Case Studies | SaaS B2B

From a conversation to a ten-minute film

A SaaS B2B client

About

A SaaS B2B business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The challenge had three focus areas, each a prerequisite for the MovieGame experience to feel coherent.

A planning conversation that holds context

The agent needed to converse with an user turn by turn to refine storyline, characters, tone, and scene breakdown, maintaining full context across the session and only transitioning to generation on an explicit user confirmation. Losing context mid-plan would break the creative collaboration the product is built on.

Coherent long-form video, not disconnected clips

A ten-minute film is well beyond a single model generation, so it had to be produced as five sequential clips with visual continuity

Autonomous, reliable orchestration

Once a plan is confirmed, the platform had to drive the entire generation pipeline without manual intervention, tracking the state of every clip and processing long-running generation jobs asynchronously so users are never left with a silently stalled render.

Solution

In the planning phase, a single Amazon Bedrock AgentCore orchestration agent converses with the user through a chat API, persisting each exchange to AgentCore short-term memory so the full plan stays in context. When the user confirms, the agent extracts a five-scene plan and records a generation job in Amazon DynamoDB, signalling the transition to generation.

1 agent

Single Bedrock AgentCore agent handling both planning and generation

5 clips

Sequential Nova Reel clips chained by automatic keyframe handoff

~10 min

Continuous stitched film produced per confirmed storyline

By the numbers:

  • 1 agent - Single Bedrock AgentCore agent handling both planning and generation
  • 5 clips - Sequential Nova Reel clips chained by automatic keyframe handoff
  • ~10 min - Continuous stitched film produced per confirmed storyline
Changes

The Proof of Concept met its success criteria, demonstrating the complete MovieGame journey end to end: a multi-turn planning conversation, a confirmation-triggered handover, autonomous sequential generation, and a single stitched film. Acceptance was based on that end-to-end feasibility being shown on AWS.

  • Context-aware planning — The orchestration agent conducted a multi-turn conversation, maintained context through AgentCore short-term memory, and transitioned to generation only on explicit user confirmation.
  • Visual continuity — Keyframe extraction between clips carried the final frame of each scene into the next, so the five clips assemble into one continuous story rather than disconnected fragments.
  • Asynchronous by design — Every generation job ran asynchronously with per-clip state tracked in DynamoDB and artifacts accessible via S3 pre-signed URLs, keeping long renders observable rather than opaque.
  • Modular and swappable — Video generation was encapsulated in a dedicated Lambda tool, so the underlying model can be replaced or supplemented with little to no change to the agent or pipeline.
  • Handover — Delivery included AWS CloudFormation templates for the full stack, CloudWatch observability, Loom walkthroughs, architecture documentation, and a sign-off call.

The result is a modular AWS foundation that proves the core MovieGame capability and gives the client a solid base to build the production service on, including story branching, multi-user support, and frontend integration.

AWS Stack

Amazon Bedrock AgentCore

For the orchestration agent that runs both the conversational planning and generation phases, with short-term memory for context.

Amazon Nova Reel

For asynchronous AI video clip generation with keyframe-anchored visual continuity.

AWS Lambda

For the API handlers and the agent tools that submit, poll, extract keyframes, and stitch the final video.

Amazon API Gateway

For the chat, job status, and pre-signed URL endpoints.

Amazon DynamoDB

For per-clip job state tracking and pipeline coordination.

Amazon S3

For secure storage of reference images, clips, keyframes, and final videos, served via time-limited pre-signed URLs.

AWS CloudFormation

And Amazon CloudWatch for infrastructure provisioning and observability.

YOU MIGHT LIKE

Related success stories

View all case studies

Case Studies | Insights

Utilising Language Recognition, Speed, and Enhanced Security to Make Social Media a Force for Good

  • Here, we take a detailed look at how the Cloud Combinator team collaborated with another cutting-edge AI service provider that provides intelligent systems to “make social media more social” for brands and users alike.
  • Arwen AI is a UK-based startup specialising in AI solutions to manage and enhance brands’ social media interactions. Founded in 2020 by Matt McGrory, Dr. David Cole, and Joel Bailey, Arwen. AI focuses on using AI to automatically detect and remove spam, toxic comments, and other unwanted content from social media platforms.
  • The team at Arwen have three core products. ‘Moderate’ is focused on identifying and removing toxic content from social media channels. ‘Engage’ helps brands identify and engage with meaningful conversations on social media, and ‘Customize’ allows brands to apply bespoke algorithms to their channels - creating an even more effective moderation and engagement.
Read more
CONTACT US

Ready to turn AI into impact?

We'll help you spot the highest-value opportunities, reduce risk around your first AI initiative, and define a clear path to results from day one.

Why talk to us:

Outcome-driven recommendations

AWS-recognised delivery expertise

Risk-aware AI adoption

Clear next step, not a sales pitch

Start with a focused 20-minute conversation about your goals — no pressure, no commitment.

We use cookies

This website uses cookies to enhance user experience and to analyze performance and traffic on our website.

See our Privacy Policy for details.