AI-ENABLED EXPERIENCES
€2.5M in Projects from One Platform POC
6 Months → 3 Weeks Deployment | 5 Industry Verticals | Nexi Virtual Assistant Among Outcomes
RESULTS
- Revenue generated: €2.5M in follow-on enterprise AI projects
- Deployment acceleration: 6+ months reduced to 3-4 weeks (85% faster)
- Industry adoption: 5 verticals (Finance, Energy, Real Estate, Commerce, Sales)
- Platform components: Reusable component library eliminated redundant development
- Business model: Created new service line with modular pricing
The platform became the foundation for enterprise AI delivery at Sketchin. Projects like the Nexi AI Virtual Assistant emerged directly from capabilities we proved with this POC.
THE STORY
Enterprise AI in 2023 was stuck in demonstration limbo. Every company had impressive prototypes. Almost none could get them into production. The gap between "this demo is amazing" and "this is live and making money" killed most initiatives.
The problem was structural. Every project required months of custom development. Compliance frameworks had to be rebuilt from scratch. Measurement systems were afterthoughts. By the time something was production-ready, the business case had often evaporated.
I proposed building a platform that solved these problems once, then deploying it repeatedly across verticals. Not another demo. A production-ready foundation.
The problem was structural. Every project required months of custom development. Compliance frameworks had to be rebuilt from scratch. Measurement systems were afterthoughts. By the time something was production-ready, the business case had often evaporated.
I proposed building a platform that solved these problems once, then deploying it repeatedly across verticals. Not another demo. A production-ready foundation.
WHAT MADE THIS DIFFERENT
Most consulting firms sell custom development. Every engagement starts from zero. The client pays for the same compliance framework, the same measurement infrastructure, the same architectural patterns that were built for the previous client.
The platform approach inverted this. We built the foundation on our own time, then sold deployments that started at 70% complete. Clients got faster time-to-value. We got repeatable revenue and compounding expertise.
The risk was real: we were investing unbillable hours building something that might not sell. The bet was that "production-ready in 3 weeks" would be compelling enough to justify the upfront investment.
The platform approach inverted this. We built the foundation on our own time, then sold deployments that started at 70% complete. Clients got faster time-to-value. We got repeatable revenue and compounding expertise.
The risk was real: we were investing unbillable hours building something that might not sell. The bet was that "production-ready in 3 weeks" would be compelling enough to justify the upfront investment.
MY ROLE
Executive Strategy Director, Sketchin (Bip Group)
I conceived the platform strategy, designed the modular architecture, and led the team that built the initial POC. I also developed the business model and pricing structure that turned the platform into a sellable service line.
I conceived the platform strategy, designed the modular architecture, and led the team that built the initial POC. I also developed the business model and pricing structure that turned the platform into a sellable service line.
WHAT I ACTUALLY DID
The Pitch Problem
The hardest part wasn't building the platform. It was convincing leadership to let us try.
Consulting economics reward billable hours. I was proposing weeks of unbillable work on a speculative bet that we could productize AI delivery. Every hour spent on the platform was an hour not spent on paying clients.
I spent three weeks building the business case. I mapped competitor offerings, calculated the cost of repeated custom development, and modeled scenarios for platform-based pricing. The pitch that finally worked: "We're rebuilding the same compliance framework on every engagement. That's waste we can eliminate."
We got approval for a limited POC. Four weeks, small team, prove or kill.
The hardest part wasn't building the platform. It was convincing leadership to let us try.
Consulting economics reward billable hours. I was proposing weeks of unbillable work on a speculative bet that we could productize AI delivery. Every hour spent on the platform was an hour not spent on paying clients.
I spent three weeks building the business case. I mapped competitor offerings, calculated the cost of repeated custom development, and modeled scenarios for platform-based pricing. The pitch that finally worked: "We're rebuilding the same compliance framework on every engagement. That's waste we can eliminate."
We got approval for a limited POC. Four weeks, small team, prove or kill.
The Architecture
We built on three pillars that addressed the reasons enterprise AI projects failed:
Adaptive Interfaces: UI components that assembled themselves based on context, user, and moment. Not a fixed interface but a system that configured itself.
Agentive Technologies: Goal-driven agents with strict permissions, audit trails, and bounded autonomy. The AI could act, but within guardrails that satisfied compliance teams.
