AI in Post-Sales: Support Is Moving Forward. Customer Success Is Hesitating.
There’s a lot of noise right now about what AI can and can’t replace in SaaS. Most of it misses the mark by being too broad, or too theoretical.
When you narrow the lens to post-sales, something interesting shows up.
I largely agree with the current thinking around AI in customer support. We’re already seeing meaningful impact there. Ticket deflection is real. Internal knowledge retrieval is faster. First responses are better. Resolution times are coming down. Support leaders are having the conversation, experimenting, and moving.
Customer Success is a different story.
Two very different post-sales worlds
From recent experience, I’m seeing two completely different perspectives on AI in post-sales, depending on where the function sits in the org.
When post-sales reports into Sales
Inside HubSpot’s Health Score configuration, Preview Distribution allows teams to model how their scoring logic plays out across real customers—before the score goes live. Despite its importance, it’s one of the most overlooked tools in Customer Success Operations.
The thinking is pragmatic and execution-focused.
The conversation usually sounds like this:
Let AI handle internal knowledge so humans stop acting like walking wikis
Let agents answer “how do I” and “where do I find” questions instantly
Use automation to remove noise, not to upsell
Bring CSMs in only where judgment, context, and value creation are required
In these environments, AI is not viewed as a replacement for people. It’s viewed as a way to create capacity.
No one is trying to automate relationships. They’re trying to automate friction.
The goal is simple. Stop burning expensive human time on low-leverage work so CSMs can actually do the work we say they should be doing.
When post-sales reports into Customer Success
I’m not hearing the same openness.
In many cases, I’m not hearing the conversation at all.
CS leaders are hesitant to even broach the subject of AI taking on pieces of the post-sales motion. Not because the technology doesn’t work, but because it challenges a long-standing belief that everything post-sale must be human-led to be valuable.
That belief used to make sense.
Products were simpler. Customer expectations were lower. Books of business were smaller. CSMs could afford to be involved in almost everything.
That world doesn’t exist anymore.
The uncomfortable gap
Support teams have already accepted that AI can handle a meaningful portion of their workload without damaging the customer experience.
Customer Success hasn’t crossed that bridge yet.
And I don’t think it’s because the use cases aren’t there.
I think it’s because CS has historically struggled to clearly define:
What is truly value-add
What is repeatable and automatable
What only a human should do
If you can’t draw those lines, any automation feels threatening instead of enabling.
What AI should and should not do in CS
Let’s be clear about a few things.
AI should not be upselling customers.
AI should not “own the relationship.”
AI should not replace strategic conversations or executive alignment.
But AI absolutely should:
Answer internal questions instantly
Surface account context without a human digging through tools
Handle routine post-sale interactions that do not require empathy or judgment
When done correctly, this does not cheapen Customer Success. It sharpens it.
It gives CSMs the time and space to focus on:
Strategic alignment
Executive-level conversations
Value realization
Identifying risk before it becomes obvious
The choice CS leaders need to make
Support figured this out first.
Sales-adjacent post-sales teams are figuring it out now.
Customer Success needs to decide whether it wants to lead this shift, or wait until it’s forced upon them by budget pressure, burnout, or executive mandates.
AI isn’t coming for Customer Success.
But it is coming for unclear roles, undefined value, and inefficient motions.
And honestly, that might be exactly what CS needs.