Large enterprises move slowly. Decades of legacy systems, lengthy approval cycles, and complex integration debt make it structurally hard to move fast. Add healthcare regulation and global scale to the mix, and the barriers multiply. Most organisations today still have a ’POC graveyard’: promising AI pilots that never made it to production, never reached the people who needed them, and never delivered measurable impact. However, Philips, a 135-year-old, $20 billion global health technology company, built an AI Agent with a team of fewer than 20 people, generating instant results.
There’s a number that confronted Philips’ digital pathology sales team every morning: 45. That was the average number of days it took to get a quote to a customer. Forty-five days of chasing systems, digging through Excel sheets, losing momentum, and sometimes losing deals. “Sellers would wake up in the morning wondering if today was finally the day they would have that quote in their inbox to share with the customer,” said Christina Murphy, VP of AI and Business Operations at Philips, speaking at the AWS Summit in Amsterdam on May 27, 2026. “We no longer wanted to accept this.” So Philips asked a simple but audacious question: could AI compress that 45-day timeline down to just a couple of minutes?
The Real Challenge
In most large enterprises, solving a workflow problem like this would mean months of requirements cycles, heavy integrations, and lengthy approvals before any user gets their hands on something new. Philips knew the problem. The harder question was whether they could build a solution fast enough for it to actually matter. They decided to find out. A cross-functional team of fewer than 20 people — drawn from Philips, AWS, and software engineering firm EPAM — took an AI-first approach, using AI to build AI. The team used Kiro, AWS’s new AI-powered development environment, to accelerate the entire development lifecycle from planning and design through to testing and deployment.
This wasn’t just faster code generation. Murphy was emphatic on that point. The team wired user requirements directly to code and test cases for full traceability. And because this is healthcare — where governance isn’t optional — they built standards directly into the development cycle from day one: accessibility, security, logging, UX consistency, all standardized. “This was truly enterprise development at speed,” Murphy said. The result was a 75% faster development cycle compared to traditional approaches. A working prototype was ready in just two weeks. Six weeks in, they had their first proof of concept. And in under five months, the solution was live in production. For a 135-year-old, $20 billion global health technology company, those timelines were, in Murphy’s words, “before unheard of.”
Meet Sensai
The product that came out of this sprint is called Sensai, a new AI-powered sales experience built on Amazon Bedrock with agentic capabilities. Instead of navigating hundreds of systems and spreadsheets, sellers now interact with a simple chat interface. Behind it, a network of AI agents does the heavy lifting: dedicated pricing agents handle configuration and quoting, coding agents manage product-specific customization, and a manual activity orchestrator takes care of updating the CRM, sending out quotes, and drafting final customer emails — automatically.
Sellers can now generate a personalized customer preparation in minutes. A configuration that previously took days or weeks can be completed on demand. And the tasks that used to eat up their day — CRM updates, sending out quotes, writing follow-up emails, are handled by agents working in the background. “AI agents working for our sellers, so they can focus on customers,” Murphy said. “This isn’t just one task getting faster — it’s an entire workflow being compressed.”
“You Can Double My Quota”
Philips announced Sensai at its global sales meeting. The reaction, Murphy said, was immediate. “Sellers were begging for access.” Her personal favorite moment: a seller who told her, “If you give me Sensai, you can double my quota.” That quote landed. “That’s when we knew this was not about efficiency,” Murphy reflected. “This was about capacity for our sellers.” Murphy closed with a message she clearly wanted the audience to carry out of the room: if you want real impact from AI, you have to build AI with AI — and you have to bake governance in from the very beginning.
In healthcare, time isn’t just money. Time is patient outcomes. The faster Philips can turn an idea into a solution in the hands of a customer, the faster that customer — a hospital, a clinician, a health system — can deliver better care to more people. “If a 135-year-old enterprise with a team of under 20 can go from an idea to a product in the hands of sellers in under five months,” Murphy said, “then I know that today nothing is impossible anymore.”
