Why a Bad Process Becomes Bad Automation — AI Readiness, Part 4
- Trisha MerriamHost
- Chris CarolanHost
- Erin WiggersHost
October 28, 2026
Part 4 of AI Readiness asks why a bad process becomes bad automation. Chris Carolan, Trisha Merriam and Erin Wiggers on teams that can't say what their process is, what to do before you automate, and where AI helps and where a person still has the edge.
Moments from this episode
15 short cuts from this conversation.
Landscape
6 of 15
Key takeaways
Wanting everything automated is a natural first reaction after buying a new system. Chris says it is not that hard to show people why that is a terrible idea.
In Erin's words, the worst process is one you can't articulate. If you can't say what the process is, you are nowhere near thinking about how to automate it.
Few teams arrive knowing their process. Erin's number, if she's being generous, is like 20%, and for Trisha the majority need a conversation just to articulate it, because they have never had that conversation as a team.
Walk the process in real time. In Chris's telling, what someone calls three clicks is actually 20 in four different windows, and the rest is someone just knowing the answer, which you can't build a process around.
You may not need to automate yet. In Chris's HR example, a digital form placed right next to the hours available lets someone click approve, and in his words like 80% of the process gets better, because a process is usually bad when people don't have the context they need to make the decision.
Erin: if you built your existing process around constraints your new system doesn't have, you don't have to bring that baggage. Chris adds that making the new system look like the old one brings the old habits too.
Trisha's open question: will a person go as deep and put the time in with AI as with a human, and will AI ask the right questions on the fly? She already gets conversation fatigue from hard questions.
Chris says the healthiest organizations have been measuring time to value, while the entire business world is built to measure time to done.
Are you ready to accept AI telling you no? Chris expects that level of coaching very soon, and says the readiness question sits with the humans.
In a first conversation about process, Chris takes the tools off the table: no HubSpot or any other tool word, just how you sell. He believes that space creates some trust.
Ask AI where the holes in your process are. Asking how to make a bad step better tends to keep the bad step going, though Chris thinks that could change next year.
Erin sees the human edge in the client's perception of empathy and in getting ideas out of people's heads. Chris says the empathy to support and the empathy to challenge are completely different, and he has never experienced AI challenging consistently at the right time.
Human data entry is, in Chris's words, probably a good candidate for being classified as a bad process, but you don't get to touch that process until you can talk about it out loud.
Trisha and Erin both describe AI as especially helpful on edge cases: running real data through a process to find where it will break, with dry runs before production.
Chris's low-risk "let's just see" also applies to a claimed risk: test it, and show up to have a productive conversation about it.
Chapters
- 00:19Part 4: why a bad process becomes bad automationTopic
- 02:47What percent of teams know their process?Question
- 06:50Three clicks, actually 20Topic
- 08:23A stack of papers and a spreadsheetTopic
- 09:26Migrating into HubSpot means rethinkingTopic
- 11:36You actually don't need to automateTopic
- 13:14Don't bring that baggage with youTopic
- 15:22Will a human go as deep with Breeze?Question
- 16:25The gas station and conversation fatigueTopic
- 18:13Time to done or time to valueTopic
- 19:35Is it a good idea to enrich 1,000 contacts?Question
- 20:59The first session, with no tool wordsTopic
- 23:10Ask AI where the holes areTopic
- 25:02Empathy and reading the roomTopic
- 26:56A clunky experience, and where humans have the edgeTopic
- 29:22The empathy to challenge at the right timeTopic
- 31:26Two pictures from NectarTopic
- 34:40Find one spreadsheet, and human data entryTopic
- 36:48Walking it back: AI and edge casesTopic
- 39:26The low-risk let's just seeTopic
- 40:45Is it really a risk?Question
- 42:37Part 4 in the books, and what's nextTopic
Transcript
This transcript was generated by AI and may contain errors. It is provided to help you skim and search the conversation.
Keep learning
You just watched the work. Now build it.
The next step is doing it yourself — live, with people on the same path.
Get the next conversation in your inbox
One signal-dense note when a new episode lands. No noise.