Empathetic Algorithms: The Role of AI in Understanding and Enhancing Human Emotional Intelligence. Journal of Electrical Systems, 20-3s, 2051–2060. https://doi.org/10.52783/jes.1806
- “Woebot Health – Mental Health Chatbot.” Woebot Health. Accessed: Jan. 24, 2025. [Online.] Available: https://woebothealth.com/
- “Wysa – Everyday Mental Health.” Wysa. Accessed: Jan. 24, 2025. [Online.] Available: https://www.wysa.com/
1.6 How AI Is Reshaping Mental Health Support
Artificial intelligence is revolutionizing mental health care by breaking down barriers like cost, stigma, and accessibility. With features like chatbots, biofeedback, and voice analysis, AI offers innovative solutions for mental health support. While AI can’t replace human therapists, its ability to complement traditional care makes it a valuable tool.
Venture capital reports reveal that mental health is the fastest-growing marketplace category, with a growth rate exceeding 200% in 2023. This surge reflects a rising demand for accessible mental health solutions as AI continues to play a critical role in meeting that need.
How AI Powers Mental Health Apps
AI-Driven Chatbots
AI chatbots provide immediate, tailored support for users in need:
- Wysa offers CBT-based exercises and mindfulness prompts, creating a safe space for users to manage stress and anxiety.
- Woebot adapts its conversations to users’ emotions, providing tools for real-time mental health management.
- Cass combines emotional support and psychoeducation, offering adaptive responses that cater to individual needs.
In May 2024, Inflection AI launched Pi, a bot designed for emotional support and conversational companionship. Unlike other chatbots, Pi openly acknowledges its limitations, avoiding the pretense of being human while focusing on honest and straightforward interactions.
Wearables and Biofeedback
Wearable devices enhance AI’s ability to provide real-time insights into users’ mental states:
- Moodfit and Spring Health use wearable data, like heart rate and stress levels, to deliver personalized mental health strategies.
- Kintsugi analyzes vocal biomarkers to detect signs of anxiety or depression, offering users actionable insights based on their voice patterns.
These integrations bridge the gap between physical and emotional health, empowering users to take control of their well-being.
Opportunities in AI Mental Health Care
AI’s advantages lie in its ability to make mental health support more accessible, personalized, and inclusive:
- Immediate and affordable: tools like Headspace’s Ebb and Wysa provide around-the-clock support at a fraction of the cost of traditional therapy.
- Engagement and effectiveness: a 2022 review found that AI tools could improve engagement and reduce symptoms of anxiety and depression. However, experts emphasize that AI works best as a supplement, not a substitute, for traditional therapy. As Dr. Chris Mosunic of Calm explains, “Having a human in the driver’s seat with improved therapy AI tools might be just the right blend to maximize engagement, efficacy, and safety.”
- Personalized support: apps like Woebot and Youper adapt their recommendations to the user’s changing emotional needs, creating a more tailored experience.
Challenges and Ethical Considerations
While AI offers promising solutions, it also presents challenges:
- Limited empathy: AI tools often lack the emotional depth of human therapists, which can leave users feeling unsupported in complex situations.
- Bias and inclusivity: non-diverse training data can lead to biased responses, potentially failing marginalized communities that rely more heavily on these tools due to systemic barriers.
- Privacy concerns: AI tools require access to sensitive data. Apps like Talkspace use encryption to protect user information, but trust in data security remains a significant hurdle.
As these tools evolve, balancing innovation with ethical responsibility will be critical – a topic that will be explored further in upcoming articles.
Sources
- A. Fiske, P. Henningsen, & A. Buyx. (2019). Your robot therapist will see you now: Ethical implications of embodied artificial intelligence in psychiatry, psychology, and psychotherapy. Journal of Medical Internet Research, 21(5), e13216. https://doi.org/10.2196/13216
- A. Thakkar, A. Gupta, & A. De Sousa. (2024). Artificial intelligence in positive mental health: A narrative review. Frontiers in Digital Health, 6. https://doi.org/10.3389/fdgth.2024.1280235
- “Can AI help with mental health? Here’s what you need to know.” Calm. Accessed: Jan. 4, 2025. [Online.] Available: https://www.calm.com/blog/ai-mental-health
- “Meet Ebb | AI Mental Health Companion.” Headspace. Accessed: Jan. 4, 2025. [Online.] Available: https://www.headspace.com/ai-mental-health-companion
- P. Gual-Montolio, I. Jaén, V. Martínez-Borba, D. Castilla, & C. Suso-Ribera. (2022). Using artificial intelligence to enhance ongoing psychological interventions for emotional problems in real- or close to real-time: A systematic review. International Journal of Environmental Research and Public Health, 19(13), 7737. https://doi.org/10.3390/ijerph19137737
- “Rise of AI therapists.” VML. Accessed: Jan. 4, 2025. [Online.] Available: https://www.vml.com/insight/rise-of-ai-therapists