516 Days and Counting on the AI Startup Journey
516 Days and Counting: The Rollercoaster Journey of a Customer Support AI Startup

516 Days and Counting on the AI Startup Journey

Yuma.ai is my 3rd startup, but it all started by accident.

In December 2022, one day before the release of ChatGPT, I launched a prototype to bring AI to ecommerce customer support. Immediately, my inbox was flooded with Shopify merchants wanting to try it out. That's when I knew I had to jump back in.

Fast forward to today, the journey has been hectic. It's not my first rodeo, so the scares of the ups and downs don't affect me much anymore, but still, what a rollercoaster. There was a startup world before AI, and there is one after. Building with AI is a whole new ballgame, everything moving at breakneck speed in a gigantic race to stay ahead. Plus, adding the non-deterministic outcomes of LLMs and you are in for quite an intense ride.

Nevertheless, the task at hand can be daunting, especially when we consider our goal: fully automating customer support for ecommerce through the deployment of autonomous AI agents. Nobody has done that before, but we are up for the challenge; even though we know the road will be long.

One particular aspect I'm neurotic about is our speed of iteration. There are days when I feel exhausted and unhappy because I believe our rate of iteration is not where it should be. Speed is a major advantage for a startup. It's actually absolutely key; especially in the AI age. We need to move faster than anyone else if we want to stay ahead of the pack. This particular high velocity mindset is critical to me and something I emphasize a lot - just ask the team :)

So yes, it’s frustrating at times, but damn, when I look back and can enumerate what we all accomplished as a team, I can actually be quite proud of all of us at Yuma.

A year ago, we served merely as co-pilots for merchants, not having deployed a single autonomous AI agent. Now, we run them at scale, and our largest merchants had over 150,000 support tickets in December alone last year. Our top merchants can automate more than 50% of their support tickets through us. And this is through True automation, not simple Q&A. Our AI agents can take actions and save hours for each support team using Yuma. Every single day.

Today, over a hundred merchants trust us with their customer experience (CX) AI automation needs.

Broken

From the inside, startups often feel incredibly broken. Everything is crumbling, and there are just gigantic lists of to-dos everywhere. Through sheer will and harsh prioritization, we make progress, bit by bit. We go from truly broken and useless to useful and a bit less broken.

The broken part is crucial. If it wasn't somewhat broken, it wouldn't be fun, as it's part of the ride. Plus, every time we fix something, we make our customer happier, which is the best part.

Through that iterative process, we have gone through a product that I was initially ashamed of—due to its deficiencies glaring at me non-stop—into something I'm truly proud of. Of course, we have always been delivering value throughout, but damn, it is painful to look at your product and see all the tiny broken things, even if they are invisible to most…

Over a year later, I'm thrilled to do demo nowadays to showcase what Yuma can offer to potential merchants. We've built a product that is at the cutting edge of AI, allowing brands to automate at their own pace in a safe and controllable way due to our significant emphasis on safety.

But let's not fool ourselves; the road ahead is long. However, as an entrepreneur, taking a few minutes to reflect and look back is critical. The future is bright, and Yuma AI (YC W23) will be the first product to truly automate customer support, effectively bridging the communication and engagement gap between brands and their customers. Could 2025 be the year of full automation?


Final thought. While I'm on it, do you think the best way to build a truly exceptional product is to continually raise the bar and always find ways to be somewhat ashamed of it? Perpetual dissatisfaction is most likely one of the keys to greatness.

I'm not a pessimist, though. You can't be to start such a journey.

Oleksandra Ivanenko

Head of Operations

10 个月

Guillaume, thanks for sharing!

回复
Leonidas Papadopoulos

Founder & CEO at Viable | Scaling Startups into Global Ventures | Venture Builder & Investor | Forbes 30 Under 30

10 个月

Intriguing ride. AI innovation meets entrepreneurial grit - inspiring stuff. Guillaume Luccisano

Cyril Vial

Head of Account Management at Yuma AI (YC W23)

10 个月

Starting Yuma by accident sounds like a nice accident to have ?? ; as for product excellence, the more we keep our standards high about the product, the more we'll find ways to improve it and the more we'll improve the experience for clients; so yes!

Damien LEGAVE

Vice President of International Sales & Marketing - Hinacom Software and Technology, LTD.

10 个月

awesome read. Keep going ! I often think of Yuma when I watch latest news regarding new models and AI tech ... must be a crazy ride !

要查看或添加评论,请登录

Guillaume Luccisano的更多文章

社区洞察

其他会员也浏览了