Tristan Douville

Context

When I joined Pivott, there was no product function — no process, no team, and no defined way of taking work from idea to shipped. The engineering team was experienced and capable, but had come from a stable, profitable business where the primary mode was maintenance and incremental improvement. The mandate was to build a brand-new SaaS product at startup pace.

Problem

Launching a 0→1 product without a product function isn't just a process gap — it's a structural risk. Without defined ownership, prioritization, and delivery workflows, teams default to ad hoc decisions, unclear requirements, and reactive engineering.

Core tension: Move fast enough to build a competitive product while establishing the process rigor needed to sustain it.

There was also a cultural challenge: the engineering team's instincts were calibrated for stability, not speed, and process change without credibility rarely sticks.

My Role

• Assessed the full gap between where the team was operating and what a startup delivery model required
• Designed a complete SDLC spanning idea submission through to post-release review and success monitoring
• Leveraged my CSPO certification to bring in proven agile frameworks and external perspectives on team rituals
• Built the business case for dedicated product and QA hires
• Partnered directly with the CEO to secure the backing needed to drive cultural change with the engineering team

Key Decisions

1. Design the full lifecycle before hiring
Rather than hiring first and figuring out process later, I mapped the complete SDLC — 17 phases from idea submission to review and success monitoring — before growing the team. This gave new hires a functioning system from day one.

2. Secure CEO alignment before rolling out to engineering
Given the cultural gap, I recognized that process changes needed to come with executive weight behind them. Getting the CEO aligned first meant the shift wasn't just a PM's opinion — it was company direction.

3. Use certification frameworks as a starting point, not a constraint
Applied CSPO frameworks (planning poker, sprint rituals, backlog governance) as a baseline, then adapted them to Pivott's stage and team dynamics rather than implementing them by the book.

Execution

• Designed and documented a full SDLC covering idea submission, prioritization, requirements, design, hand-off, development, testing, release, and post-launch review
• Hired and onboarded a PM, a product designer, and a QA engineer — growing from a solo PM to a 3-person product team
• Established agile best practices and team rituals from the ground up
• Introduced Jira-based defect tracking to create a measurable quality baseline

Outcome

• Fully operational product function built in under 6 months
• Production bugs reduced by 25% following introduction of QA and formal acceptance testing
• Engineering team successfully transitioned from maintenance-mode to startup delivery cadence
• Operating model became the foundation for the 0→1 launch of Pivott's core contract management platform
• Established scalable team and process infrastructure that supported subsequent engineering growth from 5 to 8

Context

Leadership proposed a new feature to help customers manage renewal terms within contract documents — a known pain point for users managing long-term agreements.

Problem

Initial designs aimed to model every possible renewal scenario across all document types. While comprehensive, this approach introduced significant technical complexity, a high cognitive load for end users, and a timeline that would delay a widely-requested feature.

Core tension: Deliver a complete solution vs. deliver a usable solution quickly.

There was also internal misalignment — leadership favoured the more robust approach.

My Role

• Analyzed session recordings and support tickets to understand real user behaviour
• Used LLMs to review OCR'd contract text and identify patterns in renewal terms
• Found that the majority of users had simple, repeatable renewal patterns
• Facilitated discussions with leadership to refocus on core user needs
• Managed stakeholder expectations during a decision conflict

Key Decisions

1. Reduce scope to match real usage
Prioritized a recurrence-based model covering the most common renewal scenarios. Avoided building for edge cases upfront.

2. Leverage known patterns
Drew from existing SaaS conventions and competitor solutions to reduce the user learning curve.

3. Position as an MVP
Framed the solution as a test of user demand, keeping the door open for expansion. This reduced resistance by aligning on iteration over perfection.

