Pre-purchase questions
Answers sizing, stock, shipping and compatibility questions from your real catalog data, in the customer's chat, at the moment of intent — before they open a competitor's tab.
The question was buying intent
the eight-hour reply was the goodbye.
Intent decays by the hour
"Does this fit a 15-inch laptop?" at 10:40 PM is a sale at 10:41 PM and a closed tab by morning. The answer's value has a half-life of minutes.
Answers live in six places
Size charts in images, stock in the store, shipping rules in a doc, compatibility in the team's heads. The agent stitches it per question — slowly, sometimes wrongly.
Most askers never ask
For every customer who writes in, several just leave. The ones who do ask are handing you the objection — and the queue hands them silence.
They told you exactly what's stopping the purchase. The answer came after they'd left.
The objection is answered in seconds. With facts, not enthusiasm.
- Product: Metro backpack · viewing for 4 minutes
- Spec sheet sleeve fits up to 16", padded
- In stock 23 units, ships next day
"Will the Metro fit a 15-inch MacBook with a sleeve on it?"
Reads the spec, stock and shipping promise — answer assembled from data, not vibes.
- Fits: yes 16" max, padded compartment
- Honest caveat bulky sleeves over 3cm are tight
Yes — the Metro's padded compartment takes up to 16". With a thick sleeve (3cm+) it's snug but fits. In stock; order in 40 min and it ships tomorrow.
Answer logged against the product — every question improves the next one.
- Order #7781 placed 10:47 PM, Metro backpack
- Question-to-order 6 minutes, one channel
Order confirmed! It ships tomorrow — tracking will land right here.
Question, answer, order — one thread, attributable revenue.
- 3 questions unanswerable — spec missing from catalog
- Draft catalog fixes attached per gap
Approve the catalog additions — next week those answers are instant too.
The question log becomes your conversion roadmap.
Same questions. Answered in time to matter.
| Moment | Before Clarwiz | On Clarwiz |
|---|---|---|
| Reply time | Hours — mornings after, mostly. | Seconds, at the moment of intent. |
| Answer source | Whatever the agent remembers. | Catalogue, specs and stock, cited. |
| Night and weekend asks | Wait for Monday. | Answered at 10:40 PM like it's 10:40 AM. |
| The honest edge case | Skipped — enthusiasm sells. | Stated — trust sells more. |
| What you learn | Nothing; asks vanish into tickets. | A ranked log of what blocks purchases. |
What stops it from making up product details?
Can it recommend between two products?
Does it handle discounts and price-match asks?
Which languages does it speak?
What do we need to integrate before a demo?
See this run on your own data.
Thirty minutes, nothing connected.
We take this process, connect nothing, and show you the run end to end on your real conversations and orders. When you're ready, go live at the autonomy level you choose.