Marketing automation has acquired a stale aftertaste in recent years. Too many tools promised “more touch points in less time” — and produced, in practice, annoyed recipients and worn-down lists. The right question is not “how much can I automate”, but “what works better automated, and what does not”.
Three observations from practice.
What automation actually makes better
There are marketing activities that run, automated, in higher quality than manually. They have in common that they are rule-based and operate under time pressure:
- Onboarding sequences for new newsletter subscribers, new trial users, new customers. The first three to five touch points often decide everything — and they happen reliably only when they do not depend on daily mood and availability.
- Reminders of unfilled actions — abandoned registrations, unread relevant content, expiring trials. A well-placed reminder at the right moment is helpful, not pushy.
- Confirmations and receipts — who signed up for what, when does what happen, where is the Zoom link. These mails are not marketing in the narrow sense, but they shape the impression.
In all three cases, the alternative to automation is not “a better manual version”, but “does not happen at all”. Here automation makes the difference between a contact that gets cultivated and one that cools off.
What automation tends to make worse
There are also activities that lose quality through automation — sometimes dramatically. They have in common that they need context and relationship:
- Cold acquisition first contacts. An automated first contact — however cleverly personalised — almost always reads like what it is: a mass contact. With cold B2B leads this often backfires and burns the relationship.
- Reactions to substantive replies. If a recipient replies to an automated mail with a real question, a human must answer. If after three days a second automation trigger fires, the relationship is dead.
- Complaints and grievances. Even though technically possible — an automated reply to a complaint is always an escalation of the complaint.
The rule of thumb: the more individual the email content, the less suitable automation is. Standard mails yes; reactions to individual behaviour often not.
Where AI makes the difference
The strength of modern AI in marketing automation is not in sending more mails — but in hitting the right mail at the right moment with the right content.
Concretely:
- Trigger recognition beyond explicit clicks. Who shows real interest through reading and click behaviour? Who right now does not — even if still subscribed? Who is in a phase where our content is relevant?
- Content selection instead of content dispatch. From a library of existing posts, cases and whitepapers, the most likely relevant piece is chosen for each recipient — based on what they read or clicked in the past.
- Frequency adjustment. Whoever signals “too much” (opens nothing, clicks nothing) gets less. The active ones can handle more. A static cadence (“every Tuesday at 10”) is coarse by comparison.
- Topic drift detection. If a recipient’s click profile changes over months, that hints at shifting topic interests. The content is adjusted accordingly.
Externally this often looks less automated — and that is exactly the point. Good automation is the kind the recipient experiences as personal attention.
Preconditions without which it does not work
Three points are, in our experience, minimum preconditions for a productive setup:
- Clean consent. GDPR-compliant opt-in, documented, with a clear description of what the person expects. Without this foundation, everything that follows is legal risk, not marketing.
- Content with depth. If the library consists of three whitepaper stubs and a webinar recording link, AI has nothing to choose from. Only a minimum content depth makes intelligent recommendation possible.
- Clear marketing goals. What should the programme achieve — more newsletter engagement, more conversion to trials, more pipeline enrichment? Without clarity of goal, it cannot be measured whether AI helps or harms.
What becomes visible in the end
In the setups we have seen, absolute send volume often decreases — and at the same time open, click and conversion rates rise. That is not paradoxical: recipients get less but more relevant communication. They unsubscribe less. They respond more often.
Marketing automation is no longer a volume game — it is a relevance game. Whoever internalises that early has a structural advantage.