A Conversation with Chris Wollerman
By Jason Diamond Arnold, Director of Leadership Solutions and Executive Leadership Coach, Inspire Software
Inspire Executive Article
Overview: An AI workflow is a repeatable path that turns a defined trigger into a finished piece of work, such as a brief, an agenda, a plan, or a draft, by pulling from set data sources and following steps you decide in advance. It is different from an AI agent, which is handed a goal and reasons its own way to it. That distinction is where a real AI strategy begins. Companies are adding AI far faster than they are getting value from it. In McKinsey’s 2026 State of AI survey, only about 37 percent of organizations could attribute any earnings impact to it. Inspire CEO and co-founder Chris Wollerman argues the gap is rarely about the models. It is about how the work is organized. This conversation covers what an AI workflow is, how it differs from an AI agent, and where a leadership team should start.
Watch the podcast instead: https://youtu.be/ICeZdDO53Y4
Key takeaways
- An AI workflow is a repeatable path from a trigger to a finished artifact. An AI agent is handed a goal and decides its own steps.
- Most organizations see little measurable return from AI. The return comes from building AI into how work flows, not from scattered tools.
- Start with the process. Diagnose the repetitive, administrative work before you choose a tool.
- Build the first workflows around the rhythms of execution: the daily brief, weekly planning, and one-to-one and team meeting preparation.
- A team leader can build these. The people who know where the real inefficiencies live are often line managers and front-line staff, not the executives.
- Done well, AI workflows connect strategy to performance by holding the cadence of execution in place every week.
Why isn’t AI adoption outpacing AI results?
Almost every organization is racing to add AI. Reactions range from executives who see it as the defining advantage of the decade to employees who are wary of it, and the topic has become a cultural flashpoint. This spring, graduates booed commencement speakers who brought up AI, including former Google CEO Eric Schmidt at the University of Arizona (NBC News, May 2026). What most leadership teams share, whatever their enthusiasm, is a quieter frustration. They have adopted AI, and the business still runs about the way it did a year ago.
The numbers describe the same paradox. In McKinsey’s The State of AI in 2026: On the Road to ROI, nearly nine in ten respondents report regular use of AI in at least one business function, and 37 percent report that AI has contributed positively to their organizations, essentially unchanged from 2025. Adoption is close to universal. Return is not guaranteed.
When a CEO or an executive leadership team comes to Chris Wollerman looking for more than experimentation, the question is usually some version of the same thing. They have the tools, so how do they build a real AI strategy that changes how they operate? His answer surprises them, because it does not start with AI at all.
What the evidence says about the AI ROI gap
The most useful recent look at the gap between using AI and using it effectively comes from MIT’s Project NANDA, whose 2025 report The GenAI Divide: State of AI in Business drew on more than 300 publicly disclosed AI initiatives, structured interviews with 52 organizations, and survey responses from 153 senior leaders. Its headline finding is striking. Around 95 percent of organizations were getting no measurable return from their generative AI investments, and just 5 percent of integrated pilots were capturing real value.
The more instructive part is what separated the small share that worked. According to the research, deployments that bought or partnered for capability succeeded far more often than those built entirely in house, adoption moved faster when line managers drove it rather than a central lab, and the tools that paid off were the ones that integrated deeply into existing workflows and adapted to them over time. Back-office and administrative automation tended to return more than the flashier front-office use cases.
McKinsey’s data points in the same direction. Organizations that translate individual productivity gains into enterprise results are the ones redesigning how work flows, not simply layering tools on top. Among its high performers, nearly three-quarters now report fundamentally redesigning workflows because of AI, up from 55 percent a year earlier, while most other organizations have barely started. The learning across both studies is consistent. Value shows up when AI is built into the way work moves, and it stalls when AI is bolted on as one more thing to open in a browser tab.
Where do we begin when implementing an AI workflow?
Before you build anything, look at where the work is breaking down. That is where Wollerman begins. “They hear about the potential, but not a lot are getting the ROI yet,” he says of the companies now asking for help. So before he talks about models or agents, he studies the work itself.
“You really look at where your inefficiencies are,” he explains. His AI consulting team examines the day-to-day and repetitive tasks inside an organization, the monotonous and administrative work that keeps capable people from doing something more valuable. Sometimes the diagnosis does not even point to a problem AI can solve. “Maybe it’s just a process improvement, and some documentation and some communication improvements.” AI earns its place most reliably when it connects systems and removes administrative drag, not when it is adopted for its own sake.
