CIOs need to demonstrate near-term business value from AI investments and experiments.

The WSJ reported that AI is starting to bite tens of thousands of jobs, while MIT reported that just 5% of integrated AI pilots are extracting millions in value. Other ROI challenges include insufficient investment in change management (86%) and the belief among some organizations that AI agents are overhyped (27%).

AI efficiencies in IT don’t have to be realized through headcount reduction. If you want ROI from AI in IT, CIOs should improve developer productivity and quality, thereby increasing agile team velocity and delivering backlog results faster.

Muhammad Alam, a member of the SAP SE Executive Board and SAP Product & Engineering, provided this answer to my question about CIOs under fire. I met Muhammad and other IT executives at SAP TechEd in Berlin last week. 

The answer came during a fireside chat between Mr. Alam and Alex Kantrowitz, founder of Big Technology Podcast & Newsletter. They cited countermetrics from the Wharton Human AI Research, which showed that 74% of enterprises reported moderate to significant ROI from their AI investments. (See here for details and download the report.)

Whose data is accurate? It doesn’t matter. CEOs and Board Members read the headlines, and CIOs have until their next board meeting to update their AI strategies on both short and long-term business value and ROI from AI investments.

SAP’s AI ROI is embedded in SAP BTP

Development teams use code generators to accelerate application development, and some are experimenting with vibe coding to develop new ones. Those low-level approaches to developing business workflows and applications often require significant effort to connect to data, business processes, and user experiences.

SAP customers have an alternative with SAP’s Business Technology Platform (SAP BTP). Perhaps Muhammad and the SAP team read Driving Digital, where, in 2017, I told CIOs to develop customizations outside the ERP. SAP has Rise and Grow versions of S/4Hana to help CIOs migrate to a clean core without customizations.

But business is not one-size-fits-all. SAP’s solution for customers is to customize workflows and to extend Joule AI Agent’s capabilities in SAP BTP. The platform includes low- and pro-code app development, automation, integration, data cloud, analytics, and AI. BTP is also a very open platform where developers can

• Select AI models from Anthropic, Gemini, Mistal AI, and OpenAI.
• Develop in Visual Studio, Claude, Cline, CodeX, Cursor, or Windsurf.
• Zero-copy connect to data sources in Databricks, Snowflake, and Google Cloud.

SAP BTP announces  AI Innovations

Some of the innovations reviewed at SAP TechEd directly address how IT can achieve near-term ROI from AI. (See à for the ROI opportunity.)

• SAP-RPT-1 – A relational, pre-trained, transform model that can help data scientists consolidate multiple ML models. à The ROI opportunity includes reducing model debt and simplifying the number of independently managed ML models in production. SAP claims the model is 50x faster, consumes 50,000x less energy, and improves prediction quality by 3.5x.
• Joule Agents – 40+ agents will be available by year’s end in finance, procurement, supply chain, human resources, and other business process transformations. SAP also announced extensible AI agent capabilities, so developers can start with a ready-to-use Joule Agent, extend its knowledge, and align with enterprise AI governance requirements. à The ROI comes from developing proprietary AI agents using extendability and low-code tools, rather than custom development.
• AI governance with LeanIX for cataloging AI agents and agent mining in SAP Signavio – These will help IT departments scale how they deploy AI across the organization and measure its impact. à Cataloging provides a central tool for engaging employees and accelerating adoption. Agent mining will help trace AI decision-making paths, flag model drifts, and enable tracking KPIs.

For larger enterprises, SAP BTP enables modernizing applications, consolidating platforms, and delivering agentic AI capabilities faster. The platform may be an even bigger game-changer for smaller enterprises, where innovation, automation, and AI can be hard to achieve, either because of the required skills or the trapped costs of running multiple siloed legacy systems.

“Smaller enterprises don’t have large IT departments,  and sometimes they don’t even have developers,” says Michael Ameling, president of SAP BTP and member of the Extended Board. “We simplify and make it easy for them to get started with their first automation and AI use cases. Seeing is believing when they build an app or create a simple AI agent with a low-code experience. We take away the burden of operating all these IT systems and instead focus on improving processes and gaining a competitive advantage.”

Smart platforms, partners, and advisors enable ROI

AI is already driving the next wave of digital transformation. Finding ROI requires identifying force-multiplier use cases that scale both growth and efficiency, and having a change management plan to ease adoption.

However, the speed and simplicity of developing support AI capabilities also factor into the ROI equation. Collaborating with advisors, standardizing platforms, selecting co-creation partners, and developing efficient learning programs are key enablers of an ROI that’s faster and easier to achieve.