Speaking & Media
Talks that come from building.
I speak about enterprise AI from the position of someone who has shipped it — production systems, memory infrastructure, and the operating model that turns experiments into a durable capability. Keynotes, executive briefings, and panels.
Signature keynote
The Enterprise Was Built for a World Where Building Was Expensive
AI is not simply another technology layer to insert into the existing enterprise operating model. It is changing the economics of experimentation, software creation, knowledge retrieval, coordination, and execution all at once.
Most enterprise operating structures were designed around a single assumption: that building was expensive. Software took months, projects took large teams, and failed experiments were costly, so organizations built steering committees, long roadmaps, and consensus cycles to protect themselves from the cost of being wrong. Drop AI into that model and you get AI committees governing ideas that could be prototyped and tested in days.
The better question is rarely “do we collectively believe this will work?” It is “what is the cheapest and fastest responsible way to find out?” Organizations that keep operating structures built for expensive technology will govern away much of AI’s value. This talk is about the operating model that does not.
What the room leaves with
- Decompose problems intelligently, instead of enlarging them by involving everyone at once.
- Put the right expertise at the right point in the problem.
- Experiment inexpensively, now that building is cheap.
- Govern the risk that is actually there, not the experimentation that used to be expensive.
- Build enough to learn, measure what happens, and preserve what the organization learns.
- Compound that intelligence into the next system, and give authority proportional to accountability.
Other topics
What I speak about.
Each of these is a talk in its own right, and each can be shaped for a board, a leadership team, or a technical audience.
You rent the model. You own the memory.
Frontier models will be rented, routed, and commoditized. The durable enterprise advantage is the context an organization accumulates — its decisions, corrections, and institutional memory. Why memory, not the model, is the asset that compounds.
Govern the risk, not the experiment.
Good AI governance draws hard boundaries around data, security, legal exposure, brand, and production access. Inside those boundaries, teams should be free to build and learn. Guardrails, not permission slips — and why the difference decides how much of AI’s value survives.
Authority should follow knowledge.
Hierarchy sets decision rights. It does not establish who understands the problem best. How to decompose the work, put authority where the knowledge is, and hold someone accountable for the whole — without descending into organizational anarchy.
If you can imagine it, you can increasingly build it.
The technical ceiling has fallen. The advantage moves to identifying the right problem, designing the architecture, and building fast enough to learn from real use. Why deep understanding now comes from building, not from reading the research.
Speaker bio
Jeff Ellis
Enterprise AI strategist, systems builder, and operator.
I am an enterprise AI strategist, systems builder, entrepreneur, and business operator. My career has consistently lived at the intersection of technology, business strategy, and execution — and I am an executive who builds, not a builder trying to become an executive.
I am a degreed engineer by training. I began in technical recruiting and sales engineering, then founded a digital agency that I led for fourteen years and sold in 2017. I later held senior digital leadership roles, including Global Director of Digital Marketing for a top global law firm, before moving deeper into enterprise operations and artificial intelligence inside a Fortune 100 enterprise.
Today I focus on a question I believe will define the next generation of enterprise AI: how do organizations move beyond isolated AI tools and experiments to systems that can actually remember, learn, coordinate, and increasingly act? My perspective has been shaped as much by building as by strategy. I have designed and built production AI applications spanning account-based marketing, content, analytics, data integrity, workflow automation, and organizational intelligence.
That work led me to a simple conviction: the model itself will not be the enduring advantage. Models can be rented, replaced, and commoditized. The more durable enterprise asset is the context an organization accumulates — its decisions, corrections, operating knowledge, customer understanding, and institutional memory. That conviction underlies my independent work on Lyrn, a vision for persistent organizational intelligence and the foundation for increasingly capable, governed enterprise AI systems.
My view of AI is ultimately less about technology than organizational design. The companies that win will not be the ones that buy the most AI. They will be the ones that learn to combine human expertise, institutional memory, intelligent systems, rapid experimentation, governance, and accountability into a fundamentally different way of operating.
Speaking inquiries
Available for keynotes, executive briefings, and panel discussions on enterprise AI strategy, memory infrastructure, and organizational intelligence.
Get in touch