Sona Claims 80% Ticket Resolution and I Want to Believe It

The Macro: AI Customer Support Is the Most Overcrowded Category in SaaS

I need to be honest about something before we get into Sona: the AI customer support space is absolutely packed. It might be the single most competitive category in B2B SaaS right now. Intercom has Fin. Zendesk has AI agents. Freshdesk has Freddy. Ada, Forethought, Tidio, Kustomer, Siena AI, and at least two dozen other startups are all fighting for the same budget line item. If you are building in this space, you better have a very clear answer to the question “why you and not the other 30 options.”

The market itself is real and growing. The global customer experience AI market was valued at roughly $10 billion in 2024 and is expected to triple by 2030. E-commerce specifically drives a disproportionate share of support volume because the transactions are high-frequency, the questions are repetitive, and the margin pressure makes human agents expensive relative to ticket value. A $40 order generating a $12 support interaction is bad math. That is the math Sona wants to fix.

The 80% resolution rate claim is the number that jumps out. For context, most AI support tools in production today resolve somewhere between 30% and 60% of tickets without human intervention, depending on how you define “resolution.” The gap between 60% and 80% is enormous in practice. It is the difference between AI handling the easy stuff and AI handling the easy stuff plus the medium stuff. That second category includes order modifications, return processing, shipping disputes, and product questions that require looking up specific SKU data. Automating those interactions is meaningfully harder than answering “where is my order.”

The Micro: Built for E-commerce, Not for Everyone

Sona positions itself as AI customer support built specifically for e-commerce. That vertical focus is the strategic move that makes the 80% number at least plausible. When you narrow your scope to one industry, you can train your models on domain-specific data, build integrations with the platforms that matter (Shopify, WooCommerce, BigCommerce), and optimize for the ticket types that actually show up in e-commerce support queues.

Compare that to horizontal support tools that need to handle everything from software bug reports to insurance claims to restaurant complaints. Breadth is the enemy of depth in AI support. Intercom’s Fin needs to work for a SaaS company and a DTC brand and a B2B services firm. Sona only needs to work for online stores. That constraint is a feature.

The product handles the standard e-commerce support workflow: order tracking, returns and exchanges, product questions, shipping issues, discount code problems, and the endless stream of “I ordered the wrong size” tickets that every apparel brand drowns in. What makes or breaks a tool like this is not whether it can handle those categories in theory but whether it can handle them accurately enough that customers do not notice they are talking to a bot. Or more precisely, whether customers notice and do not care because the bot solved their problem in 30 seconds instead of the 4-hour wait for a human.

The e-commerce AI support category has some serious players. Siena AI is probably the closest direct competitor, also focused on e-commerce with a similar resolution rate pitch. Gorgias has deep Shopify integration and has been building AI features aggressively. Richpanel targets DTC brands specifically. Sona needs to either outperform these products on resolution quality, undercut them on price, or win on integration depth. Ideally all three.

I could not find detailed founder information or a YC batch listing for Sona through public sources, which makes it harder to evaluate the team. The website was not loading during my research, though the domain appears active. That is not ideal for a product whose entire value proposition is being available when customers need help.

The multi-language support claim is worth watching. E-commerce is inherently global. A Shopify brand selling to customers in 40 countries needs support in at least a dozen languages. If Sona handles multilingual tickets natively without requiring separate configurations per language, that is a meaningful differentiator against older support tools that bolt on translation as an afterthought.

The Verdict

Sona is entering a knife fight with a clear thesis: go deep on e-commerce instead of wide across industries. I think that thesis is sound. The question is execution.

At 30 days, I want to see the resolution rate validated by actual customers, not just marketing copy. 80% is a strong claim. If it holds up in production with real ticket volume and real customer satisfaction scores, Sona has something. If it quietly drops to 55% when you look at the data, it is just another chatbot with good PR.

At 60 days, the integration story matters. How deep does the Shopify integration go? Can it actually process returns, issue refunds, and modify orders without a human approving each action? The gap between “AI drafts a response” and “AI resolves the ticket end to end” is where most support tools stall out.

At 90 days, I want to see retention numbers. E-commerce brands are famously fickle with their tech stack. They will try anything, and they will drop anything that does not show ROI within a quarter. If Sona can hold customers past the honeymoon period, the vertical focus is working. If churn is high, the competitive moat is not deep enough.

The market opportunity is real. E-commerce support volume is not going down. But this is a category where being good is not enough. You have to be noticeably better than Intercom, Zendesk, Siena, and Gorgias, all of whom have bigger teams, more money, and existing customer relationships. Sona needs its 80% number to be real, provable, and consistent. Nothing less will work.