This page exists so you can evaluate before we talk. Some of the work is from Salesforce, where I led product strategy at scale for a decade. One case study is from a current client. All of it addresses decisions B2B software companies actually face.
Salesforce Partner Portal was competing against pureplay vendors who did nothing but channel sales. They knew the language, had their own implementation teams, and undercut Salesforce on price. Sales Cloud was the darling. Partner Portal was a legacy product in maintenance mode. The instinctive response — match their features — would have taken years and ceded the framing to competitors who'd always know the domain better. I joined Experience Cloud to change that.
What I didWe rejected feature parity and bet on two things: speed-to-value and ecosystem leverage. Speed-to-value meant rebuilding the underlying CMS around a component and template architecture — a multi-year infrastructure investment with no precedent internally, pushed by a small group of believers against real skepticism. The payoff: pre-built mobile-responsive templates, 100+ drag-and-drop CRM and CMS components, and a deployment model that got customers live within weeks. We flipped the perception overnight. Ecosystem leverage meant not trying to own everything. We partnered with select ISVs to close non-strategic gaps — LMS integration (Appinium), Loyalty Management (Fielo) — and structured pricing so both partners and customers had a fast path to adoption. I worked in parallel with marketing and pre-sales on positioning and field enablement, staying close to post-sales feedback to close the loop.
PRM grew ARR 30%+ year-over-year. The competitive frame shifted from feature comparison to platform surface — pureplay vendors couldn't replicate what Salesforce offered across CPQ, Analytics, and Marketing Cloud. The template and component approach kept us relevant in SMB and mid-market. Uncontested leader in the Forrester PRM Wave (2019).
Salesforce — Product Strategy, PRM / Experience Cloud
In 2023, leading the Admin and Security product team at Salesforce, we had a PM who had been a Salesforce admin — she understood the problem from the inside. We ran customer interviews and cross-referenced with Salesforce's own support data: admin cases were a leading case driver across the customer base. Our estimate: the average admin spent 20% of their time, one full day a week, troubleshooting user access and record visibility issues. Salesforce's complexity had grown substantially as new clouds were added: Analytics, Commerce, Marketing. We joked that becoming a Salesforce admin was like completing a PhD. The admin job requires deep accumulated knowledge.
What we didWe built a proposal for an AI agent scoped to a specific, well-evidenced problem: diagnosing and resolving user access and sharing configuration issues. The scope was deliberately tight — not a general-purpose admin AI, but a focused agent for the highest-frequency, highest-cost task. The business case had two angles: direct support cost reduction and an indirect commercial benefit, since admins are routinely consulted during pre-sales evaluations. The proposal was approved but we did not get extra funding. 2023 was the year every Salesforce resource moved toward Data Cloud and Einstein Copilot. We built a pilot on effectively no budget. A v1 with modest capabilities followed in year two. The project is still running, slowly expanding capabilities release after release.
The capability is now part of Salesforce's Agent for Setup, under the Agentforce umbrella. The vision was correct. The hardest part wasn't the AI. It was the platform underneath it. Legacy enterprise systems weren't built to be reasoned over — the API coverage, the metadata structure, the data model all need work before the AI can do anything useful. Being early with new technology means building the infrastructure and the capability at the same time. Getting that right takes time, evidence, and a team willing to start before the conditions are perfect.
Salesforce — Product Strategy, Admin & Platform / Agentforce
We were launching Experience Cloud as the successor to Salesforce's legacy portal products. The old per-seat model was losing deals. For customers exposing a portal to millions of users who log in once or twice a year, the cost became prohibitive before it became useful. The new product needed a different model. But thousands of existing customers sat on legacy SKUs with different structures, and changing the pricing model without damaging trust with the installed base was its own problem.
What we didWe designed a login-based consumption model alongside the per-seat option, giving customers the choice based on their usage pattern and the flexibility to switch between both to optimize costs. The main internal concern was cannibalization of existing seat-based ARR. The answer was in the data: the deals we were losing were measurable and growing. For migration, we grandfathered access to the new offering, published a migration guide, and documented explicitly which legacy capabilities didn't carry forward. Customers could evaluate the trade before committing, not discover gaps after.
