B2B outreach is increasingly resembling a tragedy of the commons.

      AI has not made it easier to reach buyers; rather, it has rendered everyone more susceptible to interruptions.

      A contemporary AI sales agent can conduct research on a company, identify a possible buying signal, create a tailored email, schedule follow-ups, assess responses, and pass a qualified lead to a human salesperson. Major CRM providers are now offering agents that can research leads, draft and send personalized outreach, and follow up with minimal human involvement. Microsoft introduces an autonomous sales qualification agent capable of qualifying leads at scale, while HubSpot presents a prospecting agent for automated research and outreach. This functionality is now transitioning from a speculative sales feature to an established component of sales infrastructure.

      For individual companies, the economics are appealing. Tasks that once took hours of human effort can be completed in mere seconds, leading to a rational response of reaching out to more prospects. However, when every company adopts this approach, the collective outcome is not increased relevance but rather heightened interruptions.

      B2B outreach is evolving into a tragedy of the commons.

      In this scenario, the shared resource is not the inbox itself. Traditionally, a tragedy of the commons involves individuals acting in self-interest, depleting a resource that is shared. In outbound sales, the shared resource isn’t email, which is virtually unlimited. Instead, it is buyer attention, channel trust, and sender reputation that are scarce resources. Every unsolicited message chips away at each of these resources; a valuable message can offset this cost by helping the recipient find a relevant solution, while an irrelevant message withdraws without providing any value in return.

      No individual sender owns the recipient's attention, nor do they control how mailbox providers handle unwanted messages. Yet, every sender has a vested interest in utilizing a bit more of both resources. Initially, the damage seems manageable—reply rates might slightly decline, more messages end up in spam, and buyers become less open to unfamiliar senders. In response, teams ramp up their outreach, creating a cycle where reduced production costs lead to increased messaging volume, which results in diminished attention per message, lowering response rates and driving further volume. Complaints and disengagement escalate, causing mailbox providers to strengthen their filtering systems, rendering the channel less effective for all.

      This phenomenon can be termed interruption inflation: as the volume of automated outreach outpaces human attention supply, each subsequent message garners less attention than its predecessors.

      Data indicates a market under strain. At Belkins, we have observed this market at a significant scale. For our 2025 cross-channel benchmark, Belkins and its research partners analyzed 16.5 million cold emails, over 20 million LinkedIn outreach attempts, and more than 5 million cold calls. This extensive dataset illustrates that the pressure is not confined to any single campaign, industry, or channel.

      A separate analysis by Belkins of over 7.5 million cold emails sent in 2025 applied strict methodologies: replies were evaluated against all messages sent to new contacts, not solely those who opened an email. The average reply rate was 0.45%, with the first half of 2025 showing an average of 0.50% and the second half dropping to 0.40%, marking a 20% decline within the same year. December concluded with a low of 0.35%.

      This data does not definitively attribute the decline to AI, as factors like buyer behavior, economic conditions, targeting quality, deliverability, seasonality, and messaging play a role in performance. Additionally, the newer study employs a different denominator than prior Belkins reports, meaning its absolute reply rate should not be compared directly to earlier findings.

      What the data illustrates is that buyer response does not exhibit infinite elasticity; increasing the number of messages does not result in a proportional increase in buyer attention. Furthermore, the infrastructure is becoming less forgiving. Google recommends that bulk senders maintain their user-reported spam rate below 0.1% to avoid negative impacts on inbox delivery, with rates exceeding 0.1% potentially detrimental, and those of 0.3% or higher having an even greater negative effect.

      In simpler terms, a spam rate of 0.1% indicates approximately one complaint per 1,000 messages delivered, while 0.3% indicates three complaints per 1,000. This leaves very little margin for error.

      AI can generate countless persuasive sentences, yet a small fraction of recipients can still harm the sender's ability to reach others.

      Personalization is increasingly becoming a commodity. For years, the sales industry viewed personalization as the solution to ineffective outreach. The fundamental idea was sensible: research the recipient, highlight something specific, and confirm that the message was tailored for them.

      However, AI is swiftly transforming surface-level personalization into a commoditized concept. Almost any system can identify that a prospect has recently acquired funding, hired a top executive, launched a product, appeared on a podcast, or published a LinkedIn post