UX STRATEGY · INTERACTION DESIGN · LIVE CLIENT · WEB PLATFORM
Every ecosystem platform fails the same way. It becomes a directory.
TG10x is a state-backed effort to build the digital backbone of Telangana's startup ecosystem ‚connecting founders, investors, government enablers, mentors, and service partners in one place. With only a timeline of 1 month, I narrowed the scope to the two that unlocked everything else ‚and designed a system that helps founders and investors evaluate each other, not just find each other.
Role
Interaction Designer
Timeline
5 Weeks
Project Type
Live Client
Team
2 Designers
THE CLIENT
TG10x is part of Telangana's T-Hub initiative, a government-backed effort to build the digital backbone of the state's startup ecosystem. Founders, investors, mentors, corporates, government enablers, and service partners, all in one platform.
THE PROBLEM
Founders and investors weren't struggling to find each other. They were struggling to evaluate each other. The ecosystem had no standardised way to compare startups, assess investor fit, or move from "found someone interesting" to "confident enough to reach out."
HOW WE KNOW IT'S A PROBLEM?
I mapped how startup-investor discovery works offline: warm referrals, accelerator demo days, manual spreadsheets, social signals. Every channel had the same breakdown - fragmented signals, context lost between stages, repeated qualification questions. The problem wasn't access. It was structure.
THE GOAL
Design a platform that helps users make decisions, not just browse listings. Structure the right signals early, make intent visible, compress the time from discovery to action.
THE SCOPE
Two designers. One month. No user research budget. Eight personas in the brief, narrowed to two (founders and investors) through a prioritisation framework. A client expecting breadth. A problem that demanded depth.
RESEARCH
Ecosystem platforms don't fail because they lack features. They fail because they don't help users make decisions.
I spent the first week mapping how startup-investor discovery actually works offline: warm referrals, accelerator demo days, manual spreadsheets, social signals. Every channel had the same breakdown - fragmented signals, no standardised way to compare startups, context lost between stages, repeated qualification questions.
The insight: founders and investors aren't struggling to find each other. They're struggling to evaluate each other efficiently. The platform's job wasn't to replace existing discovery, it was to compress it. Structure the right signals early. Make intent visible. Reduce the time from "found someone interesting" to "confident enough to reach out."
The reframe: ecosystem platforms replicate directories. TG10x needed to be a decision engine.
What investors use to judge a startup:
What founders use to judge an investor:
-
Sector and stage fit
-
Traction metrics (MAU, revenue, burn, runway)
-
Founder and team background
-
Problem-solution clarity
-
Funding ask and use of funds
-
Cohort or incubator affiliation
-
Legal verification (CIN/MCA)
-
Sector and stage alignment
-
Cheque size range
-
Portfolio and reputation
-
Value-add beyond capital
-
Identity verification
-
Follow-on policy
STRATEGY
Eight personas in the brief. I argued for two.
I built a prioritisation matrix across four dimensions: frequency, impact, effort, and MVP necessity. The output was unambiguous. Discovery & Matchmaking and Progressive Onboarding were the only two experiences that every other feature depended on.
Community tools, event listings, mentor matching, service partner directories — all worthless if users couldn't find the right person and trust that they were who they said they were.
I presented the matrix to the client with the scoring visible. They agreed to proceed with the focused MVP. What I didn't anticipate: "agreed to proceed" and "genuinely bought in" aren't the same thing. That gap would show up at delivery.
What I gave up: Job seeker flows, mentor matching, and service partner directories — scoped out and documented as Phase 2.
FRAMEWORK
Designing for judgment, not just browsing.
Most platforms show you information and hope you act on it. I wanted TG10x to structure information in the sequence decisions actually happen, not in the order that's easiest to build or most impressive at a glance.
Every profile, every card, every field was organised into three tiers:
PRIMARY
Instant Scan
Identity, sector, stage, one-line pitch, traction highlight, founder name.
Answers: "Is this even relevant to me?"
SECONDARY
fit assessment
Portfolio examples, geography, cohort affiliations, value-add signals.
Answers: "Do I want to look closer?"
TERTIARY
Decision triggers
Pitch deck, engagement preferences, compliance verification, contact CTA.
Answers: "Am I confident to reach out?"
The tiering wasn't a layout exercise. It was the platform's logic layer, the same structure that informed card design, profile pages, onboarding fields, and filter hierarchies.
DESIGN
Seven artifacts. One principle: show value before asking for effort.
The final deliverable was a Hi-Fi prototype covering Homepage, Get Started (signup), Role Selection, Startup Onboarding (3 steps), Investor Onboarding (3 steps), Explore / Discovery feed, and Startup and Investor Profile pages.
Every screen was delivered with documented rationale: user goals, business goals addressed, screen justification, and known limitations.
01 ENTRY POINT
Join something with momentum, not starting from zero.
Ecosystem proof through real numbers, role-based segmentation to start a personalised journey, and sector discovery to reduce decision fatigue for exploratory visitors. No forced signups. The platform sells itself before the first form field.
