Use the AI suggested reply

Every inbound inquiry on Yesoma arrives with a draft reply already written for you. It lives on the case page and in the inbox preview panel so you can review, tweak, and send without switching screens. This guide walks through every control around that draft — what it means, when to trust it, and when to take the wheel.

1. Where it lives

Open any case in the inbox (the /inbox/[id] page). The Suggested Reply card sits in the right-hand reply panel, bordered in green. The same card appears in the preview panel when you hover over an inbox row — you can read, edit, and send without ever opening the full case.

The card title changes to Using template in violet when you have swapped the AI draft for one of your saved templates (see section 7). Switch back to the AI draft at any time with the Reset to AI button.

2. Why-this-reply transparency

Beneath the draft text you will see a small chip — a one-line explanation of which saved data the AI pulled. It might say something like "Based on your bridal package pricing + 48-hour deposit policy" or "Using your FAQ: turnaround time for edits."

This is the trust line. Before you send, glance at it. If the chip cites data that fits the customer's question, the draft is likely solid. If it cites something unrelated — or says nothing at all — that is your signal to read more carefully before sending. The AI never silently guesses; if the chip looks thin, the Business Brain probably has a gap worth filling.

3. Edit before sending

Click the pencil icon on the Suggested Reply card to open the inline editor. Type your changes directly in the text area, then click Save edit. The send buttons below the card — Send branded reply, Reply on WhatsApp, and Send from Yesoma — all pick up your edited version the moment you save.

You do not need to re-draft from scratch to make a small change. Most owners send the AI draft as-is about 70-80% of the time and make light edits on the rest. The editor is there for those cases where the draft is 90% right and needs one sentence adjusted.

4. Regenerate

The Regenerate button re-runs the AI using the full current thread as context. Use it when:

• You just added a missing service, policy, or FAQ to the Business Brain and want the next draft to reflect it.
• The first draft came out flat or off-tone and a fresh run might do better.
• A new message from the customer changed the direction of the conversation.

Regenerate does not erase your edits — it produces a new draft and replaces the card content. If you preferred an earlier version, you can type it back in using the editor.

5. Voice reply

The small microphone button on the card lets you dictate your reply instead of typing it. Speak naturally; the AI rewrites your dictation in your brand voice and drops the result into the reply field. From there, save and send normally.

For setup instructions and language notes, see Voice reply and language translation.

6. Language translation

The language pill next to the card lets you translate the current draft into another language before sending. Click it, pick the target language, and the card refreshes with the translated text. You can edit the translation before sending.

Full details in the same article: Voice reply and language translation.

7. Template pills

Above or below the Suggested Reply card you will see pills for templates Yesoma considers relevant to this inquiry. These come from your saved template library, matched by topic.

Click a pill to swap the AI draft for that template's body — the card title flips to Using template in violet. All send buttons use the template text. Click Reset to AI to go back to the AI-generated draft.

Templates and AI drafts work side by side. The template pills do not replace the AI suggestion; they give you a one-tap override when you have a carefully crafted reply already written for a situation the AI handles regularly.

8. When to trust the draft vs. when to edit

Send as-is when:

• The customer asked a simple question that is fully covered in your Business Brain.
• The why-this-reply chip cites the right data for this conversation.
• The tone matches how you normally write to customers.

Edit before sending when:

• The customer asked something the AI did not have context for (the chip looks vague or the draft hedges).
• You want to add personal warmth — name-drop their child's birthday party date, reference a conversation you had on the phone, or match the energy they brought.
• The situation is delicate — a complaint, a refund request, a price negotiation, or any moment where a generic response would do more harm than good.

The AI draft is a starting point, not a final answer. It saves you the blank-page problem; you stay in control of what actually goes out.

9. What the AI knows about

Knows:

• Everything in your Business Brain: services, prices, policies, FAQs, tone samples, and any templates you have saved.
• The full message thread for this specific case — every message, every reply, from the beginning.
• Corrections and edits you have made to past drafts on similar inquiries (Yesoma learns from these over time).

Does not know:

• Anything that is not in your Business Brain. If you verbally agreed on a custom price during a phone call but never logged it, the AI does not know it happened.
• Activity in other cases. A conversation you had yesterday with a different customer does not carry over here.
• Real-time availability, calendar details, or live third-party data unless you have explicitly connected those sources.

If the draft says something surprising, check the Business Brain first. Nine times out of ten, either the data is missing or a policy entry needs updating.

10. Sending

The reply area below the Suggested Reply card has three send paths:

• Send branded reply (Business plan and above) — sends via your Yesoma-connected From address, with your logo, signature, and footer applied automatically. The customer sees your brand, not a generic sender.
• Reply on WhatsApp — opens a wa.me link with the draft text pre-filled. You review it in WhatsApp and tap send there.
• Send from Yesoma (the dialog that appears when sending from the inbox card preview) — same branded send path, usable without opening the full case page.

Your edits and any template swap apply to all three paths. Whichever button you click, it sends the text currently in the card.

11. Common questions

The AI keeps suggesting something wrong. Why?

Almost always a Business Brain gap. If the AI is citing an old price, a policy you changed, or a service description that no longer fits, the fix is in the Brain — not in editing every draft by hand. Open Business Brain, find the stale entry, update it, and the next draft will reflect it. See Build your Business Brain for a refresher.

Can I disable the AI for one customer?

There is no per-customer toggle. The suggested reply always appears. If it consistently misses the mark for a particular customer, the underlying issue is usually a context gap in the Brain or an unusual situation that just needs a manual reply. Edit before send and, if you make a change that applies broadly, save it back to the Brain so future drafts get it right.

Does the AI ever send a reply automatically without me clicking?

No. Every outbound message from Yesoma is owner-initiated. The AI drafts; the human sends. Nothing goes to a customer unless you click a send button. This is intentional — the AI is a drafting assistant, not an autopilot.

Related

• Build your Business Brain — keep the data the AI draws from accurate and current.
• Voice reply and language translation — dictate replies and translate drafts.
• Use community templates and playbooks — add pre-built templates your team can use as pills.
• Assign cases to teammates — route cases so the right person reviews and sends each draft.