Inspiration
At Dial2AI, we work with enterprise customers where sales cycles often run from three to nine months. These conversations rarely move in a straight line. They usually shift between long quiet periods and sudden bursts of activity — follow-up emails, file sharing, product documents, video meetings, new stakeholders joining, internal customer discussions, and proposal updates.
We kept seeing the same problem: the customer context was getting scattered. One part of the conversation lived in an old email thread, another in a Zoom or Google Meet transcript, another in a shared file, and another in a short reply from a new stakeholder who did not have the full history. Before every follow-up, our team had to search across inboxes, meeting notes, documents, and video calls just to understand where the relationship stood.
We knew there had to be a better way — a way to keep every customer conversation, document, meeting, transcript, and decision connected, even when communication pauses for weeks and suddenly becomes active again. Out of that frustration came a simple conviction: customer context should not be buried across old email threads, disconnected files, and separate video conferencing tools. That is how the idea for SyncRoom was born.
We wanted to create a simple collaboration space where email, messages, video meetings, transcripts, files, meeting summaries, and follow-up actions stay connected in one shared room. At the same time, we did not want to force customers into a new communication habit. Email is still the most natural channel for external business communication, so SyncRoom works with email instead of replacing it.
With SyncRoom, a sales team can invite a client simply by sharing a secure room link over email. The client can participate without creating a SyncRoom account, receive updates in their inbox, reply by email, share documents by email, and join video meetings from email links. SyncRoom captures those replies, files, meeting transcripts, summaries, and next steps back into the same centralized room.
The result is a shared customer workspace where context is never lost — even when the relationship pauses for weeks, restarts suddenly, expands to new stakeholders, or moves from email to video meeting to proposal discussion. SyncRoom was inspired by our own need to make long-running customer conversations more organized, continuous, and actionable.
What it does
SyncRoom helps enterprise sales teams manage customer communication in one centralized, room-based workspace. Each SyncRoom acts as a shared space for a specific prospect, customer, opportunity, project, or partnership.
Sales teams can invite customer stakeholders into a room where they can participate without needing to create a password-heavy account or learn a complicated new tool. Customers can receive updates by email, reply directly from their inbox, join video meetings through calendar invites or email links, and access the room securely when they need the full history.
SyncRoom centralizes the key parts of the customer engagement process:
Messages and customer updates Email replies and threaded communication Video conference scheduling Meeting invites and calendar coordination Shared documents and files Video meeting transcripts Meeting summaries and action items Follow-up notes and next steps Stakeholder participation and role context AI-generated room summaries, analytics, and proposal support A core design principle behind SyncRoom is that it walks side by side with email. Email remains the easiest and most natural communication channel for external customers, so SyncRoom does not try to replace it. Instead, room updates are reflected over email, and email replies are automatically captured back into the correct SyncRoom space. This keeps the full conversation connected without forcing customers to constantly log into a new application.
SyncRoom also supports video conferencing. Sales teams can schedule meetings directly from the room and invite internal and external stakeholders. Meeting transcripts are generated, summarized, and posted back into the room. Key takeaways, next steps, and action items can also be shared with stakeholders by email.
Each room includes access to a SyncRoom AI agent. The agent can generate room summaries, identify risks and opportunities, analyze engagement patterns, suggest next steps, and support proposal writing. It can also create downloadable artifacts such as documents and PDFs. This helps sales teams prepare more effectively for follow-up meetings, understand where an opportunity stands, and respond with better context.
SyncRoom is built primarily for enterprise sales teams, but the underlying model can extend to many other cross-organizational workflows, including channel partnerships, R&D collaboration, customer success, legal collaboration, implementation projects, and strategic business development. Any relationship that depends on multiple stakeholders, long-running conversations, quiet periods, sudden bursts of activity, and shared context can benefit from SyncRoom.
By bringing together email, meetings, transcripts, files, stakeholder context, and AI insights, SyncRoom helps teams save time, reduce information loss, improve follow-up quality, accelerate proposal development, and maintain a clearer record of every customer relationship.