Proactive Models: Context engines that predicted user needs and triggered actions ahead of time. Not waiting for requests but anticipating them.
We built on three pillars that addressed the reasons enterprise AI projects failed:
Adaptive Interfaces: UI components that assembled themselves based on context, user, and moment. Not a fixed interface but a system that configured itself.
Agentive Technologies: Goal-driven agents with strict permissions, audit trails, and bounded autonomy. The AI could act, but within guardrails that satisfied compliance teams.
Proactive Models: Context engines that predicted user needs and triggered actions ahead of time. Not waiting for requests but anticipating them.
The Vertical Modules
Rather than a generic tool, we built five domain-specific modules that demonstrated the platform's flexibility:
Finance (Context-Aware Apps): Personalized financial advice and intervention based on transaction patterns and life events.
Energy (Power Assistant): Predictive management for home automation, anticipating needs and optimizing consumption.
Real Estate: Automated mortgage guidance combining income data, asset data, and property listings.
Commerce (Purchase Assistant): Intelligent shopping facilitation that understood context beyond the immediate transaction.
Sales/Advisory (Productivity Assistant): Relationship intelligence that surfaced opportunities and recommended actions to advisors.
Rather than a generic tool, we built five domain-specific modules that demonstrated the platform's flexibility:
Finance (Context-Aware Apps): Personalized financial advice and intervention based on transaction patterns and life events.
Energy (Power Assistant): Predictive management for home automation, anticipating needs and optimizing consumption.
Real Estate: Automated mortgage guidance combining income data, asset data, and property listings.
Commerce (Purchase Assistant): Intelligent shopping facilitation that understood context beyond the immediate transaction.
Sales/Advisory (Productivity Assistant): Relationship intelligence that surfaced opportunities and recommended actions to advisors.
The Compliance Layer
This was the piece that made enterprise sales possible. Most AI demos skip compliance because it's hard and unglamorous. We built it in from day one: data contracts, permission boundaries, and native audit trails.
When prospects asked "how do we explain this to our compliance team," we had answers. When competitors showed flashier demos, we showed audit logs. That distinction closed deals.
This was the piece that made enterprise sales possible. Most AI demos skip compliance because it's hard and unglamorous. We built it in from day one: data contracts, permission boundaries, and native audit trails.
When prospects asked "how do we explain this to our compliance team," we had answers. When competitors showed flashier demos, we showed audit logs. That distinction closed deals.
WHAT HAPPENED
The POC became a sales engine. €2.5M in projects within the first year, all built on the platform foundation.
Deployment times collapsed from 6+ months to 3-4 weeks. Clients weren't buying custom development anymore. They were buying configured deployments of a proven system.
The Nexi AI Virtual Assistant project emerged from this work. I didn't lead that engagement, but it demonstrated exactly what the platform was designed to enable: a major financial institution deploying production AI quickly because the foundation already existed."
We'd seen a dozen AI demos. This was the only one we could go live with in a month, track results, and build on. Our teams finally felt AI could work at enterprise scale."
- Technology leader, major client
Deployment times collapsed from 6+ months to 3-4 weeks. Clients weren't buying custom development anymore. They were buying configured deployments of a proven system.
The Nexi AI Virtual Assistant project emerged from this work. I didn't lead that engagement, but it demonstrated exactly what the platform was designed to enable: a major financial institution deploying production AI quickly because the foundation already existed."
We'd seen a dozen AI demos. This was the only one we could go live with in a month, track results, and build on. Our teams finally felt AI could work at enterprise scale."
- Technology leader, major client
WHAT I TOOK AWAY
System thinking beats feature thinking. The real leap happens when companies stop building point solutions and start building platforms.
The internal pitch was harder than the external one. Getting approval to invest unbillable hours required a business case as rigorous as any client proposal. That discipline made the eventual product better.
Compliance isn't a constraint. It's a competitive advantage. The teams that build compliance in from the start close deals that demo-only competitors can't.
The internal pitch was harder than the external one. Getting approval to invest unbillable hours required a business case as rigorous as any client proposal. That discipline made the eventual product better.
Compliance isn't a constraint. It's a competitive advantage. The teams that build compliance in from the start close deals that demo-only competitors can't.
