Big Splash Advertising

Applied AI & Business Systems

Your company has AI.
Is it actually changing the business?

AI is being added everywhere—customer service, sales, marketing, operations, product experiences, internal workflows. The models are getting better. The demos are getting better. The investment keeps growing.The results are another story.
95%

MIT's Project NANDA found that roughly 95% of the organizations it studied were seeing no measurable P&L impact from their GenAI initiatives.

Source: MIT Project NANDA, 2025

That doesn't mean the technology doesn't work. It raises a much more interesting question:

What if your AI problem isn't an AI problem?

AI investment isn't the problem.
Diagnosis is.

Adding a better model doesn't necessarily give a system a better understanding of the problem.

It can summarize a customer without knowing what matters about that customer right now.

It can automate a workflow that never should have existed.

It can optimize a metric while making the larger business perform worse.

Before asking what AI should do, shouldn't we understand the problem it's entering?

There's a better way to solve AI problems

We start with the business system—then place AI where it creates measurable value.

01

Problem

What outcome and constraint actually matter?

02

System

How do customers, people, economics, and workflows interact?

03

Context

What information matters to this decision now?

04

Decision

Is the decision deterministic, AI-assisted, or human?

05

Action

What should happen, and who or what has authority to act?

06

Learning

What happened, what did it prove, and what should change?

REAL SYSTEMS.
REAL RESULTS.

See how this thinking has been applied to real businesses.

Expedia Group↗

Traveler intelligence & decisioning

The profile wasn't the intelligence.

We organized massive traveler profiles across lodging, flights, destination interests and behavior. Journey state determined which context actually mattered for the decision happening now.
Signature Concrete↗

AI voice intake & qualification

Trust comes before automation.

The voice agent doesn't pretend to be human. It earns permission to continue, handles the conversation, and qualifies demand against what the business can actually deliver.
Driftwood Retreat↗

AI & deterministic automation

Knowing when not to use AI.

Predictable guest and crew communications run through ordinary automation. AI has one specific job: turn unstructured guest language into structured operational context.
Adorama / Printique↗

Experimentation, UX & AI

Let customers experience value first.

Promotion testing pushed us toward product-led growth. Instead of surveying customers, we let them upload their photos first—then used AI to turn those photos into an immediate product experience.

About David

Built from the operator's side of the table.

Operator-level applied AI, shaped by an enterprise product background and applied at small-business scale. The focus isn't adding AI for its own sake. It's understanding the system, finding the decision that matters, and building something the business can trust.

Start a conversation →