2024
When I interviewed with Pulley in January 2024, Yin Wu (CEO) and Grant Oladipo (COO & Head of Product) emphasized their mission to help founders start and scale their companies. They had made signficant progress on the mission to help founders start their companies, offering a compelling alternative to Carta for cap table management, and also being the leader in token management for crypto-native companies. Now, they were ready to figure out how to support customers as they began to scale.
The opportunity to help Pulley identify the next market opportunity for growth toward decacorn status was an amazing opportunity. In addition to my functional leadeship on Design (hiring, operations, design systems, and IC work on the core product), I advised and aligned leadership on an approach for customer discovery and product discovery.
After months of research and exploration, we have started to converge on a direction that allows executives to accelerate strategic alignment, consolidate data from multiple sources, and increase trust and credibility, all while leveraging an AI-native product experience. This central UX for strategic collaboration is what we came to refer to as a "canvas".
This particular exploration came after about a month into product discovery. We spent the previous weeks exploring a component and system architecture, with a focus on API integrations for strategic business context, agentic collaboration, and iterating on strategic artifacts. Grant acknowledged that while many executive collaborators would continue to create strategic artifacts in their preferred systems, what we would ultimately sell to customers was an opinionated framework that would consume those artifacts.
"Likely, the functional leaders are gonna create plans, and then we're gonna import them into the structure. But I'm missing the structure of the doc. And if we had that, then I could create new hiring plan on pulley. I could create new annual plan on pulley. I create new equity plan on pulley. And to me, that's the thing we're gonna sell."
The resulting exploration is an extensible framework that allows executives to scaffold collaboration for any strategic job to be done, from annual planning to department-level headcount planning, to financial projections.
Prior to embarking on customer discovery or product discovery, I facilitated many brainstorming sessions with leadership focused on articulating our North Star mission and hypothesizing an ICP. A theme emerged, as observed by Grant: "The CFO is the only person that’s aligned 100% with the founder. Everyone else on the exec team may have their own agenda, but the CFO is responsible for the business health of the whole company."
Through this lens, the team moved to customer discovery, interviewing 20 finance leaders and surveying 679 executives. We identified the following opportunities when it comes to how finance leads collaborate with executive teams:
Aside from these top-level opportunities, we also identified some task-level pain points in the finance leader's planning journey that would inform how we approached product discovery.
Through our interviews and surveys, we repeated heard concerns about: fragmented insights between departments, difficulty driving alignment on realistic revenue goals, and siloed decision-making processes.
We also heard concerns about: misalignment on hiring priorities, difficulty justifying headcount growth, and inconsistent understanding of resource capacity
And probably most contentious of all: competing priorities across the executive team, lack of transparency with financial and operational data, and difficulty aligning ROI expectations.
The following exploration is one of many iterations where I evaluated how a permissioned collaborative space that leverages agentic collaboration might address these pain points.
Executives can sync buiness context to the space via API integrations. They might connect real-time financial data from Quickbooks and Stripe, human capital data from Rippling, and strategic context from Google Drive and Notion, for example.
They could also import strategic documents and as a jumping off point for iterating on their strategic plans. Each artifact could be re-genereated by selected any element within the document and asking the agent to re-generate the information based on the underlying data.
Another key aspect of the experience if proactive requests for alignment driven by the agent. Instead of needing to read 7 different pre-reads ahead of a weekly executive leadership sync and manually identifying which topics required alignment during the call, the agent could triage the documents and call attention to topics for alignement ahead of time. This would free up executives to use their synchronous collaboration more efficiently.
During product discovery, lots of new agentic coding products were coming to market. We wondered — how quickly might we build a solution we could use internally and perhaps give to customers for evaluation?
One such tool was Replit Agent, launched in September 2024. Even though the tool was still in beta, I was able to build a fully functioning and deployable app in less than 20 hours. The app includes user auth, permissioned user collaboration, document upload and markdown parsing, file support for PDF, XLS, and DOCX, and an integration with GPT that includes document context with the prompt. The prompt feature can be initiated either from the chat or by highlighting a string on the canvas.
It became readily apparent that agentic prototyping is the way forward, as it allows "non-technical" members of the team to create truly interactive experiences in a dramatically short period of time ◼︎