For ecommerce & DTC brands
Returns and support are quietly draining your margin.
Every refund, every "where's my order?", every exchange eats a person and an hour you don't have. Commonhouse sets up AI that drafts those replies from your own policies and waits for your yes before anything reaches a customer, and you own all of it.
A lean team can't out-hustle the ticket pile.
Post-purchase is where small brands bleed time. The questions repeat, the answers live in your head and a few docs, and the work arrives faster than two or three people can clear it. Most of it isn't a judgment call. It's the same handful of replies, written again.
Tickets never stop
Sizing, shipping, "did it ship?", the same questions every day across email, chat, and DMs. Each one pulls someone off growing the brand.
Returns cost twice
First the refund or replacement, then the time to decide, write, and process it. Slow replies turn a return into a chargeback or a bad review.
Follow-ups fall through
Order updates, delay notices, the gentle nudge on a stalled exchange, the things that keep a customer calm are exactly what slips when everyone is busy.
Here's how it actually handles support and returns.
Three steps. It learns your business, drafts the work you repeat, and waits for your yes before anything goes out.
It works from your stuff, not guesses
Your return policy, prices, voice, and the calls you've already made, written down once, so every reply is built from what's true.
Nothing sends without your yes
It writes the reply or resolution you'd normally type by hand, then stops. You approve, edit, or hold, it never sends on its own.
Everything's written down
Every draft, approval, and send is logged, so you can always see what happened, what went out, and who said yes. Exportable, and yours.
Return request → AI drafts → your yes → logged.
Here is the honest version, a single return, start to finish. The AI does the typing. You do exactly one thing: approve.
A return lands
A message arrives: "The jacket runs large, can I exchange it for a medium?" Captured where the AI can see it.
Drafted from your policy
It checks the order against your return policy and writes the resolution in your brand's voice.
You approve
It stops. Nothing has reached the customer. Read it, change a line, approve, or hold.
Sent & logged
It sends, starts the exchange, and writes the whole run to your action log.
A guided install, not a software drop.
Start where you are. You own every install from day one. Start small, see it run on one workflow, then widen.
AI Ops Audit
Best fit: you want to see where AI fits before you commit.
- Review of your tools and sources
- The work you repeat, mapped
- Your highest-leverage workflow
- Where AI is safe, where it needs a gate
- A prioritized action plan
- Yours to keep, no lock-in
Solo Install
Best fit: it is just you, and you are the bottleneck for everything.
- A client-owned AI home base
- One simple workflow, end to end
- A review step before anything sends
- An action log you can export
- Yours to keep, not a subscription
Foundation Install
Best fit: your knowledge is scattered and you want a solid base first.
- Company Brain
- Source Library
- AI home base set up
- Plan for first workflows
- Recommended first workflows
- Basic handoff
First Workflow Install
Best fit: you have one repeatable job that is costing you time, quality, or speed.
- Everything in Foundation
- One approved workflow
- Connected sources
- Review steps and action log
- Guided first runs and training
- Next-workflow recommendation
Full Commonhouse OS
Best fit: you want AI running several repeatable jobs across the business.
- Full OS setup
- Broader Company Brain and Source Library
- 3 to 5 workflows
- Review steps by role and risk
- Action logs and team training
- Steady operating rhythm
Support Plan
Keeps your home base clean and improving: system checks, source cleanup, tuning, and new workflows as you grow.
Starting prices, not final quotes. Live sends, publishes, and payments always need your approval.
The honest answers.
Will it reply to customers on its own?
No. The AI drafts, then stops at a review gate. Refunds, exchanges, and anything that touches a customer wait for your yes, you approve, edit, or hold. You can later let low-risk replies, like a tracking lookup, auto-send if you choose, but that's your call, not the default.
Where does it get the answers?
From your Company Brain and Source Library: your return policy, voice, offers, and the decisions you've already made. It works only from sources you approve, so replies are answered from what's true instead of guessed.
Do you take over our accounts?
No. You keep the accounts, the data, and the keys. Commonhouse works through your tools, not around them, and every action is written to an action log you own.
We have no AI in place today. Is that a problem?
No. Most brands start from zero. We map how post-purchase works today, set up your company brain and source library, and install one workflow end to end before adding more.
Delegate the work. Keep control.
Tell us where post-purchase hurts most. We'll build the first workflow around your business, gated, logged, and yours.