Enterprise technology leaders are waking up to a blunt reality: enthusiasm for artificial intelligence is outrunning the fundamentals that make it useful. According to a recent Precisely study, many data and analytics chiefs overestimate their readiness for AI; gaps in data quality, governance and infrastructure are forcing organisations to push pilots into production before the foundations are in place, turning ambitious projects into costly experiments rather than durable capabilities.

Data foundations under scrutiny

That disconnect helps explain why conversations about AI are shifting from model performance to the plumbing around data and control. Precisely’s findings underscore that without reliable, well-governed and connected datasets, AI initiatives struggle to deliver measurable business value, leaving firms exposed to waste, poor decisions and regulatory risk.

Identity as the control plane

Alongside data integrity, identity is emerging as the central control plane for the next phase of AI. Ping Identity has been urging organisations to rethink identity and access management for a world of agentic AI, arguing that agents must be authenticated, given tightly constrained privileges and tracked across their lifecycle to ensure actions remain accountable and auditable. The firm frames agentic identities into distinct categories, personal assistants, consumer-facing digital aides, workforce digital assistants and autonomous digital workers, each requiring different trust models and oversight.

Security tools evolve for AI agents

Security vendors are responding with tools designed to secure both people and non‑human actors. 1Password’s Unified Access platform, for example, is positioned as a way to manage credential use by AI agents as well as by employees, offering visibility and control across environments that traditional login methods cannot achieve. By treating agents as first‑class identities, such platforms aim to convert a source of risk into an operational asset.

Governance and identity as prerequisites

Some vendors are building more specialised identity offerings for agentic environments. Ping Identity’s Identity for AI initiative seeks to provide visibility, governance and privilege controls specifically tailored to autonomous agents, reflecting a broader industry recognition that governance and identity are prerequisites for scaling AI safely and effectively. Together with the data integrity challenges highlighted by Precisely, these identity solutions point to a simple operational truth: models alone are insufficient; organisations must pair robust data foundations with rigorous identity and access controls to realise dependable, scalable AI.

Understanding AI and Automation is a problem throughout the industry. Join us at the Reward and Payroll Summit for our curated session to help you get the most out of AI - Click here for more information