System Alphas

The AI value gap: what the research says, and what to do about it.

Almost every company now uses AI. Very few get real financial value from it. We pulled the latest findings from McKinsey, MIT, Gartner, BCG, KPMG, Wharton and Bloomberg, and added what each one means for a growing business.

7 sources2023–2026 research8 min readBy System Alphas
1. Adoption is not the edgeNearly 9 in 10 companies use AI, but only about 6% see significant financial gains.
2. Most pilots stallMost pilots show no P&L impact, and over 40% of agent projects are expected to be canceled by 2027.
3. Winners redesign the workLeaders redesign workflows, add controls and measure growth, not just hours saved.
McKinsey · 2026MIT NANDA · 2025Gartner · 2025BCG · 2026KPMG · 2026Wharton · 2026Bloomberg · 2023
Finding 01McKinsey & Company · August 20266%of companies are AI high performersThe state of AI in 2026: On the road to ROI

What the research found

  • Nearly 9 in 10 organizations now use AI regularly in at least one function.
  • Only about 6% are “high performers” that see significant financial gains, unchanged from 2025.
  • 37% report any EBIT impact from AI at all.
  • About 75% of high performers have redesigned workflows around AI, versus about 25% of other respondents.

What it means for you

Using AI is no longer an advantage. Redesigning the work around it is. The gap is not access to models, it is how the workflow changes.

How we apply it

Our AI workshop starts with the workflow, not the tool.

See AI Workshops Read the source
Global survey of 1,719 respondents in 97 countries, May–June 2026.
Finding 02MIT NANDA · August 202595%of enterprise gen AI pilots show no measurable P&L impactThe GenAI Divide: State of AI in Business 2025 (reported by Fortune)

What the research found

  • Based on 150 leader interviews, a survey of 350 employees and 300 public AI deployments.
  • Tools bought from specialised vendors succeed about 67% of the time; internal builds about a third as often.
  • The biggest returns come from back-office automation, while over half of budgets go to sales and marketing.

What it means for you

Pilots rarely fail because of the model. They fail on integration, focus and ownership. Partnering with specialists and starting in the back office raises the odds.

How we apply it

We start with one back-office workflow and connect it to your real systems.

See AI Automation Read the source
Fortune article by Sheryl Estrada, August 18, 2025.
Finding 03Gartner · June 202540%+of agentic AI projects will be canceled by end of 2027Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027

What the research found

  • Gartner cites escalating costs, unclear business value and inadequate risk controls.
  • It estimates only about 130 of the thousands of “agentic AI” vendors are real; the rest is “agent washing”.
  • By 2028, at least 15% of day-to-day work decisions will be made autonomously by agentic AI, up from 0% in 2024.
  • By 2028, 33% of enterprise software applications will include agentic AI, up from under 1% in 2024.

What it means for you

Agents are arriving fast, but most projects die on scoping and guardrails. Clear value, cost limits and human review have to be designed in from day one.

How we apply it

We build agents with guardrails, cost limits and human review built in.

See AI Agents Read the source
Gartner press release, June 25, 2025.
Finding 04Boston Consulting Group · September 20262.4xrevenue growth for “future-built” companies vs laggardsThe Formula for Agentic AI Value (BCG Applied AI Index 2026)

What the research found

  • Survey of 1,330 CxOs and senior leaders.
  • 7.5% of companies are “future-built”, with 2.3x shareholder return and 2.8x EBITDA growth versus laggards.
  • AI spend rose from about 1.7% of revenue in late 2025 to 3.3%.
  • 42% expect to give AI agents real decision-making authority by 2030, but only 5% have the full set of controls today.

What it means for you

The gap between AI leaders and laggards is widening, and governance is now the bottleneck, not ambition.

How we apply it

We pair every build with the controls to run it safely.

See our process Read the source
BCG press release, September 30, 2026.
Finding 05KPMG · August 202649%of leaders scaled back AI agents because costs outweighed benefitsGlobal AI Pulse, Q2 2026 (reported by Forbes)

What the research found

  • Survey of 2,145 senior leaders across 20 countries.
  • 79% rank AI a top investment priority.
  • Only 53% have AI cost dashboards, and about a third cite limited understanding of AI costs, such as token pricing.

What it means for you

Agents need cost visibility from the first day, or finance will switch them off.

How we apply it

Every agent we ship comes with usage and cost tracking.

See AI Agents Read the source
Forbes article by Sandy Carter, August 9, 2026.
Finding 06Wharton Human-AI Research & GBK Collective · October 20262.7xmore departments see strong AI ROI when they track growth, not just hours savedThe Adaptive Advantage: Who’s Winning with AI and How (reported by Forbes)

What the research found

  • 35% of departments that track a growth metric report strongly positive AI returns, versus 13% that track efficiency alone.
  • 67% of departments use AI agents, but only 17% let them finish tasks without step-by-step approval.
  • Only 7% of organizations meet all three of the study’s success markers.

What it means for you

Measure AI by revenue and growth, not only by hours saved. What you measure shapes what you get.

How we apply it

We agree a business metric with you before we build.

Why System Alphas Read the source
Forbes article by Gabriel Alin Zainescu, October 13, 2026.
Finding 07Bloomberg Intelligence · June 2023$1.3Tin generative AI revenue expected by 2032ChatGPT to Fuel $1.3 Trillion AI Market by 2032 (Bloomberg)

What the research found

  • Up from about $40 billion the year before the report.
  • The sector could grow about 42% a year over ten years.

What it means for you

The market is moving whether or not you do. Early, focused adopters compound their advantage.

How we apply it

Start small, prove value, then scale.

See our process Read the source
Bloomberg News, June 1, 2023.

Sources

  1. McKinsey & Company, The state of AI in 2026: On the road to ROI, August 2026. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
  2. MIT NANDA, The GenAI Divide: State of AI in Business 2025 (reported by Fortune), August 2025. https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/
  3. Gartner, Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027, June 2025. https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027
  4. Boston Consulting Group, The Formula for Agentic AI Value (BCG Applied AI Index 2026), September 2026. https://www.bcg.com/press/30september2026-ai-starting-to-pay-off-companies-generate-value
  5. KPMG, Global AI Pulse, Q2 2026 (reported by Forbes), August 2026. https://www.forbes.com/sites/sandycarter/2026/08/09/kpmg-says-nearly-half-of-executives-pulled-back-ai-agents-over-cost/
  6. Wharton Human-AI Research & GBK Collective, The Adaptive Advantage: Who’s Winning with AI and How (reported by Forbes), October 2026. https://www.forbes.com/sites/gabrielalinzainescu/2026/10/13/strong-ai-roi-is-nearly-3x-more-common-when-firms-track-growth/
  7. Bloomberg Intelligence, ChatGPT to Fuel $1.3 Trillion AI Market by 2032 (Bloomberg), June 2023. https://www.bloomberg.com/news/articles/2023-06-01/chatgpt-to-fuel-1-3-trillion-ai-market-by-2032-bi-report-says

Figures are quoted from the publishers listed above; follow the links for full methodology. System Alphas is not affiliated with these organizations. Sections titled “What it means for you” and “How we apply it” are our own analysis.

Be in the 6%

Find the workflow where AI pays back first.

Our AI workshop maps your workflows and scores each AI opportunity, so you start where value is clearest.