Why enterprise AI needs product thinking, not just models
Most enterprise AI stalls on adoption, not capability. What changes when you treat AI as a product with users, trust, and outcomes.
Director of Enterprise Applications & Digital Transformation at OPSWAT. I set the AI and platform strategy for a global cybersecurity company, architect the systems myself, and lead the teams that ship them. Lead architect on the platforms below — and the leader of the org that runs them.
My career started in financial-services technology — trading platforms, accounting systems, SWIFT infrastructure, Salesforce — at Schroders and FPT Software. Over fifteen years I moved from supporting systems, to designing them, to building internal products that reshape how people work.
That idea drives my work at OPSWAT, where I own strategy and delivery of internal platforms across CRM, revenue operations, enterprise data, AI automation, and employee-facing products. The future of enterprise technology is not buying more SaaS — it's building intelligent products on top of trusted systems of record, where company-specific context becomes a competitive advantage. That belief shaped OSCAR, Qlari, Sugar Data Tool, and the Salesforce Metadata Archive.
I lead from the front. I set the strategy and own the outcome — and I'm the lead architect on the systems that deliver it, hands on the keyboard — then I build and grow the team around them. I take messy business problems and turn them into software that feels obvious in hindsight — the kind where users ask, "Can I just work from this instead?"
Enterprise applications shouldn't be a cost center. Combined with data, AI, and product thinking, they become the fastest team in the company.
Agentic platforms that reason over enterprise knowledge, use tools, trigger workflows, and operate with governance — RAG, orchestration, MCP tools, LLM gateways, human-in-the-loop controls.
Moving organizations from internal tools to internal products — designed around adoption, workflow fit, business outcomes, and trust.
Systems that help Sales, RevOps, CX, Finance, and executives decide — forecasting, deal health, account research, pipeline visibility, AI-generated insight.
Build-first where company-specific context compounds; buy-on-purpose where scale, maturity, or compliance make SaaS the better choice.
AI is only useful when the data foundation is trustworthy — warehouses and semantic models, APIs and sync, data quality, and AI-ready knowledge layers.
I own the outcome and the scope — and I stay close enough to the work to be the lead architect. The two aren't in tension; the depth is what makes the leadership credible.
On the flagship systems I'm hands-on the architecture and the code. I don't ask the team to build what I wouldn't.
Shifted enterprise applications from a cost center into an internal product org — built around adoption, SaaS rationalization, and AI-native workflows.
Built a 35-person organization from zero across platforms, data, AI, integrations, and QA — and developed the leaders who run it.
Build where company-specific context compounds; buy where scale, maturity, or compliance make SaaS the right call.
Each is a system I designed and built — and the outcomes it drives for the business. Click any to explore the decisions, architecture, and impact.
Consolidated fragmented chatbots into one governed, company-wide agentic AI platform.
ExploreReplaced Clari + Revic — ~$280K/yr saved — with sharper, AI-native revenue intelligence.
ExploreOne decision surface for five functions — Customer-360 & revenue intelligence on the warehouse.
ExploreThe governed data foundation for the whole company — 30+ sources, self-hosted, zero lock-in.
ExploreOne internal front door for the whole company — five functions on one identity and audit model.
ExploreA multi-tenant legal AI platform for law firms — Vietnamese-first, privacy by construction.
ExploreThe trusted data layer beneath revenue & AI — safe-by-default CRM data operations.
ExploreDe-risked a full CRM migration — years of business logic preserved, explained, and mapped.
ExploreKeyboard-first case-triage workspace for support agents, with a browser-extension OAuth bridge.
ExploreProduct-knowledge platform with AI-generated marketing assets published via git-synced review.
ExploreSelf-hosted dynamic-DNS updater for homelab and NAS — scoped tokens, sync dashboard, Docker-first.
ExploreZero-dependency port-path inspector that tells you WHY a port is unreachable.
ExploreBidirectional case↔ticket sync with HMAC-signed webhooks, polling, and comment mirroring.
ExploreA tool completes a task. A product earns preference.
AI that creates another workflow is theater. Good AI reduces steps, decisions, and cognitive load.
Don't build because SaaS is expensive. Build where company-specific context compounds.
For enterprise AI and business systems, trust is not a feature. It is the foundation.
Employees should not have to tolerate bad software just because it is internal.
Own strategy and delivery of internal products and enterprise applications across revenue operations, customer experience, employee services, enterprise data, and AI automation for a global cybersecurity company.
Led CX operations systems, Salesforce platform stabilization, AI chatbot development, and support experience modernization.
Eleven years across front-office trading platforms, APAC Salesforce, automation, and global technology operations for a leading asset manager.
Enterprise workflow, integration, Android, .NET, and data-warehouse solutions for financial-services clients — including K2 workflow platforms extending SimCorp Dimension and SSIS data pipelines.
Texas A&M UniversityB.S. Computer Engineering (primary) · Mathematics (secondary) · 2005 — 2009
Topics I speak and write about — the throughline is enterprise technology as a product and a strategic asset, not a cost center. Available for talks, podcasts, and guest writing.
Most enterprise AI stalls on adoption, not capability. What changes when you treat AI as a product with users, trust, and outcomes.
When company-specific context makes building beat best-in-class SaaS — and when buying is the disciplined call.
Governance, tools, and trust — what it actually takes to run agents against real business systems.
Rationalizing a fragmented SaaS estate into a few internal products employees prefer to use.
What a tailored, AI-native revenue platform does that off-the-shelf forecasting tools can't.
The operating model and leaders that turn internal IT into the fastest team in the company.
Enterprise AI, internal product strategy, build-vs-buy, revenue intelligence — or a messy business problem that deserves better software. I'm always up for a good conversation.
[email protected]