2023 3 months
Those considering becoming parents are uncertain about what they need to become self-efficacious parents. Those who are parents of young children are overwhelmed about the present and scared about their future. There are many resources available to offer advice and information, but none are cost effective and interactive.
Ickrus helps soon-to-be parents navigate their fears for the journey to come.
I wanted to gain insight on my core user group so I solicited interviewees through survey screeners. From posting in online Facebook and Reddit parenting groups, I received 33 screener responses and interviewed 4 parents.
A recurring theme was that parents:
1) spent hours online searching for parenting advice and
2) were uncertain the advice was credible.
It was clear that parents wanted a platform that gave them parenting advice that was relevant and credible.
Making a platform for parents to ask questions from credible sources seems like a simple solution. Sources like ChatGPT, Google, or Reddit often seemed to have specific answers, but it wasn't clear if the information presented was both credible and relevant.
Parents needed an AI chatbot that was trained from credible information. Sounds simple, but for someone who can't code- doing so wasn't so clear to me.
So I used a no-code program, Flowise, which enabled me to train ChatGPT with parenting research studies and advisory PDFs and TXT files. Flowise required using Render as a means of bringing the chatbot online and used Pinecone to host vectors. In the image above, you can see nodes between the PDFs and TXT fils with OpenAI's and Pinecone's plug-ins. The chatbot shows a history of error messages until finally receiving the answer to a question on helping children sleep.
Throughout this project, I learned to let go of my inclinations to implement a vision because of an interesting technology. Whether it was a simulation technology or complex AI, emergent technology is only productive as a medium for a solution. I experienced some roadblocks.
This stage of the project introduced and taught me to learn no-code platforms like Render, Pinecone, Flowise, Glide, and Framer. I was also exposed to Adalo and Webflow, but had not pursued them as MVP platforms. Although, they may be considered as Ickrus scales.
Made with Flowise and Glide
A USC Directed Research Project
“Intentions are half the battle.”
E., 27. Mother of two children under 5.
“When other people online comment that they are having a similar experience...it tells me that it’s not just me.”
A., 35. Father of a child under 5.
I asked 10 users to type in the same parenting related question of their choice into Ickrus, Google, and ChatGPT. From the screenshots they uploaded (and from user comments), it was clear that Ickrus provided concise feedback. Some users preferred this, others thought it was a pain point. Many users thought Ickrus was desirable, but the feedback that Ickrus needed an additional component like a parenting forum, stood out.
User Tester's Question:
"How many words should my child know at 5 years old?"
The second test presented Ickrus as a single use site for a chatbot. I asked 25 users if they would use such a service, what was missing, and how much they would pay for it. Users on average said they would pay a median of $50/year.
This felt encouraging, but I couldn’t see how Ickrus fell into the habits that users currently had with using social media daily. Ickrus was a single use “search engine” of sorts, so it wasn’t as though any other curiosity brought users to Ickrus unless they remembered.
This made the idea of a parenting forum more appealing.
Parents were asking for community in addition to advice. Yet, the market is saturated with communities. I wasn't sure there was a need for innovation in this area, so I focused on one aspect of community that groups provide: self-reflection.
Creating a workbook for parents to reflect on their concerns, ask questions, and make plans would assist parents in novel and feasible ways.