How we built it
SyncRoom is built using a multi-service architecture with four main components:
Frontend application Backend API server SyncRoom AI Agent server LiveKit AI Agent server for video meeting intelligence For the frontend, we used TypeScript, Next.js, React, and Tailwind CSS. The frontend integrates with Stripe for payments and subscription management, and it also uses LiveKit components for the video conferencing experience.
For the backend, we used Golang for the API server and PostgreSQL as the relational database. Files, meeting artifacts, transcripts, and generated documents are stored in Amazon S3.
Email integration is a central part of SyncRoom. Our backend uses Amazon SES to send outgoing emails, including OTP messages, room invitations, SyncRoom updates, meeting invites, and video conference links.
For inbound email, we use Amazon SES to receive emails sent back to SyncRoom. SES stores incoming emails in Amazon S3 and uses Amazon SNS to notify the SyncRoom backend through a secure webhook. The backend then fetches the email from S3, parses it, associates it with the correct room, and updates the SyncRoom database. This allows customers and stakeholders to participate directly from their email inbox while keeping the room context complete and organized.
Video conferencing is powered by LiveKit. Every meeting room can be attended by a LiveKit AI agent that listens to the audio and performs real-time transcription. At the end of the meeting, the agent uses an AI/LLM backend to generate a meeting summary, next steps, action items, and key takeaways. The transcript is stored in Amazon S3, and the summary and notes are sent back to the SyncRoom backend through a secure webhook.
For the SyncRoom AI agent backend, we developed a Golang-based server that streams responses over HTTP SSE. Agent-generated artifacts are stored in Amazon S3 and made available for download in formats such as DOCX and PDF. We use Claude Skills to extend the agent’s capabilities for room summaries, analytics, next-step planning, risk analysis, opportunity analysis, and proposal generation.
We also used Kiro IDE and Kiro CLI to support development across the frontend, backend, and AI agent layers. Kiro’s planning agent was especially helpful for brainstorming and designing complex workflows. Its specification-driven, task-oriented approach allowed us to review design decisions before implementation, break complex requirements into manageable tasks, and debug issues across the full stack.
SyncRoom is deployed on AWS ECS Fargate using a serverless container architecture. We use an Application Load Balancer as the load balancer and reverse proxy in front of the services. AWS CodeBuild and container image registries support our build and deployment process.
Challenges we ran into
Serverless Deployment Some issues only appeared after deployment to the serverless container environment and did not occur during local development. These included permissions, read/write access, container runtime behavior, Dockerfile configuration, buildspec settings, and service-level access requirements.
Debugging these issues required careful review of deployment logs, IAM permissions, container behavior, and AWS service configuration. Amazon Q was helpful in diagnosing deployment and runtime problems, especially where permissions, Dockerfile updates, and build configuration needed to be adjusted.
Routing Multiple Backend Services Behind One ALB SyncRoom uses multiple backend services, including the main API server and AI agent services. These services run as separate ECS services behind the same Application Load Balancer and domain.
Initially, routing rules were not always deterministic because multiple services needed to serve endpoints under the same origin. We resolved this by creating a clearer separation of route naming and API prefixes, allowing the ALB to forward requests to the correct target group reliably.
Managing Third-Party Integrations and Environment Variables SyncRoom integrates with multiple cloud and third-party services, including Amazon SES, Amazon SNS, Amazon S3, LiveKit, Stripe, and AI services. As the number of APIs, secrets, endpoints, and configuration values grew, managing environment variables manually became more complex.
ECS task definitions require accurate environment configuration, and even small mistakes can cause deployment or runtime failures. This reinforced the importance of improving automation, secrets management, and repeatable deployment processes as the product scales.
Inbound Email Processing Outbound email integration with SES was relatively straightforward, but inbound email processing required a more sophisticated architecture. We needed to receive customer email replies, store the raw messages, notify the backend, parse the email, map it to the correct room, and preserve the full communication context.