Execution

• Partnered with engineering to define a lean, flexible recurrence system
• Kept scope tightly aligned to core use cases
• Instrumented the feature with Mixpanel to measure adoption and usage patterns
• Led internal enablement and external rollout through documentation and customer education

Outcome

• Feature launched successfully and remains in active use
• Covered ~95% of customer use cases, validating the simplified approach
• Avoided significant engineering investment in low-frequency edge cases
• The "comprehensive" solution was deprioritized indefinitely, freeing capacity for higher-impact roadmap items

Context

Pivott's marketing site was built on Webflow using a purchased template. Over time, the site's needs had grown well past what the platform handled well: the marketing team couldn't independently build new pages, custom interactions required developer involvement, and a planned redesign was quietly becoming harder to justify.

Meanwhile, the product had real content needs. New feature landing pages were stalled. A long-requested pricing page redesign — with collapsible comparison tables and an interactive plan structure — had been deferred repeatedly because the effort in Webflow made it not worth prioritizing.

Problem

The root issue wasn't design quality — it was a platform mismatch. Webflow's visual editor handled basic copy changes, but broke down for anything structural. Adding a new component, building an interactive pricing layout, or launching a feature landing page each required either a developer or a Webflow specialist. Pivott had neither available for marketing work at this time.

Core tension: The marketing site needed to grow with the product, but every meaningful change was a project in its own right.

There was also a compounding risk: Webflow was retiring its legacy editor, forcing a workflow transition regardless — making the status quo a temporary solution at best.

My Role

• Identified the constraint proactively before it became a blocker and built a concrete alternative before bringing it to leadership
• Built the business case comparing Webflow and a React/Vercel stack across cost, capability, and long-term flexibility
• Led the full technical implementation — repo setup, design system, component architecture, and deployment pipeline
• Established the AI-native design and build workflow using Magic Patterns, Claude Code, and Cursor
• Designed the project plan including content audit, infrastructure checklist, and DNS migration process
• Presented the prototype to the CEO and Head of Sales and secured approval to move the 2027 redesign to 2026

Key Decisions

1. Accelerate the 2027 redesign by over a year
The effort required to build a competent Webflow site was roughly equivalent to building in React — but with worse long-term outcomes. Framed this clearly for leadership and got the go-ahead to move immediately, collapsing a year-plus timeline into weeks.

2. Use Magic Patterns as the primary design tool
Rather than hiring a designer or Webflow consultant, used Magic Patterns to produce a fully on-brand React prototype — complete component library, brand token system, responsive behaviour, and accessibility compliance — before any production code was written. This gave leadership a concrete, clickable prototype to approve before committing engineering resources.

3. Choose Next.js over a simpler Vite + React Router setup
Evaluated both options and chose Next.js for its server-side rendering (direct SEO benefit for a marketing site), file-based routing, and native metadata handling. Vercel's first-class support made deployment, staging branches, and preview URLs essentially zero-config.

4. Design for non-developer contribution from day one
The explicit goal was a codebase that a Marketing Manager could actively contribute to — not just review. Chose a named Tailwind token system, a clear folder structure, and wrote a CLAUDE.md conventions file so Claude Code and Cursor could produce codebase-consistent output without manual correction. Copy updates, new pages, and image swaps no longer require a developer.

Execution

• Ran 60+ iterations in Magic Patterns to produce a complete, approved design system covering 10+ page templates and a full reusable component library — homepage, pricing, sign-up, feature template, About, Contact, Integrations, Terms, and a brand reference page
• Set up full production infrastructure: Next.js 15 repo, Vercel hosting with staging and production branch strategy, Vercel Blob for asset storage, and Formspree for form handling with spam filtering and Zapier workflows
• Wrote CLAUDE.md — a conventions document covering Tailwind tokens, component patterns, routing, image handling, and SEO conventions, enabling Claude Code and Cursor to contribute consistently without rework
• Generated 17+ pages using a combination of Magic Patterns exports, Claude Code, and Cursor with structured CSV-driven content inputs for batch feature page generation
• Established a repeatable design-to-production pipeline: brief in Magic Patterns → export TSX → integrate → Vercel auto-deploys a preview URL for review
• Produced a full project plan and task breakdown — including content audit, infrastructure setup, third-party tool migration checklist, and DNS cutover process — for cross-functional execution

