AI Consulting · Series A–C Startups
AI consulting for startup marketing and GTM teams.
Most companies have bought the tools. Half the team is using them, badly, in isolation. Nothing is repeatable, nothing runs on its own, and the manual work everyone hates is still getting done by hand.
A Claude Project, custom GPT, or Gem gives your team instant access to your brand voice, research, and compliance-approved messaging, on demand. But it waits for you to open it and ask.
An agent executes a multi-step workflow on its own, on a schedule, and delivers a result you didn't have to ask for.
I build both, wired into the tools your team already lives in, with brand and compliance built in from day one.
Then I do the harder part, which is getting people to actually use them.
You're the entire marketing team, or close to it. AI is helping a little, but nothing remembers your positioning, and every prompt starts from zero.
Your team's already using AI, just all differently. The board keeps asking about your AI leverage, and you have a vibe instead of a number.
The AI-first mandate came from the top. Tools got bought, adoption stalled, and nobody underneath knows how to actually execute it.
Whatever stage you're at, the fix starts with a conversation, not a menu.
I built an agent last month out of pure frustration, while I was putting on my makeup. We were running a project with an outside agency, with dozens of moving parts split across owners on both sides, and I couldn't keep my arms around any of it.
By the time I sat down with my coffee, it was live for the whole team. Every morning it reads the shared Slack channel, reconciles that activity against our internal tracker, and sends a TLDR snapshot plus a detailed digest for each owner.
During its first week, it saved an estimated three hours of manual tracking. And nobody needed a status meeting to know where things stood.
That build took a morning. The part that takes real skill is knowing which problem to point it at, and getting a skeptical team to trust it once it exists.
Some work needs an agent that runs on a schedule, pulls its own data, and delivers a result nobody had to ask for. Some work needs a Claude Project, custom GPT, Gem, or skill your team reaches for on demand, trained on your voice and your sources.
I've built inside three regulated healthcare environments where one unapproved claim is a real risk. Guardrails go into the tool itself, so the compliant path is the default and your Legal team has far less to flag.
A brilliant tool nobody opens saves zero hours. I'm a Certified Professional Coach, and I use a belief-based change framework to move teams from resistance to building. Resistance is almost never about the tool but the belief underneath it.
Every system is tuned to preserve warmth, specificity, and real point of view. AI should make the human part of the work better, not flatten it.
Most AI rollouts fail for the same reason: leadership announces a mandate, buys some tools, and waits for people to change. They don't, because resistance comes from a conscious or subconscious belief, not a skill gap. "AI hallucinates" or "it'll replace me" isn't solved by a training deck.
I use a five-phase framework built on that idea:
Leaders move through their own AI beliefs before they're asked to create safety for anyone else. A fearful leader can't lead a shift they haven't made themselves.
Short, honest conversations surface the specific belief driving avoidance on each team, not a survey, a listening exercise.
Everyone gets one experience where AI visibly saves them time on their own dreaded task. That's what disconfirms the belief, not a slogan.
People move from generic prompting to the tools built specifically for their work, with friction fixed in real time.
The strongest sign adoption worked: people start proposing and shipping their own tools, without being asked.
This is the exact framework behind the 30+ people I've moved from AI skeptics to daily users, and the 15 I've taken all the way to building their own tools.
Every engagement starts with a call. That's where we figure out what's actually worth building, and whether I'm the right person for it.
Startup marketing and GTM teams who:
I've spent 18 years in marketing, more than 10 of it in B2B SaaS and healthtech, and the last three building AI systems inside high-trust, regulated environments.
I've created more than 30 custom GPTs, Gems, and Claude Projects, 10+ Claude Skills, and 8+ production agents deployed across 8+ functions, most of it with no formal mandate and no authority to require adoption. That meant every one of those numbers was earned, not required.
I'm also a Certified Professional Coach and an Energy Leadership Index Master Practitioner. That's the part that makes the difference. Anyone can hand a team a tool. Very few people can move someone from "AI is going to replace me" to shipping their own agent.
A custom GPT or Claude Project waits for you to open it and ask, which is exactly right for work you want to control. An agent runs on its own schedule, gathers its own data from your connected tools, and delivers a result without being prompted. Most teams only ever build the first kind, which is why nothing runs when nobody remembers to open it.
Engagements typically run as a monthly retainer of $5,000 to $12,000, with a three-month minimum, priced around the specific systems and agents your team needs. Some teams start with a single defined build, like a first production agent, to see how the working relationship fits before committing to a longer engagement. Every engagement starts with a call to scope the right fit.
Yes. Most of my Series A work is with one person wearing the marketing hat, drowning in manual work with no time to fix it. For teams that size, we often start with a single agent build or a foundation piece like brand and compliance guidelines, before scoping anything larger.
Using AI and running AI as a system are different things. If your team is using AI inconsistently, with no shared brand voice or way to measure impact, that's an adoption and architecture problem, not a tools problem. That's what I fix.
Start with an honest maturity benchmark, not a tool purchase. I begin by finding out where your teams actually are, then build a roadmap and the belief-based adoption work to get people there for real, not just on paper.
A first agent build is usually live within two to three weeks, and hours saved are often measurable in the first week it runs. Larger engagements are scoped in phases so something is live and useful early, rather than waiting for a full system to finish before anything ships.
Claude, including Claude Projects, Skills, and scheduled routines, plus ChatGPT Agent Builder, custom GPTs, Gems, and NotebookLM. I build in the environment your team already uses rather than asking you to adopt a new one.
Yes. Agents are delivered into your environment, in repositories you control, with documentation. Nothing depends on my continued involvement.
Guardrails go into the tool itself. I've built claims-safe and regulatory constraints directly into AI systems in a regulated healthcare environment, so the compliant output is the default rather than something a human has to catch at the end.
That's the part I'm best at, and it's the reason most AI rollouts fail. I use a belief-based change framework grounded in coaching training. I've moved 30+ people from skeptic to active user, and taken 15 of them, including leaders who were openly resistant, all the way to building their own tools.
Continued support is available after a retainer ends, and we talk about it once you know what you actually need rather than committing to it upfront.
Tell me what's taking over your week. We'll find the calm on the other side of it.