I initially tested the service with 11 women, over a period of 5 days.

Testing the tone of the chatbot, facilitating the correct matching and the rhythm of the conversations and looking into the frequency and how successful using prompts were. Looking into the Content Vs. Conversational style.

User break-down & matching

Separate groups, testing mediums and conversational styles:

Group chat on whats app. (Starting prompt being from the user asking for advice)
- Using prompts such as: pillar here! It looks like X has asked a question, please feel free to share your answer via audio on whatsapp, it is easy to use especially if you are limited with time and will share a more personal experience! Try it out :)
Group chat via text when I am the moderator and all questions go through me. More Anonymous.
Group chat on iMessage: (Start with advice from Mentor)
- Using prompts such as: Here we encourage you to discuss and share back any experiences you have with motherhood. Xdo you have any follow up questions to ask Y?
Group chat on facebook.

I wanted to learn about the characteristics of the questions being asked and to move forward from the standard questions into more provocative questions being prompted by myself or the users.

Tools:

Send fake questions as prompts found in my earlier research.
Change the tone of questions (start by more practical questions, into more stressed signals, super personal) to see where I get more push back and where they work?
Change the timing of the questions & prompts (seeing when the users are more active)
Asking more provocative questions

Using Textit to build the chatbot flow.

Allowing me to use SMS and FB Messenger channels to onboard the users, also the possibility to then save the data into a JSON file directly from Textit to then feed into the match-up process and redirect them to a personal messaging URL.

Follow up meeting all respondents to ask the following questions:

How did you feel about the questions asked to match you? What could have been improved?
How did you find the matching up? What additional information is needed to know about the person?
What did you feel about the general tone and language used from the chatbot?
What medium is preferred? (Medium such as: Social Networks they are already using, SMS, New App) Looking into if they want their data private or can be accessed via existing platforms.
What would make you stop using the service? What would make you continue to use the service?

In-depth Interviews with prototype users

Freja - Role: Mentor & Mentee

Freja used different medium, started on SMS and used iMessage with one match, and whatsapp with another.

Liked that the on-boarding questions were y/n answers or basic answers, it took the pressure off of the initial sign-up. An insight was that first-time mothers do not have alot of time and often when trying to sign-up to something it is a long process they never actually finish.
Some more questions around; pregnancy and the birth, as well as if they were late and if there were any problems to better match-up to the right women to help/seek advice from.
The chatbot was relatable and emotional, friendly, relaxed and non-judgmental.
She currently uses many apps and therefore wouldn't mind downloading another one, although liked the basic ease of SMS.

Monika - Role: Decided not to continue with the service.

Initially Monika was unsure the expectations of the service and therefore did not want to be matched up to anyone but allowed me to come and ask her why she felt this way.

Would like profession to be a key matching point.
In order to be enticed to use the service, she would have liked to have seen some tips that have come from people using the service as well as expectations of the user. Such as "on the 4th floor of Illum they have a baby feeding zone"
The matching up service should be a selection process, with an option of a few mothers in the same network as you that you can see a bit about them; profile photo and general points of interest.
Looking at the metaphor of the health visitor it would be much better of a match-up as she knows every mother in the area on a personal level. You trust the health visitor, and you would trust the matching much more than this service.

“Did Alison have her baby? I was wondering if I could text her to find out, or if that was against the service”

Alie - Role: Mentor & Mentee

Really liked how the service was a-synchronized, and that it lived purely on a message type format. However, she is worried that visually on SMS the text may intimidate the reader if it is too long, especially as mothers tend to text when the child is breastfeeding and just let all of the information ramble out.
Education actually doesn't matter in the matching, however something more like profession and if the mother is a stay-at-home mum, or if they are breastfeeding or not. Language is also something to consider when scaling the service.
"I felt more useful" when she was the mentor.
Alie can see the mentee being a useful option as and when you need it, but being prompted for a question to her future self was hard for her to come up with.
The service felt anonymous without being anonymous which actually helped, she felt like she trust the system as well as the vetting and verification of the users.
Alie would stop using the service if there were too many questions or prompts, maybe the service limits the amount of people you can match with to stop this fear. Is there a way for the user to be 'active' or 'non-active' or to send themed prompts such as "I am sorry I am busy taking a walk, I will answer you when I can"

Skype follow-up.

Line - Role: Mentor

Line felt that onboarding questions were very welcoming and friendly and she was a little surprised as to how easy and non-intrusive this was.
"Level of Education" was a weird and irrelevant question, maybe something more like, what does it mean to you to be a good mother?

“Here we encourage you to discuss and share back any experiences you have with motherhood. Freja do you have any follow up questions to ask Line?”

— prompt for Line matching

She felt that this prompt was a little closed, and therefore didn't know if she could answer and therefore didn't. Would have liked to have a more open ended prompt.
"I liked I was able to help someone" that was value enough in the service. She often felt that she was answering questions to her past self, because the service felt 'faceless' but in a good, anonymity way.
SMS she liked because it was easy, basic and felt like no-pressure as to when to respond.
She would continue to use the service if the initial questions were not so personal but as the relationship developed the questions could become more developed.

Carly - Role: Mentor

Felt that education is also not relevant as a matching data point, but instead more personal questions such as; if you were induced, breastfeeding.
Would be interesting to know about nationality of the women and also the location needs to be the same, as there are many differences between countries and pregnancy.
Carly liked being in a mentor position, she felt comfortable especially because the question was in a 'safe' field, however if she was confronted with more personal questions she would be a little more weary.
Prompts really helped, genuinely as a second-time mother you have very little time and the prompts helped remind.
During the onboarding process, include things such as : when are the times you are around the most, to utilize when to send and receive prompts.
Engagement would continue if the person needs advice, as then when they are active they will also respond to questions. Right now she felt that because she doesn't need to know anything, her engagement will drop as she doesn't have the time to commit to just answering people.

Key insights

The women preferred to be the mentor, and gave them a purpose in the service. "I liked I was able to help someone"
pillar provided a very friendly, non-judgmental tone. The women often forgot they were talking to a 'chatbot' and felt at ease with the questions and prompts.
SMS was the best form of communication, it felt easy; they were short answers; no pressure; faceless.
"It was like I was giving past-self advice.. it was easy to relate too and provided ease and comfort"
Level of Education asked was weird and not very relevant. Profession/Stay-at-home mum/problems with pregnancy and birth would have been more suitable.
An insight was that first-time mothers do not have alot of time and often when trying to sign-up to something it is a long process they never actually finish, keeping it short and concise matters.
In order for engagement to remain with the users; the service should only target women who are in need as well as who want to give advice; be available as and when needed with no time-pressure;

To find out more:
- Initial Research Methods
- Design Challenge
- Initial In-Depth Interviews: Data Matching
- Experience Prototyping: Wrappers and Value Proposition
- User Testing: Content and Conversational Style
- Building the back-end.
- Validated Concept
- Final Product: www.ourpillar.com

pillar

Prototyping 3 - Content & Conversational Style