Multi-player learning for the human skills early careers depend on
Problem
CogSolo is an AI-powered practice platform where early-career professionals build human skills like pitching, speaking up, and navigating hard conversations at work. It was single-player (watch a lesson, record a response, get AI feedback), and the objective was to evolve into a multiplayer product that drives more user value and eases the self-consciousness of practicing alone.
Outcome
I defined, designed, and launched key features including Situations, My Crew, and Crew Notes: fixed learning roles played by seven AI characters with full backstories, personalities, starting all-AI and opening to invited humans, with Crew Notes tracking how you grow.
Impact
It's now CogSolo's core product direction
Insights and recommendations across Life, Health, and Motor Insurance
"I don't want a middleman taking a cut"
"I need someone accountable if claims go wrong"
"I can research faster myself"
"I don't know enough to choose the right policy"
"A trusted friend bought online and it worked"
"A trusted friend recommended this specific agent"
Tier 1, digitally fluent, wants control
Tier 2 & 3, trusts expertise, wants minimal involvement
Both paths start with trust in a person, a friend, a family member, and not trust in a platform.
Customers choose accountability, not channels
Customers don't trust policies—they trust people. Friends, family, and agents drive decisions more than websites or aggregators. Even digitally-savvy customers who research online seek human validation before committing.
Recommendation: The 170,000+ village footprint is MIBL's trust infrastructure. Digital strategy has to work with it, not around it.
Visibility beats outcome
Even if customers aren't able to receive the full claim amount at the end, customers still cite an overall positive experience where they were kept informed.
Recommendation: Investing in reduced ambiguity, increased transparency, and clarity for customers can have a higher impact on attrition rates and brand image as compared to process optimization.
Claims shape brand perception more than purchase
Purchase journeys are relatively smooth across channels. Claims is where experiences diverge, and where lasting impressions form. Recurring pain points include: no visibility into processing status, repeated documentation requests, unexplained delays, and hidden fees.
Recommendation: A digitally-enabled claims tracking experience could be a significant differentiator, especially for their semi-urban/rural base, where physical office visits are burdensome.
Aggregators help discovery but hurt conversion
PolicyBazaar, a direct but much larger competitor to MIBL's PayBima, works for initial comparison. But aggressive callbacks, pushy tactics, and price discrepancies erode trust at decision time.
Recommendation: There's space for an alternative that offers aggregator-level transparency with agent-level accountability.
Two key stakeholders that shape CX
Hospital administrators
Network quality is about the administrative competence of partner hospitals and not just geographic coverage. In cashless claims, hospital admin teams initiate the process, gather documents, and coordinate with the insurer. A smooth hospital experience created peace of mind, even when payouts were reduced. A poor one created stress regardless of policy quality.
Recommendation: MIBL could differentiate by curating hospital partnerships based on claims-handling quality in addition to geographic coverage.
Agents and PoSPs
Agents are the single point of contact for trust, handholding, and claims support. Supported agents deliver better service. Unsupported ones create friction or charge hidden fees.
Recommendation: Investing in agent tools and training has downstream effects on customer satisfaction, especially for MIBL's village network.
We then identified key customer Jobs-To-Be-Done across verticals and mapped them to different phases of the journey, including learning, evaluation, purchase, renewal, and claims.
Approach
We had the project lead on MIBL's side join a bulk of the 70+ interviews across channels, creating an internal advocate who heard customer frustrations firsthand rather than filtered through a report. We debriefed after every few sessions, pursuing emerging threads in real-time rather than waiting for synthesis at the end. We started each interview with word associations to surface participants' emotional landscape quickly, then followed their unique story, pushed for "show me" moments, and covered the necessary details to map the journey. We shared our findings with senior executive leadership and facilitated a workshop to help them develop a roadmap based on our recommendations.
Learning
I mapped journeys for three verticals, three phases with five sub-phases, and four channels—while coordinating with three other researchers, ops, clients, and ramping fast on insurance fundamentals, regulatory context, and MIBL's product ecosystem.
Midway through, we discovered some participants were fabricating their insurance purchase and experiences. Once identified, I worked with our ops team to strengthen screening protocols, including cross-checking participant details myself against India's national vehicle registry database to verify actual policy ownership. We also renegotiated the project timeline with our client, given slower recruitment, prioritizing research quality over speed. Getting into the details of recruitment gave me a new appreciation for how difficult it is to source the right participants for a study this large and specific. I understood that research quality starts with ops rigor, well before interviews and synthesis.
All in all, this project sharpened my ability to toggle between detail and big picture, especially the discipline to capture detailed journeys across 40+ interviews, and the judgment to elevate only what answers "so what does this mean for the business."
