Most marketers think they're doing personalization right. They use merge fields for first names. They segment audiences into groups. They send "targeted" content.
But here's the reality: traditional segmentation is leaving money on the table.
Michael Roberts, an email marketing contractor specializing in SaaS and B2B lifecycle marketing, discovered this firsthand while working with Glide, a no-code app builder. Despite segmenting users into 6-8 groups based on keywords and sending relevant templates, something felt off.
When he audited his segmentation, the results were eye-opening. Users saying they wanted to build a "portal" were getting CRM templates when they actually needed inventory management tools. The misalignment was costing conversions.
The Breakthrough: Truly Individual Personalization
Roberts decided to push beyond traditional personalization. Instead of grouping users into segments, he built a system where every single email was uniquely crafted for each individual recipient.
Here's how it works:
User data flows from the marketing automation tool
That data gets sent to OpenAI via API
AI analyzes the user's specific needs and intent
AI generates personalized content and selects the most relevant template from 70+ options
The response gets logged back into the marketing automation platform
Personalized content populates the email
For Glide users, this meant the AI would:
Name the app they wanted to build based on their signup prompt
Match their use case to the most relevant template from a library of 70 (not just 6-8)
Pull the correct image, URL, and name for that template
Create messaging that truly reflected their specific intent
The Results: 3X Click Rate Improvement
The impact was dramatic. Click rates tripled at scale.
To put this in perspective: most email marketers celebrate a 10-15% lift in performance. Going from 1.2% to 1.3% click rate is considered a win. Roberts saw his click rates jump from just under 2% to around 6%.
This isn't incremental improvement. This is transformation.
The Technical Reality (It's More Accessible Than You Think)
Roberts admits he wasn't a coding expert when he started. He learned APIs through iteration, and AI itself helped him build the technical infrastructure.
The key components:
Marketing automation with webhook capabilities (he uses Customer.io, but HubSpot and others work too)
API connections to OpenAI
JSON formatting for data exchange
Willingness to iterate and learn from errors
The cost factor matters here. Using workflow tools like Zapier works for testing, but at scale, being slightly more technical saves significant money. Every personalization element requires multiple steps in workflow tools, quickly eating through limits.
The Strategic Shift: Think Different
The technical hurdles are real, but Roberts says the strategic shift is actually harder.
Marketers are trained in literal thinking: "If first name equals X, then show Y." Nuanced personalization requires completely different mental models.
You're not bucketing people anymore. You're understanding individual intent and matching it with relevant solutions in real-time at scale.
Getting Started: Start With Real Problems
Roberts's advice for marketers feeling overwhelmed: don't start with dreaming up use cases. Start with something you wish you could do but haven't been able to.
Ask yourself: What personalization would drive real impact but seems impossible at your current scale?
That's your starting point.
Then roll up your sleeves. Use AI to help you learn the technical pieces. Iterate on prompts. Set up guardrails. Test privately before going public.
The iteration process is critical. Roberts recommends starting broad intentionally—let the system give you some junk responses so you can see its tendencies and learn what limits to set.
The Brand Experience Difference
There's an important distinction Roberts makes: when a human fakes personalization in outbound sales, it's disappointing. But when a brand uses publicly available information to show they're trying to understand you? That's expected and appreciated.
Brands should be leveraging this. The effort to understand individual needs matters, even if it's not perfect.
The Future of Email Marketing
Traditional A/B testing that yields statistically insignificant results is "pretty much a waste of time," according to Roberts. He's done with incremental improvements that don't move the needle.
AI-powered personalization isn't about generating content or doing research—those are table stakes. It's about matching individual intent with relevant solutions at a scale that was previously impossible.
The marketers who figure this out won't just see incremental gains. They'll see transformational results.
