Building an AI-first healthcare solution using AI on AWS with Promptly

Specialty medical practices face a particular kind of complexity. A front desk worker at a specialty clinic typically juggles separate browser tabs: one for scheduling, one for billing, one for patient messaging, one for insurance eligibility, one for prior authorizations. For patients who are stressed or unwell, that disconnection is visible—they reenter information across systems, wait longer, and feel like just another appointment. Promptly built an all-in-one patient engagement system to replace those disconnected tools, serving ophthalmology, dermatology, orthopedics, urology, and other specialty markets. Revenue doubled year over year, but that rapid growth surfaced new internal complexity.

As Promptly’s software expanded into new specialty markets, its documentation, support tickets, and customer relationship data spread across disparate systems. Support specialists searched manually through configuration guides to answer customer questions, sometimes pausing a live call to locate the right document. Developers worked without a unified AI assistant, each choosing whichever tool they preferred, with no shared security posture and no institutional knowledge built into the workflow. The result was slower development cycles, inconsistent support quality, and a growing gap between what the product could do and what any individual staff member could quickly explain.

When DoiT introduced Promptly to the AWS Business Value Realization Initiative, Promptly saw an opportunity to take its AI ambitions further. Promptly had already worked with DoiT for several years on cloud management and consulting, building a foundation of trust through DoiT’s consistent delivery of proven, reliable solutions.

DoiT’s team conducted structured interviews with Promptly’s operations, support, and sales leaders, mapping each team’s daily friction points before recommending any technology. The first priority was consolidating Promptly’s knowledge base. DoiT connected tens of thousands of documents from approximately 50 shared drives into Amazon Bedrock—a service for building generative AI applications and agents at production scale. On the customer-facing side, Promptly used Amazon Bedrock to build a conversational chat interface integrated into the Promptly web application. The widget guides patients through appointment scheduling, bill inquiries, and insurance questions without requiring a staff member.

For developer productivity, DoiT introduced Amazon Q Developer, a highly capable generative AI–powered assistant for software development. Promptly’s more than 40 engineers now consult Amazon Q Developer throughout every sprint. The tool performs an immediate code review pass before formal quality assurance, reducing the risk that new code will break in production and accelerating the path to beta testing.

For internal knowledge management, the team adopted the Amazon Quick Suite to help answer questions and turn answers into actions using agentic teammates. Promptly staff now call the resulting AI assistant “the Oracle of Promptly”—a single interface that any team member can query without interrupting a colleague.

Amazon Q Developer now accounts for 30–40 percent of code written across the engineering team. Developers reach beta faster because AI-assisted code review runs before formal quality assurance begins—resulting in an approximately 25 percent improvement in development speed. The customer-facing Amazon Bedrock chat is live, and a version powered by Amazon Bedrock AgentCore—an agentic service for building, deploying, and operating effective agents securely at scale—is entering beta with multiple clinic customers. Support specialists report quicker resolutions and higher satisfaction scores because they can surface precise answers from thousands of pages of product documentation in seconds.

“No matter how many modules we add, anyone from any department can get the information they need quickly, easily, and efficiently using AI agents—and the team that supports that aspect of the system can keep doing their actual work,” says Dr. Anish Kapur, CEO and founder of Promptly.