Speaking & media

Four arguments, stated flat.

Kevin Owens speaks on what has to be true before AI agents are allowed to act inside a company: who owns the model version, who signs off when it changes, and what the operating model looks like once routine cognitive work stops being scarce. The talks below are the ones he is currently booking. Each one is an argument with a position, not a survey of the market.

Invite Kevin to speak Read the underlying writing

Currently booking a small number of events, remote or in person, in North America and Europe.

Talks

What he will come and say.

Each talk is adapted to the room rather than reused verbatim. The abstract is the commitment; the examples change with the audience and the sector.

Governed Execution

What has to exist before agents touch real work

Every vendor is launching agents. That is no longer the interesting part. The durable category is the layer underneath: identity, permissions, policy, evidence, reversibility, and review for work a machine performs on your behalf. This talk walks the seven questions a board will eventually ask about agent execution, shows why most organizations can answer none of them today, and lays out a maturity model for governed execution. Attendees leave able to name the owner of each control in their own company, or to name the gap.

Audience: CPOs, CIOs, CTOs, and PE portfolio operating teams.
Length: 30 minutes, or 45 with questions

Based on: The Next AI Category Is Not Agents. It Is the Control Plane.

Silent Degradation

Engineering discipline for production AI

The dangerous AI failure does not throw an error. The model is upgraded upstream, output quality shifts, no alert fires, and leadership acts on quietly wrong data for a quarter. This talk treats every model call as an unowned production dependency and gives five practices that put it back under control: trace, gate, eval, instrument, tolerate. It ends with the uncomfortable question most engineering leaders cannot answer on the spot, which is who signs off when the model changes.

Audience: Engineering leaders, platform teams, and the product leaders who share their on-call.
Length: 20 minutes, or a 60-minute internal session

Based on: Why Every Production AI System Needs a Blast Radius Plan.

The Golden Tablet

Turning operating knowledge into agent doctrine

Every company has an operator who knows the rules that were never written down. Agents cannot inherit tacit knowledge, which makes doctrine extraction, not model capability, the binding constraint on deployment. This talk gives the extraction method: shadow the best operator, capture the exceptions rather than the happy path, write escalation rules before automation rules, and version doctrine like code. It is the difference between agents that produce output and agents that produce operating force.

Audience: COOs, chiefs of staff, and product leaders deploying agents internally.
Length: 30 minutes, or a half-day working session

Draws on the VPOM operating method.

The Operating Model Built for a Constraint That Moved

What changes for product leadership when agents absorb routine work

Every CPO framework of the last decade assumes human attention is the binding constraint. Prioritization rations it, discovery directs it, roadmaps communicate it. Agents break that assumption. This talk walks what actually shifts. Prioritization moves from throughput control to signal quality, discovery compresses while synthesis stays human, roadmaps get cheaper to produce and harder to anchor. Then it walks what does not shift, which is the coordination work that still requires a person in the room.

Audience: CPOs, VPs of Product, and CEOs planning next year’s product organization.
Length: 30 minutes, or a panel contribution

Based on: The CPO Operating Model Has a Hidden Assumption.

Formats

Three shapes, and what each is good for.

The same argument lands differently on a stage, in a closed room, and in an hour of recorded conversation. Pick the one that matches what the audience is allowed to say out loud.

Keynote or conference session

Twenty to forty-five minutes on stage, built around one argument rather than a survey of the field. Slides are sparse and the position is stated flat. Best suited to product, engineering, and AI leadership conferences, and to portfolio or customer summits where the audience holds budget.

Closed-door executive session

Chatham House rules, ten to thirty leaders, half a day or less. The talk is the first thirty minutes. The rest is the room working its own problem against the framework. Runs well as a board offsite segment, a PE operating-partner day, or an internal leadership session for a single company.

Podcast and interview

Long-form conversation on AI governance, product operating models, and legacy modernization. No prepared talking points and no vendor pitch. Remote recording, one hour, with prep notes sent ahead so the host can steer toward what their audience actually argues about.

Speaker bio

Kevin Owens

Kevin Owens is a product and AI executive who has held Chief Product Officer roles at MotorK, Doodle, and FUEL CYCLE, and led product at Reveleer. He has run product organizations through platform relaunches, repositioning from consumer to B2B, and the first wave of AI capability shipped into regulated workflows. Through ProductExec he advises CEOs and product leaders who have more AI initiatives than owners.

His work centers on the layer most AI strategy skips, the control plane that decides what an agent is permitted to do, what evidence it leaves behind, and what happens when it is wrong. He writes about governed execution, blast radius, agent doctrine, and the product operating model that follows from all three.

A short bio, a high-resolution headshot, and an introduction slide are available on request for event programs.

Full background LinkedIn

Roles held

MotorKChief Product Officer
DoodleChief Product Officer
FUEL CYCLEChief Product Officer
ReveleerVP Product
Smart Biz AI HubCEO & Founder

What organizers get

Abstract, bio, and headshot within one business day of confirmation.
A pre-event call to tune examples to the audience and sector.
Slides delivered a week ahead, in the event template if one exists.
No vendor pitch and no product demo, at any format or length.

The first step

Ninety minutes on the decision you are actually stuck on.

Most advisory relationships start with a discovery call that discovers nothing. This one starts with work. Send the context beforehand — the roadmap, the org chart, the three AI initiatives nobody owns. We spend ninety minutes on the single decision blocking the others. You leave with a written point of view: what you should stop, what you should own, and what the next ninety days look like.

Format

90 minutes, one decision

Deliverable

Written point of view within 48 hours

Price

$1,500, credited against any engagement

Not a fit if you want a vendor evaluation, a staffing plan, or a deck to circulate. If ninety minutes shows this is not a fit, Kevin will say so and point you somewhere better.

Book the session

Booking a talk

Send the date, the room, and the argument you want made.

The useful first message names the audience, what they already believe, and the decision the event is meant to move. Format and length follow from that. Speaking is separate from advisory work; an event is not a sales meeting and is never treated as one.

Start a conversation Read the writing first