Support and Pre-Sales Chat for SaaS
Last updated:
Documentation is the advantage
Software businesses already have written documentation, which is exactly what a retrieval-based bot needs. Most other sectors have to write it first.
That means a software product can usually get to a good answer rate faster than anyone else.
Separate pre-sales from support
- Pre-sales: pricing, limits, integrations, security — route to sales
- Support: how do I, why is it doing that — route to support
- Bugs: raise a ticket with the details captured
- Account: billing and cancellation — always to a person
Mixing pre-sales and support in one queue means paying customers wait behind evaluators, which is exactly the wrong order.
Cite the documentation
Every answer should link the page it came from. Users who can read the full page solve their own next question without asking.
It also makes wrong answers immediately checkable, which is what builds trust in the bot.
What it should never do
- Promise a feature that is not shipped
- Give a date for anything on a roadmap
- Answer a security questionnaire
- Handle a cancellation or refund
Deflection is not the only metric
Ticket deflection is easy to measure and it is not the whole picture. A bot that deflects tickets by frustrating people into giving up is a loss disguised as a win.
Watch satisfaction and churn alongside deflection, and read the conversations that ended without resolution.
Frequently asked questions
What answer rate should a software product expect?
Should it answer technical questions in depth?
Can it help with onboarding?
What about API questions?
Support tickets that the docs already answer?
A bot over your documentation deflects those instantly. Start at spideychat.com.