Caravan x OHD: Industrial Safety Manufacturer

TLDR:

• OHD is a leader in respirator fit testing. OHD instruments (QuantiFit®, QuantiFit2®, Aerofit™) perform OSHA/ISO-required testing to ensure the proper fit of respirators for workers required to wear masks to perform their duties safely.
• OHD’s instruments are used by safety teams across manufacturing, mining, defense, law enforcement, and hospital systems in 30+ countries.
• In April, their leadership brought their top ten people together for a full day of on-site AI training with Caravan.
• It was a resounding success.

Why OHD didn’t train themselves

OHD could have figured this out internally. A couple of their technical people could have sat down with Claude, worked through tutorials, and run a workshop for the rest of the team a month later.

They didn’t.

OHD has built decades of advantage by being the best in the world at a narrow, hard thing.

They don’t try to be the best at everything. They build what’s core, and they bring in specialists for what isn’t.

Their leadership put it simply: a team assigned to “figure AI out on the side” lands somewhere useful in six to nine months.

A trainer who has done this across dozens of industries has already hit every wall the internal team is going to hit. OHD was not willing to burn six months on a foundation they could get in a day.

And their standards wouldn’t let them learn it badly.

They don’t compromise on the instruments they sell to customers around the world seeking to protect their teams. They weren’t going to compromise their standards on how their leaders learned to use the tool that’s going to shape everything those teams do.

Who did we train?

The working surface of the company:

• Leadership
• Operations
• Finance
• Sales
• Marketing
• Technical support
• Business systems

That kind of room is the hardest to teach well. The shortcuts that land for a sales floor don’t land for a service engineer. The framing that works for finance doesn’t work for operations.

How Caravan prepared

Before any private training, we do the homework.

Every attendee fills out a pre-training survey. We know who’s been using AI for a year and who has never opened Claude. We know what systems they work in, what their day looks like, and what they want to walk out with.

We build a context file — spending real time with the company’s products, customer feedback, and workflows, then packaging it so the AI already understands who OHD is before anyone opens a chat.

Nobody has to explain their job from scratch.

And we design for two layers of goals.

1. Every participant has something they personally want.
2. The organization has something it wants too.

What did we teach?

The training was hands-on from the start.

We began with AI capability — what the model can actually do, and why bad output is almost always an operator problem, not a model problem.

Then context: how to give AI memory of who you are, what your company does, and what you’re working on — so nobody has to explain their job from scratch every time they open a chat.

Then tools: connecting AI to the systems OHD actually works in — their CRM, documents, spreadsheets, internal workflows.

Then instructions: building reusable skills, agents, and workflows that turn one-time prompts into durable automation.

By the afternoon, the room had moved from chatting with AI to delegating real work inside their real systems.

The best part of a day like this is watching people surprise themselves. We push someone to try something further than they think they can go. They hit a wall. We help them through it. They come out the other side realizing they can do things they didn’t believe were for them a day earlier.

What they came away with

AI connected to the systems they actually work in — their CRM, internal tools, real workflows.

A shared understanding of which connectors to enable, which to restrict, and why.

A team that started the morning as a mix of power users and beginners and left as ten people who could delegate real work to AI inside their own departments.

A shared habit of preparing context before handing off a task — the habit that compounds most over time.

The real returns come in the weeks and months after.

“We wanted to transition from individual, siloed use of AI chat tools to a more robust, company-wide use of the tools with a goal of collectively growing our business more efficiently. The Caravan team understood our business needs, tailored the training to our audience, and unlocked the use of the Claude toolset that has significantly enhanced how every department interacts with the work we do every day. The automations, insights, and efficiencies we’ve already gained by using Claude throughout our organization are enormous. The Caravan training was just what we needed, and it has allowed us to move forward into this new world of AI with confidence.”

— Hunter Strickler, Chief Operating Officer, OHD