Only five percent of companies achieve measurable added value from AI investments

Artificial intelligence is fundamentally changing the world of work, but the economic benefits are not materializing for most companies. A study by the Boston Consulting Group of over 1,250 companies worldwide reveals a drastic gap: while a small elite of five percent of companies are recording significant increases in turnover and cash flow, 60 percent are generating hardly any measurable value despite substantial investments. A further 35 percent are in the process of building up, but are moving too slowly.

The growing value gap

The leading companies - described by BCG as «future-built» - are already achieving 1.7 times higher sales growth and 1.6 times higher EBIT margins than the lagging companies. They plan to spend 26 percent more on IT and devote up to 64 percent more of their IT budget to AI. These investments create a self-reinforcing cycle: Revenues are reinvested in new skills and technologies, continually widening the lead.

The role of agentic AI - autonomous AI agents that can observe, plan and act - is particularly noteworthy. This technology, which was barely mentioned in 2024, already accounts for 17 percent of the total AI value and is expected to increase to 29 percent by 2028.

Skill data as a decisive lever

Cornerstoneondemand, a leading provider of talent and learning management solutions, identifies skills as a key performance indicator for 2026. Without transparent skill profiles and linked data, AI cannot provide context and therefore cannot make meaningful recommendations. Current figures show the pressure to act: 35% of companies see clear skills gaps, 55% report selective deficits.

The demand for AI skills has risen by 245 percent - the most sought-after skill worldwide. At the same time, the importance of human skills is growing: The demand for emotional intelligence rose by 95 percent, as technical professions increasingly require human skills.

Five strategies of the successful

The BCG study identifies a clear pattern among the successful companies. They are pursuing a CEO-driven, multi-year AI vision with explicit revenue and cost targets. Nearly 100 percent of future-built organizations report deeply engaged leadership teams, compared to only eight percent of laggards.

Secondly, they fundamentally redesign workflows instead of just automating them. 70 percent of AI potential is concentrated in core functions such as sales, marketing, production and research. IT recorded an increase of six percentage points to 13 percent of the total AI value.

Thirdly, they implement an AI-first operating model with shared responsibility between business and IT. These companies are 1.5 times more likely to establish this shared ownership.

Fourthly, they secure the necessary talent through aggressive further training: Over 50 percent of employees receive training, compared to only 20 percent for laggards. Learning and work are increasingly merging, with learning content being embedded directly into the work context.

Fifth, they are building a technology architecture that is fit for purpose. More than half of future-built companies are working with a single enterprise-wide data model, compared to only four percent of stagnant companies.

HR and IT move closer together

Cornerstoneondemand is observing a closer integration of HR and IT. While HR has structured information such as skills and performance data, IT provides the work context from collaboration and productivity environments. A survey by Nexthink of 1,100 IT managers shows that 92% believe that this integration increases productivity and employee satisfaction.

The boundaries between technical and non-technical professions are becoming blurred. AI skills are expected in almost all roles, while the demands on empathy, communication and judgment are increasing.

High security risks

Despite the potential, 72 percent of companies report unmanaged AI security risks. Successful companies focus on designing safeguards and guardrails without compromising performance. They solve problems such as hallucination, bias and data misuse that often block laggards.

The most important insight: AI agents do not work without humans. Their performance depends on strong human orchestration in redesigned roles. The best organizations configure workflows to combine autonomous agents with human oversight.

Time is running out

Technology is developing at a rapid pace. The 10-20-70 rule for technology transformations remains valid: 70 percent of the strategic focus should be on people and processes, 20 percent on technology and 10 percent on algorithms. Most obstacles concern people, organization and processes.

Companies that have not yet started must act now. Future-built companies are pulling away, the value gap is widening and catching up is becoming more difficult with each passing week. The playbook is clearly defined and available to all - but time is running out.

Further information about Cornerstoneondemand