Minimal AI Resolves 90 Percent of E-Commerce Support Tickets, and the Math Actually Works

The Macro: AI Customer Support Is Everywhere, but Most of It Is Still Bad

I have tested at least a dozen AI customer support tools this year. Most of them fall into one of two categories. Category one is the chatbot that can answer FAQ-level questions and escalates everything else to a human. Category two is the enterprise platform that takes six weeks to implement and requires a dedicated team to maintain. Neither category is solving the actual problem, which is that e-commerce brands need support that scales without proportional headcount growth and without making customers want to throw their laptops out the window.

The market is enormous. Global customer experience management is projected to reach $32 billion by 2029. E-commerce specifically is where the pain is sharpest because ticket volume correlates directly with order volume, and order volume spikes are unpredictable. Black Friday, a viral TikTok, a product recall. The brands that win are the ones that can handle 10x ticket volume without 10x support staff.

Zendesk and Freshdesk dominate the helpdesk market but their AI features are add-ons to systems designed around human agents. Gorgias built specifically for e-commerce and has done well, but their AI automation still requires significant setup and tends to handle simpler queries. Intercom has Fin, which is solid but expensive and not e-commerce-specific. The pure-play AI support startups, Ada, Forethought, Siena, are all competing for the same enterprise slice of the market. What is underserved is the mid-market e-commerce brand. The Shopify store doing $5 million to $50 million in revenue that cannot afford a six-figure annual support platform but needs more than a basic chatbot.

The unlock for this segment is not better AI. The AI is already good enough. The unlock is implementation simplicity. If a store owner can train an AI agent in an afternoon without writing code or hiring a consultant, the adoption barrier drops dramatically.

The Micro: A Suit Entrepreneur and an Uber Data Scientist Walk Into a Helpdesk

Minimal AI builds AI agents for e-commerce customer support that resolve over 90 percent of tickets autonomously. The product works across email, chat, WhatsApp, social media, and review management. It integrates with over 60 platforms including Zendesk, Freshdesk, Gorgias, Shopify, and Magento. The agents do not just answer questions. They execute real actions. Cancel orders, process refunds, track shipments, update customer records. That action capability is what separates a useful AI agent from a fancy FAQ bot.

Niek Hogenboom is the CEO and co-founder. He started a business selling custom suits at 16, grew a company called SitRight to over $2 million in revenue during COVID, and holds a degree in Aerospace Engineering from Delft University with master’s research at Oxford. Titus Ex is the CTO and co-founder. He was previously a data scientist and ML engineer at Uber and at Soda, a Point Nine-backed startup, with a master’s in Applied Math. He has been working in AI since 2018. The team is three people, based between San Francisco and Amsterdam, part of Y Combinator’s Summer 2025 batch.

The differentiator that keeps coming up in my evaluation is the AI Manager interface. Instead of configuring rules and decision trees, store owners train their AI agents using natural language. You tell the agent “if a customer asks about a delayed order, check the tracking number first, then offer a 10 percent discount if it is more than 5 days late.” The agent updates its behavior instantly. You can watch the agent’s reasoning process in real time, see why it made a specific decision, and correct it on the spot. That transparency is rare. Most AI support tools are black boxes. You deploy them and hope they say the right things.

The traction numbers are strong. Tens of thousands of tickets resolved per day across their customer base. Case studies show Proforto at 85 percent automation, Boombrush at 88 percent, and Cloudpillo at over 65 percent. Sub-two-minute response times. 93 percent customer satisfaction scores. 80 percent lower cost per ticket. Those are not theoretical projections. Those are reported numbers from live deployments.

The 60-plus integrations matter because e-commerce support is not just about the helpdesk. The agent needs access to the order management system, the shipping provider, the payment processor, and the CRM. Minimal connects to all of them, which means the agent can actually do things instead of just saying “let me transfer you to someone who can help.” That handoff is where most AI support tools fail and where customers lose patience.

Support across 95 languages is table stakes for international e-commerce but still worth noting. A Dutch brand selling to Japanese customers needs support in Japanese. Training a human team for that is expensive. Training an AI agent for that is a configuration toggle.

The Verdict

I think Minimal AI is one of the strongest products in AI customer support specifically because of the implementation model. The natural language training, the transparent reasoning, and the action execution layer are the three things that most competitors get wrong. If I ran a Shopify store, this is the product I would try first.

The risk is the 90 percent number. That works for order tracking and refund requests. It works less well for angry customers, complex warranty claims, and situations that require genuine empathy rather than procedural accuracy. The remaining 10 percent of tickets is where brand reputation lives, and how gracefully Minimal handles the escalation to human agents matters as much as the automation rate.

At 30 days, I would want to see the edge cases. What happens when the AI gets it wrong? How fast does the feedback loop correct the behavior? At 60 days, the question is churn. E-commerce brands are notoriously price-sensitive and will switch tools quickly if results dip. At 90 days, I would want to know whether Minimal is expanding beyond support into proactive customer engagement, because the same AI that handles complaints could also drive upsells, review requests, and retention campaigns. The data it collects on customer behavior is valuable beyond the support ticket.

Niek’s background in building and scaling a direct-to-consumer business means he knows the customer he is selling to. That founder-market fit, combined with strong early traction and a genuinely differentiated product, makes Minimal AI one of the more compelling plays in a crowded category. Three people, tens of thousands of tickets per day. The unit economics speak for themselves.