12 / AI network
twocents AI.
The problem
Deals like fundraising, hiring, partnerships and inside sales come down to reaching the right person with the right message. Most of that happens in private, in email threads and warm introductions, and it is hard to see from the outside. People who needed something had no good place to say so, and people who could help had no easy way to find them.
Public social networks were noisy for this, and email was crowded. A new network had a cold start problem: a feed with nothing useful in it would be abandoned within a day. It also had to be quick to use, because the people it served were busy and would not fill out long profiles. Early on, the design also had to help a small company explain itself to investors.
Success meant a person could post what they needed in a minute and get replies that moved a deal along.
The approach
I began with a concept that fit on a handful of screens, put it in front of the team, and kept the work small enough to change quickly.
- 01
First concept builds
The first builds were a dark, low-detail web app with a sidebar and an inbox sorted into Priority, Follow Up, New Connections and Notifications. I kept the screens gray on purpose. The question I wanted answered first was whether sorting by importance made sense, before anyone argued about color or type.
First conceptual builds of the web app. - 02
Priority messages
I rebuilt the inbox around a single idea: show what needs a reply from your top contacts first. Each thread stacks its recent messages, with Reply, Archive and View on the right and an Archive All button on top. The count beside Priority Messages told a person how much waited for them.
Priority messages. - 03
The opportunity feed
The home screen became a feed of opportunities. A box at the top asks people to start a post, with Find Work, Hire, Find Customers, Fundraise and Ask Advice one tap away. Each post shows who wrote it, where they are, what they need and tags. Message sits beside the other actions on each post.
Current design: Opportunity Feed. - 04
AI replies on a post
The AI reply control sits in the row of actions under each post, beside Message. I placed it there so a reader could ask for a drafted answer at the moment they decided to respond. The AI wrote suggestions for priority messages as well, so people started from a draft and edited it.
Current design: AI Replies. - 05
Posting through a chatbot
Posting was the biggest drop-off risk, so I replaced the blank form with a chat. A page titled Create a new post opens with one question about which business goal the person wants help with. A template list for hiring, job seeking and connections sits in the sidebar for people who prefer to pick.
Early chatbot designs. - 06
A live chatbot prototype
I then built a working version that appears as a dialog titled Post a New Opportunity. It greets the person, offers suggestion chips for hiring, job seeking, investing, fundraising, selling and networking, and gives sentence starters in the text box. A person who did not know what to write could finish a sentence.
Initial chatbot prototype, live build.
The solution
twocents was a web app with a few clear parts. A feed showed opportunities people had posted: hiring, fundraising, partnerships and sales. A posting flow let people describe what they needed through a chatbot. A priority inbox surfaced the messages that mattered, and AI wrote reply suggestions so a person could answer quickly.
The design idea was to make every step start from a prompt and not a blank page. The chatbot asked for a goal, the chips offered likely goals, the sentence starters began the post and the AI began the reply. Each one cut the effort of the first move, which was where private networking usually stalled.
What I delivered
- Concept designs and first web app builds
- Opportunity feed and priority inbox
- AI-powered email reply designs
- Chatbot designs and a live prototype
- twocents logo and brand
Value added
For people using it
A person could say what they needed without filling out a form, and see the messages that mattered first. Drafted replies helped them answer quickly, which suited people who were working deals between meetings.
For the business
The product gave a young company something concrete to show. Working screens, a chatbot prototype and a brand made it easier to explain the idea to investors and carry it through seed funding.
For the team
As the founding designer I led the design effort and kept the live prototype close to what engineering was building, so decisions could be made against something that ran.
The result
twocents reached seed funding and a 1.0 launch. The company has since shut down.
What I took from it
The chatbot prototype taught me to build the real interaction early, because a clickable version answered questions a static screen could not. I would also decide sooner what a new network looks like on day one with few posts in it.