Customers blocked by per-seat economics could now enter on a model that matched their usage. The migration from legacy portal products proceeded with a clean, documented path and no forced changes. The trust work up front — being explicit about what didn't carry forward — turned out to matter more than the pitch about what did. This new consumption-based model unlocked significant usage growth. Today, Experience Cloud supports over one billion users worldwide.
Salesforce — Pricing & Packaging, Experience Cloud
CloudCraze, a Salesforce ISV, had built a B2B commerce product on the Salesforce stack. They were analyst-recognized, with an existing customer base and a sales team already in market. Salesforce had no B2B commerce presence. The question wasn't whether the market was real. It was whether to build, partner, or acquire.
What we didI led the product assessment alongside internal architects: full capability analysis, custom data model, licensing, performance and scale validation. We mapped what would carry over into our existing Experience Cloud offering and what we'd have to build. One thing we knew going in: ISV apps don't survive full integration at Salesforce scale. You acquire the team, the customers, and the market position, then rebuild the product. That was a multi-year investment we priced into the recommendation and made anyway. I was part of the post-acquisition team that executed the rebuild and the go-to-market for the combined product.
The acquisition became Salesforce's B2B Commerce offering. The rebuild was the right call. The ISV app was never the asset. The team, the customers, and the market position were.
Salesforce — Strategic Bets, B2B Commerce / Acquisitions
As I led Platform and Admin product strategy at Salesforce, we proposed transforming how customers consume Salesforce Platform — turning dozens of individual API and service SKUs into a consumption-based PaaS model, billed through Salesforce Digital Wallet. The timing was deliberate: Salesforce was already building usage-based infrastructure for Data Cloud and Marketing Cloud. The bet had multiple compounding levers: removing friction for customers forced to buy user seats upfront, unlocking consumption pricing for ISVs structurally blocked by the existing AppExchange licensing model, and positioning Salesforce as the platform of choice for custom app builders ahead of Microsoft and ServiceNow.
What we didWe built the full proposal and pitched it directly to the Platform GM, competing against a large queue of other initiatives. Three connected problems were laid out clearly: customer friction from seat-based platform pricing, the ISV monetization constraint, and the platform weakness in the custom app segment. Platform revenue cannibalization concerns were raised. We had real data from Experience Cloud's pricing transition to challenge them — not enough to close the argument in a room with limited appetite for large bets. A new GM, competing priorities, and the weight of inertia tipped the decision toward the status quo.
The proposal went against twenty years of Salesforce platform pricing. Too big to swallow, too resource-intensive to fund, slightly ahead of where the organization was willing to go. Today, Salesforce monetizes Agentforce on the same consumption-based model. The shortcomings we identified remain. Customers and ISVs can build standalone agents on Data Cloud and pay for usage — but the moment they consume Platform services within an agent use case, workflow runs, API calls, data storage, they fall back into a hybrid model. The infrastructure that would have made agents cleanly monetizable for partners was never built. The gap between Salesforce's agent ambitions and the platform economics underneath them is still visible today.
Salesforce — Platform Strategy, Salesforce Platform
A publicly documented security vulnerability — CRM data exposure through misconfigured guest users on Salesforce public sites — required a complete redesign of the security model for all public Experience Cloud portals. The scope was the entire installed base: every customer running a public site, and every ISV whose app used the guest user model. The fix was not a patch. It was a new security architecture that, when enforced, would break existing customer configurations. There was no version of this that didn't require customers to change something.
What we didWe ran queries across customer orgs to map exposure across multiple attack vectors before defining the fix. The rollout was designed in three stages tied to Salesforce's release schedule: opt-in, then automatic enrollment with opt-out, then full enforcement with no exit. Eighteen months separated the first announcement from final enforcement. At every stage, customers had a self-service diagnostic tool in the admin space, detailed release notes, direct email campaigns, recorded webinars, and success manager coverage with prepared templates. When our queries identified a strategic account at risk, we routed proactively to their success manager. We built a silent kill switch for the final enforcement phase. Never announced. Used once.
100K+ customer orgs migrated to the new security model. No red accounts. The handful of customers who hit edge cases during automatic enrollment had support protocols ready and were resolved within hours. A mandatory, breaking change across Salesforce's global customer base, executed without a single major incident. The result came from one decision made early: give customers enough time and enough tools to fix the problem themselves, and they will.