02 PROGRESSIVE ONBOARDING
The form isn't extracting data. It's building something the user can see growing.
Step 01: Identity
The founder sees their profile taking shape before they've finished filling in the first form. Name, company, problem statement, website — each field populates the live preview on the left. Profile Strength starts at 14% and climbs with every input. The form isn't extracting data. It's building something the founder can see growing. The copy reinforces it: "Preview how you'll appear to investors and mentors."
Step 02: Discovery
Stage, sector, location, traction indicators, and intent — "What Are You Looking For?" with tags like Raise funds, Find mentors, Join programs, Hire talent. Each selection sharpens how the founder appears in the discovery feed. The copy reframes the task: "Help Us Curate What Matters Most" — not "fill out this form." Profile Strength hits 33%. The founder is a third of the way to being findable.
Step 03: Credibility
The final step is framed as enhancement, not requirement at this moment. Pitch deck upload, team members with LinkedIn links, fundraising amount — each one makes the founder more investable in the eyes of anyone browsing the Explore feed. The "Boost Your Profile" nudge makes the abstract cost of skipping feel concrete.
Investor onboarding mirrors this structure with the same dual-pane logic, but optimised for a different evaluation context.
-
Step 1 captures identity and investment activity stance (Actively Exploring Deals vs. Observing the Market).
-
Step 2 captures investment preferences (sector focus, stage focus, cheque size, geography) — with each selection visibly increasing the match count in a live discovery feed on the left.
-
Step 3 captures credibility signals (portfolio examples, preferred founder traits, SEBI/DPIIT compliance, value-add areas). The same principle holds: show the user what their input produces before asking for more of it.
-
03 THE DECISION ENGINE
Every card is a judgment call. The layout had to serve that.
The Explore page serves eight different personas from one surface. An investor scanning for deal flow, a mentor looking for startups to advise, a corporate innovation lead evaluating partnership candidates, all landing on the same page, all needing different entry points into the same data.
The startup card itself had to answer one question in under three seconds: "Is this worth a closer look?" Name, location, one-line pitch, stage, sector, founding year — that's the identity layer. One human-readable traction badge that's the proof layer. Two CTAs: "View Profile" for evaluation, "Connect" for action.
The investor card mirrors this structure with different priority fields — firm name, cheque size, stage focus, portfolio highlights, and a track record badge instead of traction.
This thinking came directly from my game design background. In game systems, you design for the whole game state, every possible condition a player might be in, not just the happy path. Here, I was designing for the ecosystem state: what happens when 80% of startups have incomplete profiles? The card had to account for that, not ignore it.
04 PROFILES
High signal in minimal scroll.
Startup Profile
Structured around investor decision sequence: identity and traction proof at the top, problem-solution fit in the middle, team credibility and pitch materials below, and current needs as a persistent sidebar. An investor can determine fit within 15 seconds or go deep without switching screens.
Investor Profile
Mirrored structure, different priorities. Founders care about alignment (sector, stage, cheque size), credibility (portfolio highlights, exits), and engagement style (warm intros only, available for pitch calls). "Preferred Founder Traits" — Technical Founder, First-time Founder, Domain Expert — adds signal that goes beyond capital.
OUTCOMES
The client liked the work. They still wanted more of it.
The client received the deliverable on 7th June 2025. Their response was mixed, they appreciated the craft and the depth of the rationale, but felt we hadn't covered enough ground. They still wanted the other six personas addressed in the same timeframe.
The gap wasn't between the design and the brief. It was between the rigour of the process and my ability to make that rigour visible to the client in the moment that mattered. The work was thorough. The argument for it wasn't.
What held strong:
The information tiering gave my teammate an unambiguous structure — no screen had ambiguity about what went where or why. The progressive disclosure model kept every decision traceable to either a trust principle or a conversion goal. The dual-pane onboarding design solved a real problem: making the value of profile completion visible before the profile was done.
REFLECTION
What I'd do differently?
This project taught me something I hadn't fully understood before: being right about the problem is not the same as winning the room.
The prioritisation logic was defensible. The decision-centric reframe was correct. The information tiering held under scrutiny. But a state-level client under time pressure doesn't want to hear "we narrowed the scope." They want to see the scope they asked for — or they want to be taken on a journey that makes narrowing feel like the obvious and exciting choice.
I built that journey for users. I didn't build it for the client. If I could go back, I'd invest time early in a one-page visual that made the prioritisation argument before anyone opened a Figma file. Something that showed: here are the 8 personas, here's the dependency map, here's exactly why these two unlock everything else. Make the logic feel inevitable, not like a constraint we imposed.
The second thing I'd change: start directing the visual design earlier, with rougher outputs and faster loops. I gave my teammate a clear content logic, but too late for real back-and-forth on the interaction details.
The strategic thinking on this project is work I'm proud of. The gap between that thinking and what the client saw in the room — that's what I'll do differently next time.