Maju Kuruvilla
20 March 2026
I have seen this moment before.
In healthcare, I was the founding engineer at a company that automated payment decisions between hospitals and insurers. Completely manual, impossibly complex. We built the decision engines that replaced it. I spent ten years as CTO. We sold the company to Hearst.
At Amazon, as VP of Technology for worldwide fulfillment, I was there right when everything got harder at once. Same-day fulfillment, multiple new business lines, global expansion, robotics rolling into warehouses alongside human workers. I helped build the AI systems that powered real-time decision-making inside Amazon's fulfillment operations.
Both times, the pattern was the same. A complex industry running on manual processes. New technology arriving that made those processes both more inadequate and newly replaceable. A window opening. And a window that would not stay open forever.
I see that window in commerce right now.
The store is the bottleneck
Every ecommerce store in the world still requires a human to decide what happens next. The tools have gotten better. The stack has gotten bigger. But someone still decides what goes on the homepage, what the email says, which test to run, how to read the results. The tools execute. The decisions are still manual.
For thirty years, that was fine. But traffic does not arrive empty anymore. Shoppers land shaped by a TikTok video, a ChatGPT recommendation, a Perplexity answer. They arrive with intent already formed. And increasingly, they do not arrive at all. An AI agent does it for them.
The store treats all of this the same. Same page, same layout, same experience for everyone. The manually operated store cannot keep up. Not because the teams are not talented. Because no human team can process the volume of signals, the variety of visitors, and the speed of change fast enough. The answer is not a better process. It is a different model entirely.
The blank page
Here is what we built.
A brand hands us a blank page on their website. They send traffic to it. Their metrics improve. No templates. No rules. No configuration.
Spangle resolves purchase intent before the page renders. When a shopper arrives, the system reads the context of that visit and builds the right experience for it. Two visitors never see the same page. When a buyer agent arrives, sent by Instagram, Google, ChatGPT or Perplexity, it meets that agent in its own language and converts it.
Brands we work with have seen 50% increase in revenue per visit — within weeks, not months.
What we discovered
Something unexpected happened as we built this.
We started seeing the full chain. Not just what gets clicked or opened, but what actually gets bought. Which products, which messages, which experiences, for which visitors, from which sources. Most companies have this data scattered across a dozen tools. Nobody connects it end to end.
Our customers noticed. They started pulling us upstream. Can you write our emails the way you build our landing pages, shaped by what actually converts? Can you make sure our products show up correctly when an AI agent searches on someone's behalf? Can you feed what you know back into how we acquire customers?
That is where we are heading. A closed loop where every surface, the storefront, AI search, paid media, email, is informed by the same intelligence. What converts flows back into everything that comes before it. The brand gets smarter with every transaction, not just every session.
Why us
My co-founder Fei Wang spent 12 years at Amazon as a Principal Engineer. He was on the founding team for Alexa and built Amazon's customer service chatbot. Both products were built around the same architecture: listen to every input, understand intent, execute the right response. He has built the brain before. Twice.
Then he became CTO of Saks OFF 5TH. For three years he experienced the exact opposite. Dozens of disconnected vendors. Fragmented data. No unified intelligence layer. Every decision required manual coordination across systems that did not talk to each other.
He knew exactly what was missing. Because he had built it before.
When we compared notes, we were looking at the same problem from different sides. I had seen the transformation pattern and knew the timing was right. Fei had built the AI architecture and knew what the solution looked like from the inside. Spangle is where those two halves met.
Where this goes
Traffic will increasingly be mediated by AI. Buyer agents will shop on behalf of humans. The brands that win will not be the ones with the best templates or the most tools. They will be the ones with an intelligence layer that connects every touchpoint from discovery to purchase and gets smarter with every transaction.
Today, that starts with the blank page. A brand sends traffic, and the store responds intelligently.
Tomorrow, it runs commerce end to end.
More revenue from every visit.
If this resonates, I'd love to talk. Book a demo.
