Challenge 02: AI with a clear vision

Webinar: Achieve ROI with AI agents Go to next challenge
AI-generated image showing the paradox of a pair of glasses with no practical use to highlight how AI can be ambiguous without proper use

Say “yes” to AI strategically

Invest in and focus on where AI can have the most impact

The average company is now using 254 AI-enabled apps1—including apps not sanctioned by the business. Executives are acknowledging the sprawl and attempting to control it, as 84% of business leaders say that there are too many AI tools2.

AI and AI agents are most valuable when intentionally integrated deeply into your organizations, working across the tools used every day to complete complex task.
Focus on creating an open, orchestration layer that communicates and connects across your business data, workflows, and tools. This will effectively break siloes and create more value. A unified coordinated ecosystem will unlock a whole new level of what's possible.

Surreal blue watering can with multiple spouts making it is use impractical
View of a contemporary office environment where three individuals are engaged in a collaborative discussion.
We have to continuously evaluate how AI goes into an environment and the potential impact it will have

Matt Lyteson

CIO, Technology Platform Transformation

IBM

AI agents and assistants

When to say “Yes” to AI agents or assistants

While the terms AI assistants and AI agents are often used interchangeably, AI assistants typically handle repetitive, well-defined tasks, whereas AI agents operate with greater autonomy—using reasoning to pursue outcomes with minimal human intervention.

Here are 6 questions to help you decide where to get started with AI agents or AI assistants.

Surreal white computer mouse with metallic spikes with no pratical use
6 questions to help you get started Question #01

Which repetitive, well-defined or low value tasks are your employees spending time on?

 Question #02

Which decision-making processes follow predictable rules or patterns and can be improved or made faster with data-driven insights?

 Question #03

Are there workflows that span multiple systems or departments and suffer from handoff delays or manual coordination?

 Question #04

What high-impact goals can be accelerated if a system could act independently within defined guardrails?

 Question #05

Are there areas where customer service or employee satisfaction is lagging due to slow response times or inconsistency?

 Question #06

Do you have access to the data needed to support AI agents in this area—and is it trustworthy?
Black polka-dot umbrella with perforated design and curved handle

Wondering where to start?

If you’re wondering where to start, know that client support and customer service, human resources, digital procurement and IT modernization have all had a track record of producing quantifiable productivity gains by using AI agents and assistants.

For example, IBM’s AskHR has automated 80 different HR processes and 94% of simple tasks, such as vacation requests and pay statements. And across these various use cases IBM has realized USD 4.5 billion in cost savings.

 

Keep on reading

Challenge 03: Don’t isolate your AI Challenge 01: Find your problem first Business leaders: 3 actions to boost the impact of your agentic AI Explore watsonx Orchestrate
Footnotes

1 Marriott, M. “How Many AI Tools Are Employees Using? More Than You Think, and Often With Personal Accounts”, Harmonic (blog), 14 April, 2025
2 Adams, C. “How CIOs are planning for an AI-first future” , Canva (blog), 30 January 2024.