That caution matters because many organizations fail in a common way. Leaders get excited about the technology, or about a strategy or a performance discipline they want to install, and the effort swells into a year-long change initiative before anyone sees a benefit. That approach is slow, and it breeds resistance before people ever experience what AI can do for them. Most people first meet these tools as a better search engine, a place to ask questions. That has value, but it is a long way from changing how the business runs. The shift that produces real return is the move from asking AI things to building repeatable work with it. That shift has a name.
What is an AI workflow?
“A workflow is a repeatable path from a trigger to a finished piece of work.” Under that definition are four parts worth understanding, because together they form a practical blueprint any team leader can use to design one.
Trigger. The workflow kicks off from a defined occasion, whether a date, a meeting, or a request. It runs when that condition is met, the same way each time.
Sources. It pulls from a defined set of inputs. In practice, that is the information already scattered across the day: emails sent and received, one-to-one and team meeting transcripts, documents in progress, open tasks, and action items that come out of those meetings.
Sequence. It follows steps you decide ahead of time. “You kind of guide it, sometimes lay out a flow chart,” Wollerman says. The order is planned, not improvised.
Artifact. It produces a finished output, “like a brief, an agenda, a deck, or a draft.” Something you can use.
What is the difference between an AI workflow and an AI agent?
The cleanest way to grasp the difference is the metaphor Wollerman reaches for. A workflow is a detailed recipe, with ingredients and steps laid out to reach a known result. An agent is more like a chef. “You just say, make me dinner. And if he’s good, it’ll be good, but it’s not consistent. It’s not necessarily repeatable.” An agent is given a goal and reasons its way to it on its own.
That autonomy is genuinely useful, and the tooling is moving fast, to the point where some platforms now run an agent that manages other agents beneath it. But autonomy and repeatability are different properties, and leaders often want the second one first. Wollerman is careful not to frame this as a choice between the two. “You really want to have both,” he says. The recommendation is about sequence. Start with the workflow, because a workflow is something you can point at a specific inefficiency and measure a return on, rather than a general-purpose capability you hope pays off. Get the recipe working, then bring in the chef where judgment and open-ended reasoning help.
Which AI workflows should you build first?
The theory gets real when you attach it to the tasks that repeat for most employees. A few of those rhythms are where Wollerman sees the fastest wins.
The daily brief. A workflow assembles the briefing a leader would otherwise pull together by hand: the overnight email, the meeting and chat transcripts, the documents in play, the open tasks, and yesterday’s action items, ready on the desktop at the start of the day. Tools such as Claude, Grok, and ChatGPT already generate strong versions of this. It is far more than using AI as a search box. It also works as a bookend. “In the morning, I had all these good intentions. At the end of the day, what did I get done? What do I still have to do?” Wollerman says. That closeout brief tends to surface what slipped during a busy day.
What does an AI-supported week of execution look like?
The execution chain. One workflow rarely stands alone. The rhythm of a real week strings them together. Monday morning, set the plan against the goals that matter most. Once the plan is set, prepare for the one-to-one with a manager, using a shared leadership language so the conversation lands better. In the afternoon, prepare for the team meetings by pulling the action items and updates worth sharing. Each of these is its own repeatable path from trigger to artifact, and together they keep people aligned through the week.
Connect the systems. Inspire is one of them. Its MCP connector lets Claude or ChatGPT read and write your live objectives, key results, and check-ins, so the Monday plan starts from the goals that matter. Much of the drag in knowledge work is moving the same information between tools and entering it twice. Many of the applications teams already use have a data connector that plugs into environments such as Claude Cowork, which removes a surprising amount of that duplicate handling and frees up the harder, more detailed workflows to tackle next.
Who builds AI workflows, and do you need engineers?
You do not need an engineering team. A team leader can set most of this up. The people who know where the real inefficiencies live are often not the executives. “Executive leaders think they know where the inefficiencies are,” Wollerman notes, “then you start to talk to either the line-level employees or the middle managers, and that’s where you really start to uncover more of what’s happening in the day-to-day.” Inspire’s own consulting approach is built on closing that gap. It gathers insight from leaders and the front line, records and transcribes the conversations, and uses AI to synthesize the findings into a few workflows that deliver value, such as hours returned per week, planning or delivery deadlines met, or other cost savings achieved through structured automation.
Chase quick wins first so the return is visible early, then move to the harder automations, and keep improving from there rather than trying to automate everything at once.
How do you measure the ROI of an AI workflow?
