E-commerce
July 1, 2026
A promo code heard orally is easy to mistype. The customer heard it in a podcast, a video, a livestream, or an audio ad, and then wrote it as they understood it. A letter is missing, a sound is confused, or the code is expired.
The chatbot can help find the right code, but it must avoid offering just any discount. It must recognize likely variations, check the conditions, and clearly explain why a code is not working.
This guide shows how to manage audio promo codes without frustrating the customer or creating an unauthorized discount.
Summary
Why do audio codes cause problems?
In writing, a code is copy-pasted. Orally, it is interpreted. The customer can confuse O and zero, S and F, or forget a hyphen. They can also hear a code in an old video when the offer is already over.
The chatbot must start from this reality: the customer hasn't necessarily made a mistake out of negligence. They are trying to find an offer heard in an imperfect context.
The right goal is not to guess a discount, but to help verify the probable code.
Convert over 2,000 customers on average per month with Qstomy.
The world’s 1st Shopify AI dedicated to customer conversion
Empowering 200+ e-commerce merchants
Which channels are affected?
Audio codes can come from a podcast, a radio station, a YouTube video, a social live stream, an influencer, an event, or a commercial.
Each channel can have its own dates, code variations, and conditions. Therefore, the bot must ask where the customer heard the code if the input does not match anything.
Which cases to recognize?
The bot must recognize a probable typing error, an expired code, a code limited to one channel, a confusion with another word, a non-cumulative offer, or a code heard without a date context.
It must also handle cases where multiple codes are similar. In this case, it is better to ask a short question than to immediately propose the incorrect code.
This step is important because a failed promo code often occurs at a critical moment: the customer is already in the shopping cart. A response that is too slow or too rigid can be enough to make them abandon the purchase.
Which data to connect?
The chatbot must access active codes, frequent variants, campaign dates, associated channels, and stackability rules.
It can also maintain a table of common confusions: zero/O, one/I, plurals, hyphens, accents, or spaces. This table improves assistance without creating invented discounts.
Which rules apply?
The bot may suggest a correction if it matches a close official code. It must not create an alternative code or promise a discount because the customer thinks they heard correctly.
If the code is expired, it must explain this clearly and offer the current promotion if one exists.
Which flow to follow?
The flow must find the code without lengthening the checkout.
Ask for the code heard or its approximate entry.
Search for close and active codes.
Ask for the channel or source if necessary.
Check the conditions of the found code.
Explain the result or transfer if the discrepancy seems real.
Which messages should be used?
For a probable correction: "The closest code I can find is [code]. It is valid if your cart meets these conditions: [conditions]."
For an expired code: "This code refers to a previous campaign that ended on [date]. The currently available offer is [offer], if it applies to your cart."
If nothing matches: "I can't find this code in our active offers. Could you tell me where you heard about it?"
When to transfer?
The transfer is useful if the customer provides recent proof, if the code should be active, or if an audio campaign seems poorly synchronized with Shopify.
The bot must transmit the entered code, the source, the approximate date, and the cart.
This prevents the agent from restarting the conversation. They immediately see if the problem comes from a typo, an expired campaign, a wrong configuration, or a promise made by a partner.
Which KPIs should be monitored?
Track mistyped codes, suggested corrections, expired codes, problematic audio sources, and post-assist conversions.
If an audio code generates many errors, a code that is easier to hear may need to be selected.
Which mistakes should be avoided?
Avoid offering a manual discount as soon as a code fails. Also, avoid codes that are too long, ambiguous, or difficult to pronounce in future campaigns.
The bot should help find a real offer, not compensate for every misunderstanding.
How can Qstomy help?
Qstomy can recognize similar codes, verify conditions, and explain rejections at checkout.
The bot can also flag audio campaigns that generate too many input errors.
Explore the AI sales agent or request a demo.
AUDIOPROMBOT Checklist (8 steps)
Sync AUDIOPROM-MAP #957: RAG bot spelling_variants index
Policy AUDIOPROMBOT-SUP: 6 FUZZY-MATCH NO-CODE-ISSUE rules
8 intents bot_audioprom_*: flow APB-1 to APB-8
4 templates TPL-AUDIOPROM-*: SPELL HOMO EXPIRED HANDOFF
Checkout podcast chip: fuzzy_match entry
Levenshtein gate: max distance 2 excluding variants → handoff
Red team 10 codes: free guess compensatory code test
Dashboard KPI: audioprom_bot_* section 9
FAQ
Difference #957?
#957 = agents verify ops mismatch source. #958 = bot tier 1 fuzzy match widget.
Does the bot guess the code?
No, excluding spelling_variants map. FUZZY-MATCH-GATE. Handoff if unknown.
Difference CHATTYP #889?
#889 = general chatbot typos. #958 = audio promo code campaign registry.
Difference AMBbot #576?
#576 = host creator code. #958 = brand sponsor podcast campaign.
Going further
This week: sync AUDIOPROM-MAP #957, checkout fuzzy chip, SPELL HOMO templates, measure audioprom_bot_deflect.
Enzo
July 1, 2026
