AgentsBusiness & Market 🇷🇺 27.07.2026 17:01

Do human salespeople still matter? The reality behind AI agent promises

The hype around fully autonomous AI sales agents often clashes with reality: many such implementations fail within a year, and live sales representatives still outperform AI agents in revenue generation by a factor of 2.6. However, a new paradigm—agent-qualified leads (AQL)—is emerging as a more effective alternative, focusing on qualification rather than mass outreach.
The enterprise AI market is split between pragmatists who resist AI agents and those who over-rely on them. Many companies that replaced human sales teams with autonomous agents reverted to hybrid models within six months, as the cost savings failed to materialize. A/B tests from Dashly show live SDRs generated 2.6 times more revenue than autonomous agents. The startup 11x raised $74 million for its AI SDR Alice but faced 70-80% churn and a legal warning from ZoomInfo after poor results. Autonomous agents often flood prospects with emails—agents send 7400 emails per month vs. 1150 by humans, but reply rates drop from 4.7% to 2.9% and spam rates increase. However, a new approach called Agent-Qualified Leads (AQL), introduced by Docket, uses AI to qualify leads through dialogue rather than post-hoc analysis. In a pilot, AQL agents led to 7 times more demo meetings and increased conversion from 13% to 40%. Gartner predicts AI agents will outnumber human sales staff 10:1 by 2028, but less than 40% of employees confirm AI tools improved their productivity. The value of AI agents is limited by the skill of the human operator: experienced workers become three times more productive, while weak ones simply generate spam.
Сокращения
B2B = Business-to-Business
SDR = Sales Development Representative
MQL = Marketing Qualified Lead
SQL = Sales Qualified Lead
BANT = Budget, Authority, Need, Timeline
AQL = Agent-Qualified Lead
ICP = Ideal Customer Profile
Source: Habr — хаб ИИ — original
Our earlier posts on this topic ↓
Fresh news