AI changes inbox visibility; marketers need to rethink strategy

Musa Kalenga, CEO of the Brave Group says AI filters now sideline up to 40% of emails before customers see them (Image supplied)

Approximately 4.6 billion people worldwide use email, generating an estimated 376 billion sent and received messages every day. The average consumer actively manages 12 or more inboxes across different platforms.

Consumers lack the cognitive capacity to process this sheer volume.

AI algorithms step in

Artificial intelligence algorithms can now step in to evaluate and summarise messages before the human recipient ever sees them.

Major email providers have deployed AI features to prioritise threads and draft replies. Industry estimates suggest AI filters deprioritise up to 40% of emails before human eyes ever see them.

Brands that rely solely on technical deliverability requirements like authentication and sender reputation will reach the server but fail to earn visibility. Deliverability has evolved into desirability.

Can AI help change how businesses view customer service emails?

Ryan Falkenberg  8 Jul 2024

Machine and human readers

Every email is now evaluated by two distinct readers: the machine and the human.

The human reader scans for emotional relevance and trust. They decide to engage based on personal resonance.

The machine reader scans for clear intent and structured signals. It assesses the subject line, preview text, and alt text to rank priority. It generates a summary before the human reader opens the email.

Marketers must write for both

Marketers must write for both simultaneously.

Copy tuned purely for human emotion often gets deprioritised by the algorithm. Copy written strictly for machine logic feels cold and earns a manual deletion.

The opening lines of a campaign must state the specific value immediately so the human understands the outcome and the AI can parse the intent accurately. Clear, specific language performs better than vague curiosity.

AI-powered chatbots make buying insurance painless

8 Apr 2019

Traditional measurement tools compromised

AI summaries have compromised traditional measurement tools. Inboxes that automatically open messages to generate summaries artificially inflate open rates.

Optimising for an open rate now means optimising for algorithmic inflation. Marketers must pivot to measuring engagement depth and downstream conversion signals.

Effective personalisation relies on behavioural signals rather than static demographic labels. Inserting a first name using a merge field proves data access.

Triggering an email based on browsing behaviour proves attention. Relevant signals include recent clicks, purchasing frequency, search intent, or declared preferences.

Campaigns leveraging this personalised, behaviour-driven approach generate six times more transactions than generic email blasts.

Retail marketers face a huge AI shift in 2026

Vincent Maher  1 Oct 2026

Empathy comes first

Technology alone cannot salvage a poor message. The PET model dictates a strict operational sequence: start with Empathy, apply Persuasion, and use Technology last.

• Empathy: Define the audience's problem, need, or desired outcome.
  
• Persuasion: Craft a relevant promise supported by a credible reason to act.
  
• Technology: Deploy a scalable journey using data, automation, and controlled variants.

Building a campaign brief in the correct order is critical.

1. Teams must first describe the person and the context in plain language.
   
2. Next, they define the value exchange by stating what the recipient gains by acting.
   
3. Only then do they choose the persuasive idea.
   
4. Then they select the technology, data triggers, and reporting methods.

Move to predict-and-deliver journeys

Moving from calendar-driven batch sends to predict-and-deliver journeys transforms the operating model. A lifecycle engine uses live data from customer relationship management systems and customer data platforms to trigger messages based on specific actions.

Consider a new insurance customer onboarding process. If the customer opens a welcome sequence but does not click, the system sends a shorter variant to reduce friction.

If the customer browses a specific cover type, the system triggers a benefit-led follow-up matched to that observed interest.

How AI is challenging traditional agency billing models and defining value

Musa Kalenga  26 Aug 2026

Behaviour drives communication

If the customer calls the contact centre, the email journey pauses automatically and resumes later with relevant service content.

Behaviour, not a fixed calendar, drives communication. Automated, behaviour-triggered emails can generate up to 16 times more revenue per send compared to manual batch campaigns.