The Amazon Connect Podcast - Ep.25 - Be the 5%! w/Hannah Bloking and Brian Dawson
https://aws.amazon.com/blogs/contact-center/be-the-5-what-we-learned-by-shipping-ai-at-scale/
Most AI projects fail for the same reason: teams start with the technology, not the business problem. Brian Dawson and Hannah Bloking reveal why the companies that actually ship to production do something radically different - they work backwards from the customer experience, the data, and the controls they need before they ever reach for a model.
The conversation cuts through the AI hype cycle fast. Instead of chasing the latest acronym - RAG, MCP, A2A, and whatever comes next - Brian and Hannah explain why the real unlock is solving a specific business problem with the right mix of generative and deterministic workflows. They share the mistakes that stall deployments, from giant prompts and bad knowledge bases to teams that assume AI will magically fix broken processes.
You'll discover:
Why "be the five percent" is really about getting AI into production, not just running pilots
How to decide when an AI experience should be conversational and when it should stay simple
Why accuracy depends on the last five to ten percent of undocumented human knowledge
How observability and telemetry expose what customers are really asking for
Why security, legal, and compliance need to be involved early, not at the end
How small, bite-sized releases beat massive AI transformations every time
Hannah brings hard-won lessons from shipping AI at scale inside AWS and working directly with Amazon Connect customers. Brian adds frontline perspective from customer deployments around the world, including the patterns that make businesses move faster once they stop trying to force their problems into an LLM. Together, they show how trust, speed, and scale come from clarity - not complexity.
If you've been told to "do AI" without a clear plan, or you're trying to turn a promising proof of concept into something real, this episode gives you the framework to avoid the trap. Essential listening for customer experience leaders, contact center teams, product owners, and anyone responsible for making AI actually work in the real world.
Most AI projects fail for the same reason: teams start with the technology, not the business problem. Brian Dawson and Hannah Bloking reveal why the companies that actually ship to production do something radically different - they work backwards from the customer experience, the data, and the controls they need before they ever reach for a model.
The conversation cuts through the AI hype cycle fast. Instead of chasing the latest acronym - RAG, MCP, A2A, and whatever comes next - Brian and Hannah explain why the real unlock is solving a specific business problem with the right mix of generative and deterministic workflows. They share the mistakes that stall deployments, from giant prompts and bad knowledge bases to teams that assume AI will magically fix broken processes.
You'll discover:
Why "be the five percent" is really about getting AI into production, not just running pilots
How to decide when an AI experience should be conversational and when it should stay simple
Why accuracy depends on the last five to ten percent of undocumented human knowledge
How observability and telemetry expose what customers are really asking for
Why security, legal, and compliance need to be involved early, not at the end
How small, bite-sized releases beat massive AI transformations every time
Hannah brings hard-won lessons from shipping AI at scale inside AWS and working directly with Amazon Connect customers. Brian adds frontline perspective from customer deployments around the world, including the patterns that make businesses move faster once they stop trying to force their problems into an LLM. Together, they show how trust, speed, and scale come from clarity - not complexity.
If you've been told to "do AI" without a clear plan, or you're trying to turn a promising proof of concept into something real, this episode gives you the framework to avoid the trap. Essential listening for customer experience leaders, contact center teams, product owners, and anyone responsible for making AI actually work in the real world.
