Welcome back to ‘Tactical Tips’ by Jerel and Shuo at DECODE, the largest founder community co-hosted across Berkeley and Stanford.
Today, we’ll be answering the question, “How to automate support without losing trust?” and covering insights around how to free up your team’s capacity for higher-value work, design AI workflows with clear escalation paths, and build a knowledge base that gets better with every interaction.
Mary-Beth Donovan, former Chief Customer Officer at EnterpriseDB, has a practical framework for automating support without losing customer trust. So, here’s advice inspired by her.
And ... we’ve curated a YouTube playlist featuring our best founders, operators and investors.
🔥 Inside this issue:
✅ 3 tactics to automate support without losing trust
✅ 2 traps to avoid
✅ 1 tool to leverage
👇Let’s dive in.
Grab 30 mins with Jerel - Need personalized advice on building your startup or just want to talk? Happy to help and make intros if it’s the right fit.
💌Someone shared this with you? Subscribe here.
Nov 14: Roundtables with execs from OpenAI, NVIDIA, NEA, Cisco and more
Join the DECODE Conference on Nov 14 at UC Berkeley and bring your current startup blockers.
3 things that make this day different from every other events:
Every session targets a blocker pre-seed and seed founders are hitting right now, and sends you home with actionable tactics
Speakers stay in the room and sit down for an off-the-record roundtable with 12 to 15 founders
Every speaker has founded, funded, operated, or advised an early stage startup
Past attendees have raised pre-seed to series A rounds from investors they met through DECODE, including a16z and Sequoia, as well as found advisors, mentors, and team members to help 10x growth.
Register now with code DECODE for 30% off.
3 tactics to automate support without losing trust
🎬 Start with capacity, not cost savings
Use automation to free up time, not to cut headcount
Set the first goal as freeing up one full-time equivalent (FTE) from repetitive, low-value interactions
Shift the reclaimed time upstream into proactive, strategic customer conversations; customers accept automation if it leads to better human engagement where it counts.
Before setting an automation target, ask:
Which interactions are repetitive and rules-based?
Which conversations require judgment and context?
What higher-value work would we do if we reclaimed this time?
Where does the customer experience get better as a result?
Drop any automation goal that doesn’t lead to a better customer experience somewhere else
📋 Design aroud trust and tolerance
Treat trust and tolerance as the real success metrics
Automation doesn’t need 100% perfection. Good enough is acceptable if you preserve trust
Customers will tolerate imperfection if they feel understood and steps of progress are being made.
Tolerance is lost when customers feel stuck in a loop with no path forward and have no off-ramp to access the escalation path.
Define the exact moment a human steps in for every automated workflow, and make that handoff feel seamless, not hidden or punitive
Practical guardrails include:
A clear path to escalate to a human
Monitoring negative feedback patterns
Regular review of cases that required escalation
Automation without a visible safety net erodes confidence quickly.
🧭 Fix the knowledge base before automating
Audit current documentation for outdated articles, inconsistent answers, and unclear processes; AI exposes and amplifies these gaps as soon as automation begins.
Document edge cases and failure paths before scaling automation; poor documentation leads to poor answers and frustrated customers.
Build a learning loop:
Design the moment of being wrong as a chance to capture new knowledge, not to argue
E.g. When a customer finds an answer the bot missed, the bot asks how the issue was solved and logs the steps for review
2 traps to avoid
🚨 Chasing arbitary automation targets without reviewing impact
Don’t chase an arbitrary number like 80 percent automation without considering the downstream impact
Set a goal tied to a clear business outcome and revisit them quarterly
E.g. freeing up capacity for proactive outreach or improving time to resolution on complex cases
🚨 Focusing on deflection rates only
Measure customer satisfaction and expansion indicators in addition to deflection rates.
If satisfaction drops or renewal conversations become harder, reassess
Automation is not a one-time project; it is an evolving balance between efficiency and experience
1 tool to leverage
📖 Best practice on automating support workflow
Three lenses to evaluate every automation decision: operational efficiency, customer experience, and product value
There is no universal “right” automation percentage; the correct level depends on your customer base, complexity of product, and support model
Leverage tools like Fin when ready for automation
Bonus: 1 trend to spark startup ideas
Most AI Tools Weren't Built for This
B2B customer issues move across teams and systems, not through a single chat window. A new Harvard Business Review Analytic Services briefing paper, sponsored by Front, breaks down where AI tools fall short in B2B service and what to ask before you invest.
Get more involved ❤️
Continue learning 🎓
How to build
How to sell
How to operate:
Other perks and resources 📚
Learn more about the latest trends in startups/tech/venture
Access our crowdsourced Founders’ Prompt Library to build faster
Apply to participate in Founders Pop-up Board Advisory and receive startup feedback from execs at Google, Microsoft, Meta, Reddit
Schedule for a free consultation on structuring your equity-based compensation plan
Fill this form and we’ll get in touch for details on how we can get your brand in front of our community
Get $5,000 in credits with Deel to set up your payroll in 150+ countries in minutes, onboard contractors compliantly, manage benefits and IT access, and handle visa.
Kickstart your idea 💡
Still figuring out your startup idea?
Help us improve 🌱
Please complete this 2-min survey to help us get to know you and better tailor content for you.
“Reply” with any follow-up questions you might have, and we’ll work on covering them in a future newsletter!
Stay tuned for more startup wisdom in next week’s edition!