Outcome

• Full marketing site redesigned and launched in under 4 weeks — compared to a planned 12+ month timeline under the original 2027 approach
• Non-developer contribution enabled for the first time: Marketing Manager can draft pages in Magic Patterns, push content updates via Claude Code, and review changes on Vercel preview URLs without engineering involvement
• Two high-priority product launches unblocked — previously stalled on the old platform, built and ready for content at launch
• All future pages follow a consistent, templated system — new feature pages are a content swap, not a new project
• ~40% reduction in monthly tooling costs compared to Webflow site plan and workspace seat fees

Context

The product team was increasingly pulled into customer support — fielding interruptions via Slack, email, and calls to help resolve tickets. Comprehensive documentation, PRDs, and BI tools existed but were underutilized by the Customer Success (CS) team.

Problem

Support staff relied on the product team for answers instead of self-serving through available resources, causing:

• Constant interruptions to product work and roadmap execution
• Slower response times due to dependency on PM availability
• Underutilization of existing documentation

Core tension: How do you enable support to independently access product knowledge without making the product team a bottleneck?

My Role

• Owned internal documentation strategy as part of the product team
• Interviewed CS team members to understand their real-time support workflows
• Identified that the issue wasn't lack of documentation — it was lack of accessibility and synthesis
• Explored solutions ranging from process improvements to tooling

Key Decisions

1. Validate non-technical solutions first
Improved internal documentation and ran a lunch & learn on effective self-serve questioning techniques (adapting PM-style user interviewing for support). Helpful, but did not reduce reliance on the product team.

2. Invest in a scalable system
Identified an opportunity to use AI to aggregate and synthesize internal knowledge. Proposed an internal "Product Concierge" to enable real-time, self-serve answers.

3. Prioritize trust and accuracy over breadth
Designed the system with strict guardrails: surfacing source references, indicating the age of information, and limiting responses when confidence was low.

Execution

• Defined requirements and architecture for a Slack-based AI assistant connected to internal documentation, PRDs, code repositories, and select BI queries
• Collaborated with developers to validate the approach and ensure feasibility
• Built and iterated on the solution, refining prompts and data connections
• Implemented tracking to monitor usage and response quality

Outcome

• "Product Concierge" is now actively used, handling hundreds of internal queries per month
• Reduced direct interruptions to the product team
• Improved speed and consistency of support responses
• Created a scalable knowledge system, reducing dependency on individual team members
• Established a foundation for future internal AI tooling

Context

A high-value client relied on email and spreadsheets for a critical financial workflow and experienced delays in their year-end reconciliation due to missing records. They came to our product and leadership team with a request to build a proper solution.

Problem

The client required a highly customized workflow our product didn't support. Key challenges included:

• A time-sensitive deadline tied to their fiscal year-end
• Conflicting requirements across multiple client departments
• One-off solution risk — potential roadmap diversion without clear reuse potential

Core tension: Solve a critical client problem quickly without introducing long-term product bloat.

My Role

• Led user interviews with client stakeholders to uncover root workflow issues
• Synthesized findings into clear product opportunities for leadership
• Owned the PRD and solution design, including a functional prototype
• Liaised with Customer Success and Engineering to ensure alignment
• Acted as the bridge between client needs, business goals, and technical feasibility

Key Decisions

1. Build vs. defer

• Option A: Defer to protect the roadmap
• Option B: Build a one-off solution
• Option C (chosen): Build an MVP with a scalable architecture

2. MVP scope definition
Prioritized only the core workflow blockers for year-end readiness. Deferred edge cases to post-launch iterations.

3. Platform investment
Designed the solution to support configurable workflows (fields, statuses, naming), with clear potential for reuse across other clients.

Execution

• Aligned internal stakeholders on scope vs. timeline tradeoffs
• Collaborated with engineering to define a flexible technical design under time constraints
• Presented a client-facing proposal including problem framing, a prototype walkthrough, a delivery plan, and a pricing and support model

Outcome

• Client accepted the proposal, increasing spend by 20% YoY
• Delivered on time for fiscal year-end
• Established a foundation for configurable workflows, enabling reuse across other clients
• Project success generated an internal referral, expanding into additional opportunities within the client's organization