1. AI is Everywhere. Workflow is not affected.
in enabling Agentic AI workflows across business teams, I kept seeing the same problem. Teams have told me they are excited about AI. Then I watched their day-to-day: workflows from different data sources, multiple rounds of Excel copy-paste, and manual hand-outs before anything reached the test site. The most important intelligence was not careful to enter into any system. It lives in people's heads, unwritten, spread across emails, slide decks, and discussion threads.
People have been using AI, but the work itself is still in the old way.
I kept hearing the same solution proposed again and again. When teams talk about AI, someone says: let's put a nice chatbot on top of our files and data, and everything will be fine. I spent weeks sitting down with teams to dig deeper into their current processes. It's slow, sometimes quite painful, and it's a necessary step that unlocks the value of AI. And I'm sure these situations are no exception.
People love to discuss AI adoption at a higher level. But what I want to share may sound difficult to many people. AI only becomes business value when it reaches your products and business processes. The truth is that before introducing more AI tools and agents, you need to redesign and improve the workflow.
Recent research points to the same conclusion. McKinsey's “Talent to Value” work shows that the value of AI is growing exponentially with integrated human and agent systems, and cites nearly 900 Johnson & Johnson GenAI use cases, where 80% of the value is from 10% to only 15% of systems. BCG's 2026 AI Radar adds a CEO lens: companies expect AI usage to nearly double by 2026, and nearly all CEOs believe AI agents will generate measurable returns this year. Microsoft's 2026 Work Trend Index breaks down how work is done: its most advanced AI users are using multi-step workflow agents, rethinking workflows, and creating shared AI standards across their teams.
Your AI strategy no longer has to start with a list of use cases in a cylinder. Instead of measuring the reach of AI tools, it should start with a value map, then identify the key workflows, human judgment required, agents that can improve performance, and the operating system that proves whether the setup all works.
2. Start with Your Business Value
The right starting point for an AI enablement strategy is to identify where AI can create disproportionate benefits in cost, growth, innovation, or business model expansion.
BCG's 2026 AI Radar shows that companies expect the use of AI to nearly double by 2026, and nearly three-quarters of CEOs say they are key AI decision makers in their organization. The same study says that almost all CEOs believe that AI agents will generate measurable returns by 2026.
Johnson & Johnson's experience is the strongest case here. Extensive testing across nearly 900 use cases helped the company learn, but the impact came from reducing resources to 10% to 15% of efforts that generate the most value.
The right starting point for an AI enablement strategy is to identify where AI can create disproportionate benefits in cost, growth, innovation, or business model expansion.
- Where AI can reduce costs.
- Where AI can improve revenue, profit, or customer experience.
- Where AI can support a new product, service, or business model.
Then prioritize by asking the team one question: what 10% of our AI work can create 80% of business value?
3. Redesign Work: From the Role of the Individual to the Human Agent System
The old question was: who is right for this role? A better question is: what parts of this workflow should be human, what parts should be managed by agents, and where should human judgment remain in control?
McKinsey's agents for growth article shows organizations create greater value when agents develop end-to-end processes instead of isolated tasks. Microsoft's 2026 WTI reaches the same conclusion: AI-enhanced users are already using multi-step workflow agents and rethinking how work is done.
A customer service agent that only helps employees write quick responses improves one step. A better workflow predicts problems, causes access, routes exceptions to people, and closes the loop with personalized fixes. That's a system design decision, and no deployment of AI tools does it for you.
4. Redefine Talent: Top AI Users Are Workflow Designers
The most important workers in an AI workplace are not always the people who write the most information. They are people who can focus on the right problem, define the current process, identify weak handoffs, implement and test AI solutions, and make improved work up to others.
PwC's 2026 Global AI Jobs Barometer shows why this level of talent is rising. Jobs requiring AI skills are growing eight times faster than the overall job market, and the average salary paid for AI skills has reached 62%.
So find your biggest AI users out there. Give them authority over a workflow instead of a side project. Ask them to write about how they worked, where AI helped, where it failed, and what other teams could reuse.
Opportunity is much greater than individual productivity. The best big AI users can help a company redesign how work gets done.
5. Educate the Executive Team Before Measuring More AI Agents
While senior leaders are sponsoring many AI pilots, they may not have a consistent way to determine which tasks are appropriate for AI agents, which talent should be reassigned, and which business metrics they need to measure impact.
This is now a CEO-level performance issue. BCG's AI Radar shows 72% of CEOs say they are key AI decision makers, and half believe their career depends on getting AI right.
You need to keep AI from being a fragmented creative portfolio out of control.
- Which AI projects generate measurable business value?
- Which projects should be stopped?
- Which workflows need to be redesigned before adding more tools?
- Which leaders own business results?
- What risks require governance, auditing, or human review?
6. Measure Business Impact
The performance of an AI agent should be evaluated in terms of decision quality, reliability, speed, and cost. People should be evaluated for business impact, AI workflow development, behavioral use, and team collaboration.
The governance gap is real. Deloitte's 2026 State of AI in business research found that only 21% of surveyed organizations have a mature management model for autonomous AI agents, while nearly 80% lack mature capabilities such as decision parameters, real-time monitoring, and audit methods.
An AI agent can speed up one task while slowing down a customer's review, approval, or maintenance. The metric of success is the overall result of the workflow: faster decisions, fewer errors, lower costs, better customer experience, and stronger personal accountability.
Use three layers of measurement:
- AI agent metrics: accuracy, reliability, speed, cost, quality of escalation.
- Human metrics: business judgment, workflow improvement, behavioral implementation, collaboration.
- Business metrics: cycle time, decision quality, customer impact, operating costs, continuous improvement.
7. Final thoughts
Before your next AI review, ask one question: are we buying more AI, or are we redesigning the work that would produce a better business outcome?
First, stop asking for too many AI use cases and start identifying a few value pools that are worth investing in. Second, redesign one critical workflow by deciding who the people should be, which agents to manage, and where human reviews are needed. Third, revise the management system so that the value of AI is measured in terms of business results, workflow quality, and appropriate execution.
This is how you save time and budget while keeping AI adoption alive. It's also a test of whether your AI strategy is ready for business execution.