Building this flow with SES, S3, SNS, and our backend processing pipeline was one of the most important technical milestones of the product.
AI Agent Workflow Design SyncRoom is not just a communication tool; it also includes AI agents that analyze room activity, summarize customer engagement, identify risks, suggest next actions, and help generate proposals. Designing these workflows required us to think carefully about context, permissions, data access, streaming responses, artifact generation, and how AI-generated insights should support, not replace, human judgment in customer relationships.
Accomplishments that we're proud of
Production-Oriented Architecture We are proud that SyncRoom was built with a real-world deployment mindset from the beginning. The product runs on scalable AWS infrastructure using ECS Fargate, Application Load Balancer routing, automated build and deployment workflows, Stripe payment integration, and multiple third-party service integrations.
This architecture allows us to support pilots, offer subscription-based access, provide a free tier, and continue improving the service based on real usage. SyncRoom is not just a hackathon prototype — it is a product we are actively using internally at Dial2AI to manage our own sales and business development cycles.
Solving a Real Problem We Experience Ourselves One of the most meaningful accomplishments is that SyncRoom addresses a pain point we have personally experienced. Long enterprise sales cycles create communication fragmentation, and we built SyncRoom to solve that problem in our own workflow.
By using SyncRoom internally, we are continuously learning from real customer and partner interactions. This helps us improve the product based on practical feedback rather than assumptions.
Seamless Email Integration with AWS SES, S3, and SNS A major accomplishment was building an email-first experience that complements how customers already communicate. Our goal was not to replace email, but to make email more organized, contextual, and connected to the broader customer relationship.
Using Amazon SES for outbound email, SES inbound email processing, Amazon S3 for email storage, and Amazon SNS for event notification, we built a scalable pipeline that allows stakeholders to send messages and share documents by email while SyncRoom captures everything in the correct workspace.
This makes SyncRoom easier for external customers to adopt because they can continue using email while the sales team benefits from centralized context.
Kiro-Powered Development We are proud of how we used Kiro as a development partner throughout the build process. We adopted a specification-driven workflow where we worked with Kiro to define business requirements, review system designs, generate implementation tasks, and iterate through feedback.
Kiro’s repository understanding helped maintain context across the full stack, from requirements and architecture to implementation and debugging. Its planning mode was especially valuable for breaking complex features into smaller, manageable modules and asking clarification questions before implementation.
For a product like SyncRoom, which combines email, video conferencing, AI agents, payments, cloud infrastructure, and multiple backend services, this structured development process was extremely valuable.
AI Agent Integration We are also proud of the AI agent capabilities built into SyncRoom. The AI agent can support room summaries, customer engagement analysis, risk and opportunity analysis, next-step planning, and proposal generation.
By integrating Claude Skills and supporting artifact generation such as DOCX and PDF files, SyncRoom can help sales teams turn customer conversations into practical outputs. The agent can use real room context to help create better proposals, prepare for meetings, identify engagement gaps, and support more informed customer follow-up.
What we learned
Serverless Deployment on AWS ECS Fargate Deploying SyncRoom to AWS ECS Fargate taught us a great deal about serverless container infrastructure. During the early stages, errors appeared across multiple layers: Dockerfiles, application code, build configuration, container image creation, service permissions, and runtime access.
However, once the deployment pipeline was working, the experience became highly efficient. We could push code, trigger a build, deploy updated services, and move changes into production much faster.
We also learned that debugging in a serverless container environment is very manageable with the right observability tools. Amazon CloudWatch provided strong logging support and helped us understand runtime behavior. Managed SSL certificates and ALB integration also simplified domain setup, HTTPS configuration, and certificate renewal.
Subscription Pricing with Stripe In previous projects, we had mostly worked with one-time payments or enterprise invoice-based pricing. SyncRoom required a different model: subscription pricing, seats, feature-based access, free tiers, paid tiers, and upgrade paths.
Building Stripe subscription integration taught us how important payment automation is for SaaS products. We also learned that supporting a free tier is not just a nice-to-have feature. For a product aimed at sales organizations, teams need a low-friction way to try the service before purchasing.
