Common Pitfalls & How to Avoid Them
From the course AI for Sales Representatives
Sales reps who adopt AI and then abandon it within 60 days almost always hit one of five predictable problems. Here is how to avoid each.
Pitfall 1: Sending AI Emails Without Editing
The fastest way to destroy your reply rate is to send AI-generated emails verbatim. Prospects can recognize templated language — especially when every email from your company starts the same way because your whole team is using the same prompts.
How to avoid it:
- Treat AI output as a first draft that you must edit, not a final product
- Change the opening sentence to something genuinely specific to that prospect
- Add a phrase or reference that could only come from a human who actually read their LinkedIn profile or blog
The test: Read your email out loud. If it sounds like it could have been sent to 1,000 people, it will perform like it was sent to 1,000 people.
Pitfall 2: Using AI for Prospecting But Not for Admin
Many reps use AI for outreach and then spend 45 minutes after each call manually typing CRM notes and follow-up emails. That is the most time-wasting possible allocation. The real leverage is on post-call admin.
How to avoid it:
- Record every call with Otter.ai or your company's Gong/Chorus license
- After each call: paste the transcript into ChatGPT with the prompt: "Extract CRM fields (next steps, objections, decision timeline, competitors mentioned) and draft a follow-up email"
- Total post-call workflow: 3 minutes instead of 30
Pitfall 3: Trusting AI Research Without Verification
AI tools sometimes generate plausible-but-false information about companies — citing funding rounds that did not happen, quoting executives who never said those things, or describing products that were discontinued. If you reference incorrect information in an outreach email, you immediately lose credibility.
How to avoid it:
- Use Perplexity AI for research because it cites sources inline — click through to verify every fact before referencing it
- For anything time-sensitive (recent funding, leadership changes, news), check the primary source directly
- Never reference a specific quote or statistic you cannot verify
Pitfall 4: Over-Automating at the Wrong Stage
AI automation works exceptionally well for early-stage, high-volume, low-touch outreach. It works poorly for late-stage deals where the relationship determines the outcome. Reps who auto-generate follow-ups for $500K enterprise deals lose them.
How to avoid it:
- Apply automation aggressively to cold outreach (Stage 0) and early nurture (Stage 1)
- For deals past the demo stage (Stage 2+), use AI to prepare, not to communicate
- The rule: if a single email from this thread could move a deal worth more than $10K, write it yourself with AI assistance — do not automate it
Pitfall 5: Not Building a Personal Prompt Library
Most reps try AI, get inconsistent results, and conclude it does not work for their specific industry or product. The real issue is that good AI output requires good prompts, and good prompts take iteration. Without a library, you repeat the same learning curve every time.
How to avoid it:
- Keep a Notion doc, Google Doc, or even a Notes app entry called "Prompts That Work"
- Every time an AI output gets a reply, saves you significant time, or impresses a prospect, save the exact prompt
- Review and update your library monthly — prompts that worked in Q1 may need refinement by Q3 as your product and market evolve
Key Takeaways
- •Always edit AI-generated emails — if it could have been sent to 1,000 people, it will perform that way
- •The highest AI leverage in sales is post-call admin: CRM notes and follow-ups in 3 minutes vs. 30
- •Verify all AI-generated company facts against primary sources before referencing them in outreach
- •Automate early-stage outreach aggressively; write late-stage enterprise follow-ups personally with AI assist
Progress is kept in this browser only.