Amazon Connect – What we're excited about for 2026

Date Published

Cover image for Amazon Connect – What we're excited about for 2026

Amazon Connect continues to evolve at a pace that’s reshaping the contact centre landscape. New features are constantly being released that not only push the boundaries of AI but also empower agents, streamline operations, and enable more personalised customer experiences.

Here are the announcements Cortex Reply is most excited about.

AI-Powered Experiences

AI is, unsurprisingly, front and centre this year. What stood out to us was the meaningful progress towards making AI operationally safe, governable, and reusable at enterprise scale.

New AI Agents (with MCP!)

We’ve spent a lot of time building AI-powered experiences, and the most time-consuming part was building bespoke integrations for each use case. MCP directly addresses this.

Now, instead of building bespoke tools and integrations for every new AI agent or use case, MCP creates a standardised integration layer. Once core systems like CRM, order management, or case platforms are exposed via MCP, they can be reused safely across multiple AI agents and channels.

This is exciting because it means faster time-to-value, less duplication of effort, and stronger governance. New AI agents can be introduced by configuration and orchestration, not re-engineering.

We’ve already spent time exploring this capability, and it meaningfully lowers the effort required to build powerful, well-governed, AI-driven customer journeys—while keeping enterprises in control of how AI interacts with their systems.

AI Agent Observability and Evaluation

Another valuable release we’re excited about is how AI agent observability and evaluation addresses  one of the biggest blockers to AI adoption we see – the lack of confidence: “How do we know it’s working?” and “How do we spot issues before customers do?”

Improved visibility into AI-led interactions, hand-offs, sentiment, and performance — combined with automated quality evaluation — gives teams the tools they need to iterate safely. This shifts AI from a “set and hope” model to something that can be continuously improved, just like any other critical customer-facing system. 

More Natural Voice Experiences

This is exciting because voice quality significantly impacts the perceived effectiveness of AI-powered experiences. With the integration of Nova Sonic and support for third-party speech providers, Amazon Connect can now deliver more natural and human-like voice interactions, something particularly important to our clients. 

This flexibility allows organisations to select high-quality voices that better reflect their brand, resulting in a noticeable improvement in perceived experience and reduced friction. When voice interactions feel natural, customers are more willing to self-serve — and that directly impacts containment, satisfaction, and cost.

Bedrock Knowledge Bases Support

We’re excited because, like the benefits of MCP, the ability to reuse existing Bedrock Knowledge Bases addresses a common problem: organisations maintaining duplicate knowledge bases that create confusion, inconsistencies, and additional operational overhead. By leveraging existing, governed, application-agnostic knowledge sources, organisations can achieve faster AI agent rollout, consistent answers across applications, and reduced ongoing maintenance effort.

Personalised Product and Service Recommendations

Predictive insights and recommendations are something we’re particularly excited to see. Many of our conversations centre on moving beyond generic personalisation, and this stands out as a clear step towards intelligent hyper-personalisation across self-service and agent-assisted journeys without increasing complexity. Used well, this isn’t about aggressive upselling. It’s about enabling more relevant conversations, driven by existing customer data. At the moment, it actually matters.

Agent Empowerment

What stands out to us here isn’t just new functionality, but how these updates reduce friction in day-to-day agent work - something we think will make a real difference for our clients. These capabilities directly address common pain points we see in live engagements — helping agents work more confidently, efficiently, and with fewer distractions.

Agent-initiated Chat Workflows

This is exciting because giving agents the ability to trigger workflows directly within chat conversations is a small change with a big impact. Securely collecting sensitive data or sharing disclosures without disrupting the flow reduces friction for customers and agents alike — and we see this as particularly valuable for our clients operating in regulated environments.

Customised Workspaces

We’re excited about this because customisable, persona-based workspaces are something clients have repeatedly asked for. By surfacing only the tools and information relevant to a specific role, these workspaces accelerate onboarding, reduce cognitive load, and enable agents to focus on what matters most. And, the addition of full theming also allows large organisations to deliver a cohesive, enterprise-grade experience across teams and business units without compromising usability.

AI agent assistance and summarisation for Agentforce Service

We’re excited about this because we’re seeing increasing adoption of Amazon Connect with Salesforce, particularly where organisations want to avoid agents switching between systems. Real-time AI assistance and automated summarisation improve data quality and reduce agent workload, helping agents stay focused on the customer rather than note-taking.. For organisations already invested in Salesforce, this meaningfully strengthens the case for Amazon Connect as a native, enterprise-ready contact centre platform.

Accelerating Outcomes

These releases matter because they shorten the gap between change and impact, which we believe will be particularly powerful for our clients. From our experience, operational agility is often the limiting factor in contact centre performance — and these capabilities directly improve how quickly teams can adapt, test, and respond.

Adaptive IVRs and Operations

Adaptive IVRs are a capability we have previously designed and implemented through custom integrations, so it’s exciting to now see it supported out-of-the-box. We’re excited because automating, or allowing operations teams to adjust, routing, IVR messaging, and notifications in real time, enables organisations to respond. faster to disruptions, reducing unnecessary agent interactions while keeping customers informed. We see this as a huge opportunity to transform historically challenging scenarios into seamless, well-managed customer experiences.

Smarter Performance Evaluations

Automatic selection of relevant contacts, combined with enhanced evaluation criteria — including call avoidance, audibility, and handling efficiency — are two key features we see as transforming how supervisors approach performance management. Helping supervisors focus coaching efforts where they have the most impact, rather than manually reviewing large volumes of interactions.

Testing and Simulation Capabilities

We’re excited because this significantly lowers the risk of change. Developing new customer experiences is inherently complex, and even minor issues can have a significant impact on customers and brand trust. We see the new testing and simulation capabilities as particularly valuable to our clients as they enable teams to identify and resolve issues before they reach customers, providing a level of confidence previously only achievable with costly manual testing. This not only reduces risk but also accelerates deployment and drives faster time-to-value.

WhatsApp for Outbound Campaigns

Support for WhatsApp in outbound campaigns reflects a broader shift towards proactive, conversational engagement on channels customers already trust. This is exciting because companies can send proactive messages about disruptions or appointment reminders and handle responses while keeping all the familiar campaign features—such as scheduled delivery, compliance guardrails, and engagement tracking. We see this as a significant step toward proactively keeping customers informed, ultimately helping reduce contact volume.

Operational Enhancements

  • Several operational enhancements stood out to us as particularly impactful for our clients. These are the kinds of capabilities that improve day-to-day control, observability, and compliance — and while individually small, they make a powerful difference in how contact centres operate at scale, especially in regulated or migration-heavy environments
  • Custom metrics with no-code creation, available via dashboards and APIs
  • Improved alerting with richer operational context
  • In-flight data redaction and message processing before content reaches participants

These enhancements are exciting because they will be particularly valuable in regulated environments and migration scenarios where continuity and control are critical.

Final Thoughts

What excites Cortex Reply most isn’t any single feature — it’s the way these capabilities work together to reduce friction between idea and delivery, while giving enterprises the control they need to adopt AI with confidence.

If you’d like to talk through what this means for your own Amazon Connect roadmap, or how these capabilities fit into real-world architectures, we’d be happy to continue the conversation.

Ben Num

Ben Num

Talk to us about transforming your business

Contact Us