E-commerce
July 1, 2026
A chatbot can respond quickly, but a fast response becomes dangerous if it is no longer up to date. The customer may receive an old return policy, an outdated delivery time, an expired promotion, or obsolete product information.
The real problem is not just the error. It is the loss of trust: the customer has written proof from your assistant, then discovers that the information no longer applies.
This guide explains how to manage chatbot response versioning: up-to-date sources, validity dates, old policies, sensitive cases, and human transfer when information changes.
Summary
Why are old responses a customer risk?
Obsolete information can lead to a bad purchasing decision. A customer orders because the bot announces delivery within 48 hours, then discovers a 5-day delay. Another returns a product thinking they have 30 days, when the policy has changed.
The chatbot is perceived as an official voice of the brand. Even if the mistake comes from an old document or a forgotten FAQ, the customer will remember that the store gave them the wrong information.
An undated answer quickly becomes a permanent promise. This is exactly what must be avoided.
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
What information changes most frequently?
The most sensitive information is that which evolves with business operations: delivery times, fees, promotions, return conditions, warranties, free delivery thresholds, product availability, and B2B rules.
Legal content and customer policies change less frequently, but have more impact when they are incorrect. The bot must therefore know which sources take priority and which information must be dated.
A good rule of thumb is to consider any response related to a commercial condition as temporary, unless proven otherwise.
How to organize the chatbot's sources?
The chatbot must rely on its own document base. Each source must have an owner, an update date and, if necessary, a validity start or end date.
An old hidden page, an unpublished FAQ or a draft policy must not be indexed. If several documents contradict each other, the bot must prioritize the most official source: active policy page, Shopify configuration, promotional tool or validated support base.
Without this hierarchy, the chatbot sometimes chooses the most plausible response instead of the correct response.
Which versioning rules should be applied?
Sensitive responses must contain validity logic. For example: "Offer valid until [date]", "Policy applicable to orders placed after [date]" or "Estimated delivery time based on the currently selected carrier".
The bot must avoid absolute phrases when the rule depends on a date, country, product, or campaign. It is better to write: "For your order, the applicable rule is..." rather than "We always accept...".
This precision protects the brand and makes the response more useful for the customer.
How to respond when a rule has changed?
When a rule has changed, the response must acknowledge potential confusion. The bot can say: "This information has been updated. For your request, the current rule is [rule]."
If the customer quotes an old response or a screenshot, the bot should not deny it. It must explain that the information may have changed and escalate if a gesture or an exception needs to be investigated.
The right tone is not "you misunderstood". It is "I am checking which version applies to your situation".
Which flow to follow?
The flow must determine the applicable rule before responding.
Identify the subject: return, delivery, promotion, warranty, price, or availability.
Check the active source and its validity date.
Take into account the order date or country if necessary.
Respond with the rule applicable to the customer's case.
Escalate if the customer mentions an old promise or a contradiction.
Which messages should be used?
For an updated policy: "The policy was updated on [date]. For your order placed on [date], the applicable rule is [rule]."
For a past promotion: "This offer was valid until [date]. The currently active offer is [offer], if it applies to your cart."
For a contradiction: "I see that there may be a difference between two pieces of information. I am forwarding your request so that an agent can verify the applicable rule."
When to transfer to a human?
Transfer is necessary if the customer shows a screenshot of an old response, disputes a promise, requests an exception, or if multiple internal sources contradict each other.
The bot must transmit the subject, the source cited, the order date, the response given, and the rule currently found. The agent can then make a decision without having to redo the entire diagnosis.
Which KPIs should be monitored?
Follow conversations where the customer reports contradictory information, corrected answers, obsolete sources removed, and escalations related to an old policy.
A good indicator is the number of indexed documents without an owner or without an update date. The higher it is, the greater the risk of an obsolete response increases.
Which mistakes should be avoided?
Avoid indexing drafts, using outdated FAQs, answering without checking the date, and presenting a variable rule as permanent.
Also, avoid removing all traces of old rules without a procedure. For some orders, the old rule may still apply.
How can Qstomy help?
Qstomy can structure chatbot sources, prioritize active documents, and help avoid obsolete answers.
The bot can also transfer contradictory cases with complete context: source, date, command, and given answer.
Explore AI support or request a demo.
VERSIONbot Checklist (8 steps)
Sync CONTRABOT-MAP #883: feed intents conflict priority purge
Policy VERSIONBOT-SUP: 6 rules CANONICAL SYNC EFFECTIVE-DATE
8 intents bot_version_*: flow AVB-1 to AVB-8
4 templates TPL-VERSIONbot-*: POLICY STALE-BLOCK SYNC-NOTICE RESUME-REFRESH
version_id on KB chunks: effective_from deprecated_at
COHERENCE Link #191: sync_trigger on registry change
Red team stale: deprecated chunk still indexed test
KPI Dashboard: version_bot_* section 9 + delta contrabot_
FAQ
Difference #883?
#883 = agents process contradiction tickets. #884 = bot serves active canonical version.
Difference #103?
#103 = clean the initial corpus. #884 = continuous post-launch versioning maintenance.
Difference #885?
#884 = versioning policies KB. #885 = sync recent catalog promos.
Which sync SLA?
sync_sla_hours registry #191. Typical target 4 to 24 hours depending on customer service criticality.
Going further
This week: tag version_id shipping return chunks, activate CANONICAL-ONLY RAG, configure sync_trigger #191, measure version_bot_stale_answer_rate.
Enzo
July 1, 2026
