Ai-powered, Expert-led
The Work
The work represented here ranges from developed prototypes to smaller investigations and thought experiments.
The goal was in each case was the same: build enough of the idea to test assumptions, understand the implications, and learn.
When prompts go wrong.
Cadence
AI-Assisted PPC Maintenance
Cadence began with a simple question: can AI reduce the time and cost required to maintain mature Google Ads accounts without sacrificing account quality or giving up control over the process?
The prototype combines live campaign data with persistent client context and a defined set of rules, then uses AI to evaluate account performance and recommend what should happen next. The system reasons inside a business and operational context rather than looking at campaign data in isolation.
This is an active project.
Lamina
AI-Native Campaign Strategy
Lamina was an AI-native strategy engine designed to turn fragmented marketing data into forward-looking campaign plans.
The project combined time-series forecasting, retrieval-augmented generation, and LLM reasoning to produce strategic briefs with projected KPI ranges, cited priors, and explicit rationale.
The core question was whether AI could help agencies move beyond backward-looking dashboards and slow planning cycles toward faster, more evidence-based campaign strategy.
Pamela
Intelligent Businesses System
Pamela addresses a weakness in the consulting and business OS space. Most consulting frameworks and operating systems may help organizations define priorities and processes, but the burden of execution still falls to the organization.
Pamela explored what has to exist around an LLM before it can safely participate in real business work, providing an intelligent operations coordination layer that can actively participate in advice and execution.
The project focused on AI-enabled coordination of work across people and systems.
Brand Positioning Audit
AI-Assisted Strategic Assessment
This experiment explored whether an AI-based brand assessment could make expert strategic thinking more accessible, consistent, and scalable.
It was designed not only to assess what was present, but to identify missed opportunities, weak signals, and areas where the positioning could be made clearer or more differentiated.
More than 25 free audits were completed as part of the experiment, creating a useful test of how well AI could support a repeatable strategic framework while still leaving room for human interpretation and judgment.
Other Investigations Included:
-
A context-aware multimedia card that dynamically assembles information based on the people and situation involved.
-
A Google Places lead-generation workflow combining business discovery, enrichment, and AI-assisted outreach, along with smaller experiments around marketing, workflows, product concepts, and human-AI interaction.
Closing
The point of all of this is to get beyond speculation by building in the world.
We simply do not have enough runtime with LLMs and generative AI to rely on decades of proven patterns, reference models, or technical conventions. At the same time, the technology keeps moving, and the cycle of obsolescence is fast.
A working prototype, even a rough one, creates something that can be experienced, questioned, tested, and refined. It reveals assumptions, requirements, and limitations that are difficult to uncover in a strategy document or an abstract discussion.