At the same time, enterprise customers may still require invoice-based pricing, custom commercial terms, and procurement workflows. This means SyncRoom needs to support both self-serve subscription models and enterprise sales motions.
Spec-Driven Development for Enterprise Applications We learned that agentic coding is most powerful when it is connected to clear business requirements and system design. For a complex application like SyncRoom, with many moving parts and integrations, simply generating code is not enough.
Kiro’s multi-level approach — starting with specifications, then system design, then task lists — made it useful not only as a coding assistant, but also as a business analyst and architecture partner. It helped us think through requirements, evaluate alternatives, plan implementation steps, and maintain control over a complex codebase.
This was especially helpful when adding new features to an existing system, because Kiro’s planning mode could provide guidance, ask clarifying questions, and help us avoid rushing into implementation before fully understanding the design.
Email Remains Critical for Cross-Organization Collaboration One of our strongest product lessons was that email is still the default communication channel for external business relationships. Customers, partners, legal teams, finance teams, and executives are already comfortable with email.
Instead of trying to force users into a completely new communication tool, SyncRoom becomes more powerful by working with email. This makes adoption easier while still solving the bigger problem: scattered context.
AI Is Most Valuable When It Has Organized Context We also learned that AI agents become much more useful when they operate on structured, centralized, and relevant context. A generic AI assistant can answer questions, but a SyncRoom AI agent can analyze the actual room history, meeting summaries, stakeholder activity, files, and communication patterns.
That context allows the agent to provide more practical insights, such as opportunity risk, next best actions, engagement gaps, meeting preparation, and proposal support.
What's next for SyncRoom
Product and Platform Roadmap We plan to continue improving SyncRoom across several product areas.
First, we will enhance permission and access control so teams can manage internal users, external stakeholders, room-level visibility, and role-based access more precisely.
Second, we will improve the video conferencing experience with additional user experience features, including better meeting controls and options such as background filters.
Third, we plan to support enterprise identity and security requirements, including SAML, OAuth 2.0, and enterprise SSO integrations.
Fourth, we want to expand notification channels beyond email. We plan to explore AWS End User Messaging to support SyncRoom notifications through SMS and WhatsApp, allowing teams and customers to stay updated through the channels they prefer.
Fifth, we will continue enhancing the SyncRoom AI agent. Future capabilities may include deeper risk analysis, next-best-action recommendations, customer engagement scoring, advanced proposal generation, custom skills, dashboards, HTML artifacts, and richer downloadable outputs. We also plan to support more flexible agent execution environments, including Docker-based and microVM-based sandboxing.
Sixth, we will explore integrations with complementary tools such as Calendly, Boardy.ai, lead generation platforms, CRM systems, and other inbound or outbound sales tools.
Finally, we will continue improving the SyncRoom website with clearer product pages, solution pages, industry use cases, customer resources, videos, guides, and educational content so visitors can better understand how SyncRoom supports enterprise sales and cross-organization collaboration.
Go-To-Market Roadmap On the go-to-market side, our next steps include lead generation, prospecting, customer discovery, and running pilots with enterprise customers in target verticals.
We also plan to leverage the AWS Partner Portal and AWS Marketplace to explore customer opportunities, partner-led growth, and broader distribution.
Content marketing will be an important part of our strategy. We plan to develop blogs, user guides, product videos, customer support resources, and AI-assisted sales content to help customers understand the value of SyncRoom.
We will also invest in social media marketing and business development activities to reach sales teams, enterprise technology buyers, and potential partners.
Finally, we will explore partnerships with companies whose products complement SyncRoom, including sales enablement platforms, lead generation tools, collaboration platforms, communication tools, and enterprise workflow providers.
Our goal is to make SyncRoom the shared workspace where long-running customer relationships stay organized, searchable, actionable, and AI-enhanced from first conversation to final agreement.
Log in or sign up for Devpost to join the conversation.