Train Managers Before Implementing AI to Ensure Success

Businesses are quickly adopting generative AI (GenAI) without establishing adequate governance systems, potentially leading to serious quality and compliance issues. Implementing AI requires more than just a working knowledge of the technology; it demands comprehensive preparation and training, especially for managers and executives.

Survey Findings on AI Readiness

A recent survey by SAS of 1,600 IT decision-makers reveals a significant gap in understanding and readiness for AI implementation among senior tech leaders. An overwhelming 93% of senior tech decision-makers admit they do not fully understand GenAI or its potential impact on business processes.

Executive Familiarity with AI

The survey highlights the urgent need for executive education on AI. Less than half (45%) of CIOs and just over a third (36%) of CTOs consider themselves "extremely familiar" with GenAI adoption in their organizations. Alarmingly, only 13% of chief digital officers and a mere 4% of heads of IT or Information Systems claim a high level of familiarity with AI, with only 2% of IT managers or directors reporting the same.

Training and Governance Gaps

The survey reveals that only 7% of organizations provide a high level of training on overall AI governance and monitoring, and only 15% do so for generative AI. This is crucial as 75% of respondents express concerns about data privacy and security with GenAI.

Challenges in AI Implementation

Several challenges could hinder AI implementation. Only 5% of organizations have a reliable system to measure bias and privacy risk in large language models. Additionally, 42% are considering developing in-house capabilities for privacy risk detection, and 32% for bias detection. Continuous automated monitoring of generative AI implementations is present in only 29% of organizations, and only 25% conduct regular manual audits of AI output.

Integration and Tool Challenges

Many companies struggle to integrate AI with existing processes and systems. Nearly half (47%) of decision-makers report lacking appropriate tools for GenAI implementation. Common issues include:

• Effective utilization of public and proprietary datasets: 48%

• Absence of appropriate tools: 45%

• Transitioning GenAI from concept to practical use: 42%

• Compatibility issues with current systems: 39%

In-House Expertise and Skill Gaps

In-house AI expertise is in high demand, with 51% of organizations concerned about insufficient internal skills to use the technology effectively. About 39% of respondents identify a lack of internal expertise as a significant obstacle to implementing GenAI.

Mandates for Successful AI Projects

The survey outlines key mandates for successful AI projects:

• AI Integration: Seamlessly integrate GenAI models into decision workflows and existing business processes using decision flow tools.

• Data Protection: Ensure user privacy and security with robust data quality measures, including synthetic data generation, data minimization, anonymization, and encryption.

• Trustworthy and Explainable Results: Apply natural language processing techniques to preprocess data, explain generated output, minimize hallucinations, and reduce token costs.

• Enhanced Governance: Use built-in workflows to validate the entire lifecycle of large language models, from regulatory compliance to model risk management.

Proving ROI and Practical Use Challenges

Predicting or calculating ROI for GenAI is a challenge for more than a third (36%) of IT decision-makers. Almost half (47%) face difficulties in transitioning GenAI from concept to practical use. Additionally, 39% of organizations do not have a GenAI usage policy in place.