Salesforce — Platform Security, Experience Cloud / Salesforce Platform
A large Salesforce platinum SI — several thousand employees, global presence, deep expertise across Sales, Service, and Commerce clouds — had spent two-plus years building an AI coding assistant for Salesforce developers. Genuinely impressive: deep Apex code generation, test class automation, native Salesforce metadata understanding, Knowledge Base integration, built from the ground up with AI scientists. The team had used it internally on customer engagements. The question they brought to the Lab: can we turn this into a standalone product — sell it to Salesforce customers, to other SIs, and potentially position it as an acquisition target?
What I didI reviewed the full technical documentation alongside the pitch deck — not just the demo layer. I ran multiple product walkthroughs and interviewed the technical team to get below the surface on how the system reasoned over Salesforce metadata. I pulled in contacts from my network to pressure-test the demo independently. My assessment drew on direct Salesforce product knowledge: I knew where Salesforce's own coding tools stood at the time, what their trajectory looked like, and what the competitive dynamics in developer tooling actually were. The capability was real. The competitive position was not. The product ran on OpenAI's models — every new foundation model release reset the capability gap independent of what the SI had built on top. The surface area was too broad, the ICP undefined, and the go-to-market had no path to scale: one-on-one customer relationships don't become a product business without structural investment they hadn't planned for. I also identified a pivot worth exploring — the Salesforce admin persona — an underserved audience of 200,000+ customer orgs, an untapped AI opportunity, and a fundamentally different competitive dynamic than developer tooling.
The client decided not to build a standalone product business. What they did instead was the right call: they kept the AI assistant as an internal productivity tool and as a pre-sales differentiator, evidence of technical depth they could demonstrate to prospective clients. The Lab's value wasn't validating the original thesis. It was delivering a clear-eyed assessment before a company spin-off and heavy investments for a go-to-market path with structural problems. The exec came in with high conviction. The conversation that followed was direct. That's what the engagement is for.
The Fabs Way — Fabs Lab, 2024
PRM wasn't the priority product inside Salesforce. Specialist vendors owned the category. We were under-resourced and late. We modernized the platform, extended through the Salesforce ecosystem instead of over-building, brought in strategic ISV partners to fill gaps, and rationalized a SKU structure that was killing deals. We also killed bets that weren't working — including some we made. In three years: uncontested leader in the Forrester Wave. $350M+ ARR.
I owned pricing and packaging for Salesforce's external user licensing — the model every cloud and industry vertical built on. We introduced a usage-based model alongside the existing seat option — fully executed: contracts, provisioning, monitoring, edge cases, sales playbooks. My boldest bet: price Salesforce Platform services like AWS — consumption-based, charged through Salesforce Digital Wallet. Not adopted. Agentforce runs on that model today.
I've started companies, worked across six countries, and built in markets where there was no safety net. None of those businesses became venture-scale. These experiences taught me something the Salesforce years couldn't: what it means to have skin in the game. I care about the work. It's why I'm selective about who I work with. Founders who hire me don't get strategy advice filtered through the comfort of a large org. They get someone who has been where they are and knows which tradeoffs are real and which ones are noise.
Twenty-plus product strategy projects across my career. Many didn't make it: the proposal didn't win, the budget didn't come, the thesis didn't survive scrutiny. Early on, that stung. Over time, I understood it differently: the failures were the reps. You don't develop your craft by winning. You develop it by being wrong enough times to know where the traps are. What changed: I learned to go deeper before committing to a direction, think three steps ahead, and break my own plan before the client does. The engagements that didn't ship taught me more than the ones that did. You get the version of me that was built on both.
I ran The French Cookie Guys, my last venture, entirely on AI (ChatGPT, Claude, Gemini) and built three custom GPTs. This website is 100% built with Claude. My publishing pipeline runs on Claude-powered automations and skills. When I assess an AI opportunity for your product, I'm not reading analyst reports. I'm sorting hype from real value from inside a working system. One filter: does it create real value? I write about it publicly. Read what I think before we talk.
The page exists so you can evaluate before we talk.
If something resonated, the scoping call is a working conversation — not a sales call.
You'll know within 30 minutes whether it's worth continuing.