Users Don't Want Two Tools – They Want One That Does Both

01 /The Concept

HubSpot backed Doola. Y Combinator incubated it. Four years later, the platform handles US company registration, bookkeeping, taxes, and financial tracking for founders worldwide – a quiet workhorse that just added something noisier: an "AI co-founder" for e-commerce companies, announced on the Y Combinator blog.

The stated goal is "the fastest path from US company formation to first sale." The functionality will almost certainly expand over time – this is a v1.

Right now, the AI co-founder handles four categories: e-commerce operations ("How do I connect Shopify payments to my bank account?"), company formation rules across states ("What does it cost to file an annual report in Delaware vs. Wyoming?"), accounting ("Cash or accrual for my online store?"), and taxes ("What filings do I need as a foreign owner of a US company?").

The AI co-founder is embedded inside the Doola platform, not sold as a standalone product. Standard subscriptions include a fixed number of credits per question; once you burn through them, you can top up for additional cost.

02 /Why It Matters

Nothing about this sounds revolutionary at first glance. There are plenty of business management platforms, and plenty of standalone AI assistants that answer questions about accounting, taxes, and finance. What caught my attention is the combination – and here's why.

A few weeks ago, a partner at the venture firm a16z posted something that crystallized the shift:

"It's remarkable how AI has changed my expectations of ANY software. Why would I use a mood tracking app if I can't talk to it about the insights I might draw from my mood graph? Or use a calorie counting app that isn't also an AI coach who can weigh in on what I'm eating and suggest how to improve my diet and routine?"

The implication: a platform that just automates a routine task is no longer enough. Users now expect to be able to have a conversation with the platform about what's happening – to understand it better, to improve it, on either side of the equation.

At the same time, a standalone AI assistant that only answers questions falls short too. Before every question, you have to load it with context – explain your situation from scratch. Without context, the answers are generic. And crafting a thorough, well-structured prompt takes real effort.

Beyond that: why use two separate services for the same domain?

The ideal is a single system that works like a good employee or partner – handles its own responsibilities quietly, stays out of the way, but is ready to answer questions about the current state of affairs and suggest improvements when you need it.

Having a separate platform and a separate AI assistant is like having an in-house operator and an outside consultant who've never met. Everything moves slowly and through a game of telephone.

Doola's move is straightforward: keep the platform doing what platforms do, and add an AI partner that can answer substantive questions about the actual work happening on that platform.

The integration between platform and AI partner could obviously be tighter, and the functionality will grow. But the direction is right.

03 /Opportunities

The general opportunity here is building platforms, services, or apps that don't just automate a routine – but that you can have a real conversation with about that routine.

To be clear: this isn't about embedding a help-center bot that walks users through the interface. It's about an AI that can answer substantive questions about the substance of the work and surface actionable insights for improving it.

The beauty of this direction is that almost any existing service or app can be upgraded into this kind of AI partner – one that doesn't need to be briefed every time, because it already has all the relevant context.

So while most services and apps haven't made this leap yet, there's still a real window to pick your domain and ship an AI partner that's better than just a platform, and better than just a standalone chatbot.