Shipping a product within 2 weeks | In-app customer support
Get in touch
Ask anything...
What can I navigate for you today?
Powered by
vector
How do I customize a template?
Click on the Template Center, find the one you want, edit to personalize the note, and Build.
Show me
Execute for me
Vector
An AI co-pilot for SaaS that automates in-app support, helping users complete workflows instantly and reducing support tickets.
Project Duration
2.5 weeks
Industry
B2B Saas
Customer support
Productivity
Team
Tools
Arman Arya (PM)
Figma
Notion
Brennen Ho (Developer)
Vishnu Kadaba (Developer)
CONTEXT
Starting as founders
Problem exploration
Before building anything, my team and I spent a week collecting frustrations from our own lives.
We mapped industries we understood and talked to people to see what problems were actually worth solving.
Our guiding questions
Is this a problem space we are equipped to create within?
Does solving this problem make sense within 8 weeks, our incubator timeline?
We had a shared, recurring story
A lot of our early pain points centered around bottlenecks.
PROBLEM (END-USERS)
Users abandon software the moment friction appears
Users get stuck. People trying to complete workflows inside SaaS tools hit moments of confusion.
Work grinds to a halt
When users can’t move forward, productivity drops and simple tasks take far longer than they should.
Support becomes the bottleneck
Instead of finding answers in-product, users default to opening tickets, flooding support teams with questions.
PROBLEM (CUSTOMERS)
Businesses pay for the tickets, confusion, and churn
When users get stuck, the impact doesn’t end with them.
$75B
Confusion is expensive
Users bounce between tabs, outdated docs, and unclear instructions, then give up and ask for help, creating long clarification threads that drain time.
Support teams down in repeats
Without clear in-product guidance, support teams answers the same questions over and over instead of solving real product issues.
Interruptions slow product velocity
Early-stage teams feel it most: every user question goes straight to engineers, breaking focus and shipping.
THE OPPORTUNITY
Why is support separate from the product?
We wondered what would happen if guidance didn’t live in docs, tickets, or threads but inside the UI itself. Users aren’t technical, so we can’t rely on exact specification on the issue.
However, we can rely on technological advances to alleviate the problem.
When users get stuck, they click Vector’s entry point and describe what they want to do.
SOLUTION
Vector is an AI co-pilot that completes tasks
Vector offers three layers of help
Tell me: a chat-like explanation (“Click here, then here.”)
Show me: It hovers over elements in your UI and guides you step-by-step.
Do it for me: it executes the workflow automatically, taking users from the beginning to the end of their workflow.
Halo Medical Solutions helps practices automate intake, documentation, and bill with AI.
REAL IMPACT
Our first partner: Halo Medical Solutions
The current system
Halo’s software required users to remember multi-step templates workflows and forms.
The team had to hop on call, answer questions, and guide doctors through all the processes.
What changed with Vector
Once embedded, Vector helped doctors complete workflows all without hand-holding or confusion.
Halo saw
drop in support tickets.
80%
Users stay focused on why they opened the software, not how to operate it.
VALUE PROP
Businesses stay focused on growing the product, not managing confusion
If users can’t figure out your product, theu churn. Vector makes sure they don’t.
WHY NOW
Users expect immediacy
When people can truly experience what you’ve built, that’s when it can change the world.
vector turns confusion into clarity.
Agentic AI’s recent advances
Only recently did browser automation and large-model reasoning become reliable enough to run real workflows.
SaaS landscape
SaaS teams are moving faster, hiring smaller support teams, and relying more on retention.
Customer success and onboarding is a growing market that represents over $300B across 17,000+ SaaS companies. We focused on early-stage teams who already spend on support and retention.
MARKET
Customer success and onboarding is a growing market
We generated
in contracted ARR
$7.8k
Vector is a system that has to feel trustworthy while taking action on behalf of a human.
PROCESS
Designing for behavior
Control handoff clarity
No rapid jumps, no unexplained clicks, and always pause before any pivotal actions take place.
A few key takeaways
RESULTS
Shipped a product within 2 weeks; won Audience Choice ($500)
Start broad, commit narrow
Exploring many directions early helped us land on the right one. We spent 6 weeks exploring an idea that we pivoted out of. It was called Folio.
Agentic design is behavior design
I wasn’t just designing UI. I was designing how the AI should act. I was crafting a system users could trust to act on their behalf.
Design around the moment of need
The most valuable features are the ones that appear exactly when users need them. Prioritizing context over clarity guided every design decision.
Be opinionated
Focusing on beachhead users meant making deliberate trade-offs. We designed for clarity over optionality, choosing workflows that mattered most.
Reach out if you want to see earlier explorations, alternate flows, or more of the thinking.
ADDITIONAL EXPLORATIONS
Plenty more lives in the design archive
2025 @ Maya Parthasarathy
X