Datatonic says poor AI implementation and weak human-AI collaboration are slowing productivity and contributing to workforce cuts. Human-in-the-loop models are key to sustainable enterprise AI.
Poor AI implementation may be contributing to workforce reductions across enterprises, according to cloud data and AI consultancy Datatonic. The firm argues that organizations are undermining productivity, competitiveness, and efficiency by deploying artificial intelligence without integrating it into human workflows.
Datatonic’s research suggests that companies failing to embed AI into day-to-day decision-making processes are experiencing productivity slowdowns rather than gains. Scott Eivers, CEO of Datatonic, stated that the core issue is not the technology itself but how it implemented. He emphasized that AI must redesign how work gets done, warning that productivity leakage occurs when AI operates in isolation from business teams.
The company advocates for human-in-the-loop (HiTL) systems as the next phase of enterprise AI. In this model, AI delivers speed and scale, while humans retain oversight, judgment, and accountability. For example, in agent-assisted software development, AI systems generate modular code components, but human teams define requirements, validate plans, and approve outputs before deployment.
Datatonic points to finance operations as evidence of hybrid success. AI-powered document processing has reportedly reduced invoice-processing costs by up to 70%, yet finance professionals continue to approve final decisions. This partnership model, according to CTO Andrew Harding, creates sustainable enterprise value by combining governance frameworks with automation efficiency.
However, Datatonic warns that many enterprises are attempting to scale autonomous agents without adequate security controls or evaluation systems. Governance checkpoints, performance benchmarks, and compliance validation are required before autonomy can expand responsibly.
The broader implication is clear: AI adoption alone does not guarantee productivity gains. Organizations that prioritize governance, embed AI into human workflows, and build trust through structured oversight are more likely to unlock long-term value. Companies that skip these steps risk inefficiency and unintended workforce disruption.
ソース：
https://www.artificialintelligence-news.com/news/ai-workflows-need-human-in-the-loop-say-datatonic/
