AI and technology consulting / Los Angeles
We help YouTube creators, Hollywood studios, and media teams decide where AI belongs and build the workflows to use it. Our work connects audience insight with generative AI, production, VFX, and new formats, including microdramas and shorts.
01 / Experience
David Carroll has built media technology through Hollywood’s shift to digital cinema, the growth of the creator economy, and the arrival of production generative AI.
Source: David Carroll’s career account. Figures describe work at Spotter and its predecessor.
David Carroll’s career
Digital cinema
Co-architected Sony’s 4K digital cinema projection system and conceived Media Backbone Conductor.
CTO office
Next-generation delivery for services including NFL Sunday Ticket.
Founding CTO, twelve years
Built a cloud platform that analyzed more than a billion YouTube videos.
Head of AI R&D at Spotter
Production AI for top YouTube creators, including a thumbnail generator that delivered 4.1 million images.
Founder, Los Angeles
Bringing that experience to companies navigating their own media and AI transitions.
02 / Services
We connect business priorities, audience knowledge, and production requirements to a practical technology plan.
Choose where AI can improve your business and define what a useful first investment looks like.
Deliverables
Turn channel performance and creative knowledge into tools your team can use.
Deliverables
Design the path from generated material to reviewed production assets.
Deliverables
Build a repeatable process for testing and producing emerging media formats.
Deliverables
Engagements are scoped around the decision you need to make and the systems you need to operate.
03 / Applications
The work is most useful when it begins with a specific decision, bottleneck, or creative requirement.
Your team needs more useful ideas without losing the voice that built the channel.
We can structure your creative knowledge, connect it with audience signals, and design tools for concepts, titles, thumbnails, and editorial review.
Your teams are experimenting with AI, but leadership needs a clear investment decision.
We can assess candidate workflows, compare implementation options, and define a pilot with measurable outcomes and accountable owners.
Generating an image or shot is one step. The production still needs consistency, revisions, approvals, and usable handoffs.
We can map those requirements into a workflow that gives artists and producers clear control over the result.
You need to test stories and formats while keeping characters, visual choices, and production decisions consistent.
We can design a process that connects the series bible, production tools, release cadence, and audience feedback.
04 / Approach
We agree on the decision, test the assumptions, and document what your team needs to operate the result.
Establish the business question, creative requirements, and constraints.
Typical output
A written scope and measures of success.
Examine the workflow, available data, tools, and handoffs.
Typical output
An opportunity map and architecture recommendation.
Evaluate a focused pilot against real examples and agreed criteria.
Typical output
Evidence about quality, cost, and operational fit.
Document the system and prepare your team to run and improve it.
Typical output
An implementation plan, operating guidance, and training.
An engagement can focus on a single decision or continue through pilot design and implementation. Scope and responsibilities are agreed before work begins.
05 / AI architecture
A useful AI system needs access to the right information, clear instructions, and a way to judge its output. It also needs people who can see what happened and decide what happens next.
Exceptions and revisions return to the model
Our architecture approach considers how models use approved source material, call tools, pass work between steps, and return results for human review. We define how the system will be evaluated, what it will cost to operate, and how it should respond when a step fails.
David plans to pursue Claude architecture certification as part of his continued work in applied AI systems. His technical focus includes tool use, context design, evaluation, and workflows that keep creative teams involved in key decisions.
06 / Founder
David Carroll has spent four decades building the systems Hollywood runs on, from the digital projectors in your local theater to the AI tools behind today’s biggest YouTube channels. He founded Customer Insights LLC in Los Angeles to bring that experience to companies navigating their own media and AI transitions.
At Sony Electronics, David worked at the center of Hollywood’s conversion from film to digital. He co-architected Sony’s 4K digital cinema projection system and helped restructure its economics as a service rather than an equipment sale, a shift that drove $1 billion in new revenue and put 4K projection into theaters worldwide. He also conceived and architected Media Backbone Conductor, the workflow orchestration platform that NBC used to process its streaming content for the 2012 London Olympics, at the time the most-watched television event in U.S. history. Sony still sells Conductor’s descendants today as part of its Ci Media Cloud platform. Between Sony and his startup years, he served in the CTO office at DIRECTV, working on next-generation delivery for services including NFL Sunday Ticket.
David then spent twelve years as founding CTO of Spotter and its predecessor, the SoftBank-backed company that reached a $1.7 billion valuation financing YouTube creators. He built its technology alone at first, then scaled it into a cloud platform that discovered, warehoused, and analyzed more than a billion YouTube videos. The valuation models running on that platform priced the deals behind over $500 million deployed to creators.
As Spotter’s Head of AI R&D, he moved from analyzing creator content to generating it. His team built the production AI systems behind Spotter Studio, the SaaS product used by top YouTube creators, including ideation tools and a thumbnail generator that delivered 4.1 million AI images. Central to that work was the writer’s guide, a structured “creator bible” that captured each creator’s voice, style, and creative rules so that every generative tool in Studio produced work the creator would recognize as their own. He filed two AI patents along the way.
The thread through all of it: understanding what an audience wants, at the scale of a billion videos or a single creator’s channel, and building the systems that act on that understanding.
07 / Contact
Tell us about the decision, workflow, or production challenge in front of you. We can discuss where Customer Insights may be useful and what a first engagement could cover.
