TL;DR
In this first episode of Spark Sessions, Chris Murphy, MD and Founder of Tuncarp, joins Spark AI for a live demo of what embedded AI capability actually looks like. Tuncarp is a 50-person design agency based in the Philippines. As their AI production work grew, jumping between multiple platforms became increasingly difficult to manage. So they built their own solution. Chris walks through the operating system his team created, covering everything from brief intake and multi-model generation to approvals and IP governance.
Episode 1 of Spark Sessions
Efficient teams: from fragmented tools to a unified AI operating system
This is the first episode of Spark Sessions, Spark AI's new monthly series. The question we hear most from agencies right now is: what does good actually look like? This series is our answer to that. Each session, an agency walks us through exactly how they've embedded AI into the way they work.
In this session, EJ (Spark AI coach) and Chris Murphy, MD and Founder of Tuncarp, walk through the operating system Tuncarp built to replace fragmented, multi-platform workflows. Chris does a live demo of the system in action, so you can see exactly how it works in practice.
Duration:
30 Minutes
Who its for:
Agency Founders, Creative Directors, and Production Leads
Who its for:
AI Workflow Design, Creative Production, IP Governance
Format:
On-Demand Video
The production problem Tuncarp solved
As Tuncarp's AI production work grew, so did the complexity. They worked across various AI platforms, as each was suited to something different, so managing across them became increasingly difficult. The breaking point came when they were producing a campaign that required 8,000 images and 1,500 video clips. The team decided to do something about it and built their own solution: a single workspace layered across multiple LLMs, designed around how they actually work.
Inside Tuncarp's creative operating system
From a single interface, their team can run a brief through a conversational agent, generate assets across several models simultaneously, compare outputs side by side, iterate on brand-consistent visuals, and route work through a structured approval flow, all without leaving the system.
How Tuncarp handles IP and client accountability
For large clients, the main blocker to AI is risk. Tuncarp have built accountability into the system itself: every generation logs the prompt, the model, and the timestamp. Prompts are scanned for IP risk before generation, and watermarking is built into the output workflow.
FAQs
-
Yes. You may not be able to build a bespoke system, but the principle is the same. Find your biggest friction point and fix it. At the most basic level, that means shared standards, documented processes, and making sure everyone on the team is working the same way. A well-built SOP shared across your team is a meaningful step in the right direction.
-
It was a team-led decision. The production team identified where the friction was and built something to fix it. That instinct - to start with the biggest internal pain point rather than looking for an off-the-shelf solution - is one of the most useful things you'll take from this session.
-
Around four weeks to reach a working MVP. The team started experimenting with aggregator tools in August, and by the Christmas period had built the first version of the creative workspace. Three months later, it had become a production-grade system.
-
Each session is free to attend. We announce upcoming episodes on our LinkedIn and through the Spark Intelligence newsletter. Recordings are available on demand via the Spark AI website, or directly through our youtube.
Proven AI literacy results for agencies
After working with Spark AI our clients tell us:
9.3/10
Average course rating
100%
Agency Leaders would recommend Spark AI
100%
Felt more confident
4/5
Felt more or much more optimistic (nobody felt less!)