Most AI work stops at the slide deck.
You have probably already sat through the assessment. Maybe you bought licenses. Somebody on your team is using ChatGPT for email. And the actual work, the estimating, the submittals, the proposals, the reporting, looks exactly like it did eighteen months ago.
The gap is not strategy. It is that almost nobody in this business actually builds anything.
I do. I sit with your team, find the thing that is genuinely slow or expensive, and build the fix with them. The education happens on your real work instead of in a training session nobody remembers.
The 90-Day AI Sprint
We sit down with your leadership team, rank where AI creates the most impact against your actual goals, and pick the first workflow together. Ninety days later something is live in the tools you already own, someone on your team runs it, and we measure the result against a baseline we captured before building anything.
Prioritize
Leadership working session and discovery interviews. We build the ranked list of pain points and put a number on each one, hours or dollars. Then we baseline the workflow we are going to fix.
Build
Weekly build cadence, inside your existing tools and permissions. Your team is in the room while it gets made. That is not a courtesy, it is how the knowledge transfers.
Prove
Test, roll out, train, and measure against the baseline. Results reviewed with leadership.
- A ranked AI roadmap for the whole business
- One workflow live in production
- A trained owner on your team, not a dependency on me
- Before and after numbers
The cycle repeats each quarter with the next item on the list, so capability compounds instead of stalling after a pilot.
Two ways in. Both end in something built.
Most engagements start with one and pull in the other. Both are how the sprint gets delivered.
Working software inside the tools you already use.
Custom internal tools, workflow automation, role-specific assistants, dashboards and knowledge systems. Built with your team, in your environment, on your real data.
- Workflow automation
- Internal assistants and agents
- Dashboards, portals, knowledge systems
- Custom application development
- Integration with the systems you already run
- Ongoing support and refinement
Find the right problem before building anything.
Leadership working sessions that surface the real friction, put numbers on it, rank it against your goals, and name an internal owner for each item. You get a ranked plan, not a wish list.
- Leadership working sessions
- Workflow discovery and baselining
- Use case identification and ranking
- Roadmap development
- Team enablement and adoption
- Governance where it is actually needed
Prioritizing tells you what matters. Building is the part most people skip.
The smallest team creates the largest impact.
Owners and leaders are the smallest group in the business and their decisions shape how everyone else works. When leadership builds real confidence with AI, adoption moves faster and the value compounds down through every tier.
This is also why I do not run generic AI training. Leaders do not learn this from a slide on prompting. They learn it when we take a real problem off their desk and build it together while they watch.
What this looks like in practice.
The patterns that come up most often across the teams I work with.
Best fit for teams with real workflow complexity.
Coordination overhead, workflow friction, and a real need for something to get built.
AEC is where I go deepest. The broader fit is any team with genuine workflow complexity and a real need for something to get built.
Brennan Gerle
I build. That is the short version.
I spent fifteen years in video production and creative leadership, including seven years running teams at an agency, before founding POLR AI. What I do now is sit with leadership teams, find where the work actually breaks down, and build the thing that fixes it.
I work in Claude, Cursor, Next.js, TypeScript, Supabase, and n8n. I build custom applications, agents, and internal tools, and I am usually building something the same week we talk about it. That matters because the gap between “here is what AI could do” and “here is the thing running on your data” is where most engagements die.
- Member of the Claude Partner Network
- Board member, AZIMA
- Speaks on AI for leadership teams
Questions worth answering first.
Let's find the thing worth building.
If your team is trying to figure out where AI fits, or you already know and need someone to actually build it, let's talk.