One Pager – Team 7

Team Members: Diego Sanchez Tirado, Felicia Yan, Isaac Were, Itai Reingold-Nutman, Janus Tsen, Victor Ahona, Carolyn Lee

Description: What is it?

Index is a self-growing documentation system that can take action to find context that it doesn’t have to make cross-team collaboration and alignment seamless. When PMs in large multi-team enterprise companies want to plan new features, but struggle with conflicting or duplicated work across teams, Index proactively searches for the missing context behind other teams’ plans by searching through files or reaching out to team members to fill in details, maintaining up-to-date and reliable company-wide knowledge of ongoing work. 

Problem: What problem is this solving?

PMs can’t see what other teams are working on, so conflicts downstream break their work. 

Why: How do we know this is a real problem and worth solving?

We talked to more than a dozen PMs, and half (7/14) reported this as a problem they face in their work. It’s a problem worth solving because PMs today already use workarounds like organizing workstreams into focused Slack channels (“You might have 20 to 30 active Slack threads going on”) or searching existing company documentation like Notion and Confluence, but these methods are limited by bandwidth to monitor multiple information streams and unreliability of outdated information. 

Success: How do we know if we’ve solved this problem?

We conduct a review of the product usage as well as the feedback from team members on the significance that our product has had on the team. The following metrics can be used to quantify success: number of people using the product, the number of pings recorded, the growth rate of the knowledge base. More users would mean our product is getting a good amount of traction; a higher number of pings would imply that the rate of usage of our product is high; and the larger a company’s volume of information is contained in the knowledge base makes it more efficient as a resource to team members.  In the long run, the number of pings reduces while the volume of knowledge increases, and success is measured by the number of queries issued to the product. 

Audience: Who are we building for?

We are building for PMs who want insight and information across other teams’ projects to prevent conflicts in their own work, especially targeting PMs at large companies (500+ employees) where multiple teams ship on the same product. Our business model will focus on selling to enterprises using a seat-based strategy. 

What: Roughly, what does this look like in the product?

Our product aims to solve the problem by creating a shared AI-based knowledge base that is more efficient and effective than current solutions. Users have an AI chat that they can ask questions about other teams and orgs. If the AI can find the answer in the files, code, and memory of the knowledge base, it quickly responds. Otherwise, it pings the relevant people across the company and asks them questions that are important to answer in the initial query. Once they answer, it responds to the initial user with a well-supported and up-to-date argument, and generates more files to store that information in memory so that future queries can be answered with it as well. In this way, it is an ever-growing knowledge base that stays up to date on the company’s practices, initiatives, teams, projects, and processes. 

How: What is the experiment plan?

We’ll run our three test cards (responding to pings, visibility of work, and workstream discovery) with 2-3 PMs (if finding PMs is hard, we’ll just work with our friends) and teammates per round, using the v0 prototype. Each test has a metric and a pass threshold set in advance. If a test passes, we keep that feature. If it fails, we revise or cut it before the next round. We’ll also have a different LLM than the one we built to review the prototype for usability issues, as a quick check alongside real users, not a replacement. We plan two prototype rounds, with round 2 changes based on round 1 results, before shipping the MVP.

When: When does it ship and what are the milestones?

  • Week 8: run experiments to validate assumptions
  • Week 9: run pilot sessions with beta testers

One line from the reading: We drew from Lenny’s emphasis on clearly defining success criteria in a one-pager, which we applied to our success section by identifying measurable indicators of Index’s adoption, usage, and knowledge base growth.

Avatar

About the author

Leave a Reply