Part 1: Separate Signal from Noise
The Problem
You're in a vendor demo. The rep says: “Our platform uses advanced AI to deliver hyper-personalized experiences at scale.”
Do you:
- Know what that actually means?
- Know how to verify the claim?
- Know what questions to ask?
Most marketing leaders can't. And that's costing them.
The 3-Question Filter
Every technical claim should pass these tests:
Can you replace "AI" with a simpler word?
The Test:
If the claim still makes sense after replacing “AI” with “software” or “automation,” it's probably not AI.
"Our AI personalizes email subject lines"
Replace with "software" → "Our software personalizes email subject lines" → This could just be A/B testing. Ask: "What type of AI model are you using?"
What does it DO vs what does it COST?
The Test:
If the vendor can't explain the specific action the technology performs, they're hiding behind buzzwords.
"AI-powered marketing orchestration"
"Our system analyzes email open times from past 90 days and schedules future sends within ±2 hours of each contact's most common open window."
Questions to Ask:
- →“Walk me through exactly what happens when I click 'send.'”
- →“What data inputs does the system use?”
- →“What outputs do I get?”
Has ANY human reviewed the output?
The Test:
If AI is making decisions without human review, ask what happens when it's wrong.
Real Scenario:
A martech vendor claims “AI writes your email campaigns automatically.”
What to ask:
- →“Who reviews the emails before they send?”
- →“What happens if the AI writes something off-brand or incorrect?”
- →“Can I see examples of mistakes it's made?”
Why this matters:
AI that runs without human oversight = risk you're inheriting without knowing it.
Try It Right Now
Here are 3 real vendor claims from recent demos. Apply the 3-Question Filter:
Vendor Claim #1
“Our predictive AI identifies high-intent buyers before your competitors do.”
Q1: Replace “AI” with “software”
Replace "AI" with "software" → "Our software identifies high-intent buyers..." Still makes sense. This might just be lead scoring.
Q2: What does it DO?
Ask: "What signals are you tracking? How is this different from traditional lead scoring?"
Q3: Human review?
Ask: "Do these leads go straight to sales, or does someone vet them first?"
Vendor Claim #2
“Our AI-powered chatbot resolves 80% of customer inquiries automatically.”
Q1: Replace “AI” with “software”
Replace "AI" with "software" → Could be keyword matching. Ask what happens with edge cases.
Q2: What does it DO?
Ask: "What types of inquiries? Show me examples of the 20% it can't handle."
Q3: Human review?
Ask: "When it fails, how do customers escalate to humans?"
Vendor Claim #3
“Machine learning optimizes your ad spend in real-time.”
Q1: Replace “AI” with “software”
This could be ML, or it could be rule-based bidding. The word "real-time" is doing heavy lifting.
Q2: What does it DO?
Ask: "What data inputs drive the optimization? How often does the model retrain?"
Q3: Human review?
Ask: "Can I see a report of decisions it made and why?"
What You Just Learned
- You can now identify when “AI” is actually just automation
- You know the 3 questions that expose empty technical claims
- You can demand specifics instead of accepting buzzwords
Next: Learn how to translate the technical jargon you'll hear in every vendor meeting.
Want daily intelligence like this? Join thousands of marketing leaders getting AI Ready CMO.