Good, Better, Best
Never let it rest! Until your good gets better and your better gets best!
The Bears have been on a tear this season. Ben Johnson has Chicago fired up behind “Good, Better, Best!” We’ve brought that phrase into Cargado’s culture — half our company is in Chicago and we’re all riding this run, even with a heartbreaking loss last night to the 49ers.
Everyone is posting their 2025 Year in Review. Even ChatGPT gives you a Wrapped summary now. A year-in-review for a startup in Year 2 feels weird. So I’m skipping it.
Let me share what’s changed over the year and what we’ve learned building Cargado.
Technology is dramatically different than it was six months ago.
AI has completely transformed the world, whether everyone realizes it yet or not. It can 10x, maybe even 100x, the strongest performers if they lean into it. Whether it’s improving your writing, accelerating content creation, writing code, or building prototypes in apps like Lovable - we’re in a different world now.
I got early access to Notion’s new Agents about a week ago. I’ll admit — I hated Notion until a few months ago when the structure finally unlocked for me. The way Notion AI and Agents can access centralized knowledge across our entire company is a game-changer. Ivan Zhao recently wrote about how Notion has 1,000 employees and 700 agents deployed. I’ve built 15 agents in the past week. It’s like having a bunch of really smart junior employees who know everything across Notion, Slack, and email.
The way companies are built is evolving right now. How Cargado will look in 2026 will be different than how it looked for most of 2025. How we execute internally will continue to evolve as our team learns to use Claude and ChatGPT, as we build workflow automation that lets our business scale to the impact of a 250-person org with less than 40 people.
Company building will be done with people who can harness this technology and use it to do several jobs at once. A marketing department should never need 15-20 people, regardless of company size or funding. A great marketing function includes the right system architecture, incredible prompts, and a strong understanding of how powerful AI tools can be.
Our sales motion is scaling through technology. We have three incredible sellers right now. We may add a couple to cover new markets, but I don’t feel like we need a sales org with tons of SDRs and regional sellers. You need an incredible outbound engine built with the right infrastructure to rapidly scale how you reach prospects and the stories you tell them. This all ties together — marketing and a modern GTM engine.
Our pricing model changed dramatically throughout the year.
We started with simple seat-based pricing. That’s how other products in the market are sold today. While it made sense initially because people could equate a dollar amount to individual access, it was clear seat-based pricing wouldn’t scale. The ideal scenario is paying for what you use. If six people use the same amount as one power user, you should be charged fairly.
We shifted to usage-based pricing. Customers got unlimited seats but paid for marketplace postings and rate lookups. While this seemed more logical and fair, unlocking the seat limitations, we heard customers talk about holding back. They were posting only their most critical freight to avoid using up their postings. They were experimenting with Mexico business and weren’t ready to go all in on a marketplace when every input wasn’t guaranteed to result in the outcome they wanted. Tools like Claude are consumption-based because every time you ask it to do something, it does exactly that. It’s a utility. A marketplace is a path for coverage, but it’s not guaranteed yet. People tried to balance value with cost.
This feedback led to our latest pricing iteration. We still use these usage amounts we modeled out, but we’ve moved customers to tiered unlimited plans. They can post and quote as much as they want across Mexico and Canada. They can add everyone in their company. Based on their usage over the next year, they’ll land in a new pricing tier that’s fair for both sides. I’ve never loved this model because it’s less formulaic for the end user, but it gives business leaders a clear budget number rather than worrying about overage.
Our product has evolved, a ton.
We started the year as a load board for cross-border Mexico freight. We’ll end 2025 as the only modern marketplace built for Mexico and Canada freight, with a robust market rates tool built on all the data generated by our marketplace activity. We have a true pulse on the market based on carrier bids we collect. Our product enables brokers to understand the true range of cost for a lane. Not just what someone spent weeks or months ago — what all the carriers are asking for right now. A true pulse of the market.
When we started building Cargado, the first thing our engineering team did was map out how our platform would be architected. We mapped out how we would store data and why it was important to structure it. Everything you could want to know about pricing or moving a load lives in a structured format, just like most systems brokers use across the industry. It keeps things clean in a messy industry. Historically, most systems have only housed the most basic data points. Origin, destination, buyer, seller, a block of notes people tend to fill with the same garbage instructions everyone already knows (like trailers being clean).
There’s a craze right now in logistics to apply AI agents — especially voice agents — to solving all of freight’s problems. A lot of capital has been raised across startups working to bring AI to freight. They’re taking different approaches, but they all sit at a similar cross-section, applying AI as a bandaid over old systems. Whether it’s supporting a flood of inbound calls from legacy load boards, cold calling carriers off lists to book freight, or crawling legacy TMSs in search of data and making life easier for freight operators.
But there’s a battle ensuing. Does this industry need more bandaids on old systems or do we need new systems? Mexico freight has long existed mainly in spreadsheets and WhatsApp, with a side of LinkedIn for coverage support. Rather than building tools that sit on top of those systems, we’re building tools to replace them. Brokerage operators should not be living in Excel and WhatsApp. They deserve a system built for their needs, structured to understand how freight moves. It’s not just about filling up Notion pages with context so agents know what to do in different scenarios. The ability to flex between structured and unstructured data is what will differentiate the true difference-makers in this industry.
2026 is going to be a big year for Cargado.
We’re working on our 2026 forecast right now. The plan currently calls for 3-5 hires. Almost all those hires are people who can build and operate without needing a lot of guidance. Building at a startup means BUILDING. You roll up your sleeves and start digging in until you solve a problem, then you move onto the next problem. We’ve already seen several people who came from long careers in freight now spending all day in Claude, ChatGPT, Notion, and VAPI (👀). I hope to continue converting more people inside Cargado into operators with agency — people who can just go and build without needing a ton of guidance.
We have a lot of customers asking for a lot from our product and our team. We only launched our first product in April 2024 and the whole world has changed since then. What we never thought was possible, or what we thought might take six weeks, now could take six hours or six days with the right technology supporting the work. Our company, team, and product will continue to evolve as the world evolves.
Good, Better, Best.
That’s the mindset we’ve brought into Cargado from the Bears’ season. We’re always pushing to improve. Good isn’t good enough when better is possible. Better isn’t the goal when best is within reach.
What’s different now is the timeline. The AI tools available today — Claude, ChatGPT, Notion Agents, voice agents — are compressing the distance between good and best. What used to take quarters now takes weeks. What used to require hiring five people now takes one person with the right prompts and infrastructure.
We’re not settling for incremental improvement. We’re using every tool available to accelerate our path to best. That’s how we’ll win in 2026.