Before you automate a workflow, measure the manual version, because you cannot claim a return you never baselined. Capture three numbers as the work happens today: how long it takes, how often it runs, and the fully loaded cost of the person doing it. A weekly planning routine that takes a manager 90 minutes every Monday is about 78 hours a year. If an AI workflow brings that to 20 minutes, you have returned roughly 60 hours per manager per year, and you can price that against their loaded rate.
Time returned is the easiest number to defend, but it is not the only one that matters to an executive buyer. Cycle time is the next, when a board brief that used to take two days is ready the same morning. Quality and consistency is the third, meaning fewer missed action items, fewer errors, and the plan actually happening the same way each week, which you can track as a simple omission rate before and after. The fourth is strategic yield, the data the workflow creates that did not exist before, such as every one-to-one and team meeting leaving a structured record tied to objectives.
For an audit engagement, reduce each workflow to one line: baseline cost, cost after automation, hours returned, and payback period. A workflow that takes a day to build and saves an hour a week pays for itself in a few weeks, and that arithmetic is what moves a skeptical leader from interest to budget.
How do AI workflows connect strategy to performance?
Put the pieces together, and the strategic point comes into focus. The reason most AI adoption underdelivers is that it stays a set of disconnected conversations, each one useful in the moment, with little of it compounding over time. A workflow is what makes the value repeatable, and the rhythms of execution, the weekly plan, the daily focus, and the one-to-one and team conversations, are what connect that repeatable value to the goals a company is trying to reach.
Strategy produces better results when it is linked to a continuous performance management process that generates meaningful data. That link lives in the recurring cadence of planning, doing, and talking about the work. AI workflows are a powerful new way to hold that cadence in place, reinforcing the connection between strategy and performance every week rather than revisiting it once a quarter, while producing powerful new sources of data. The cost of leaving AI as a library of one-off prompts is not just a missed efficiency. It is a missed chance to make execution more consistent.
For leaders who want to get serious about using AI to build meaningful workflows, the honest starting move is a diagnosis, not a purchase. Look hard at the repetitive work, find one rhythm worth automating, build the recipe, and measure what it returns. The bigger payoff is consistency: the plan, the one-to-one, and the team meeting happen the same way every week.
Watch the Podcast:
Frequently Asked Questions
What is an AI workflow?
An AI workflow is a repeatable path from a trigger to a finished piece of work. It starts on a defined occasion such as a date, meeting, or request, pulls from a defined set of data sources, follows a planned sequence of steps, and produces a usable artifact like a brief, agenda, deck, or draft. Because it runs the same way each time, a business can count on it and measure it.
What is the difference between an AI workflow and an AI agent?
A workflow is like a detailed recipe, with the ingredients and steps mapped out ahead of time to reach a known result, so it is consistent and repeatable. An agent is like a chef you ask to make dinner. You give it a goal, and it reasons its own way there, which is flexible but less predictable. Most organizations benefit from both and often start with workflows because the return is easier to target and measure.
Why do organizations need AI workflows?
Scattered AI use rarely changes how a business runs. Adoption is now near-universal, but MIT and McKinsey studies find that most organizations see little measurable financial return, and that the value appears when AI is built into how work flows. Workflows turn one-off AI conversations into repeatable, measurable work.
How do you build an AI workflow?
Start by diagnosing where the repetitive, administrative, or inefficient work is, rather than starting with the tool. Then design the four parts:
- the trigger that starts it
- the data sources it pulls from
- the sequence of steps it follows
- the artifact it produces
Begin with a quick win so the return is visible early, then move to more detailed automations.
How do you measure the ROI of an AI workflow?
Baseline the manual process first: how long it takes, how often it runs, and the loaded cost of the person doing it. After automating, measure four things: hours returned (time saved times frequency times loaded rate), cycle time (how much faster the artifact is ready), quality and consistency (fewer missed items or errors), and strategic yield (new data the workflow creates, such as structured records tied to objectives). Express each workflow as baseline cost, new cost, hours returned, and payback period. A workflow that takes a day to build and saves an hour a week typically pays back within a few weeks.
What are examples of AI workflows in business?
Common early examples include a morning briefing that assembles email, meeting transcripts, tasks, and action items into one view, an end-of-day closeout that reviews what got done and what remains, a Monday planning workflow that sets the week against goals, and preparation workflows for one-to-one and team meetings.
How can AI improve weekly planning?
An AI workflow can pull a leader’s goals, open tasks, and recent action items into a single starting point for the week, so planning begins from an organized picture rather than a blank page. That plan can then feed the next workflows in the week, such as meeting preparation.
What is a daily brief, and how can AI generate one?
A daily brief is a consolidated view of what a leader needs to start the day, drawn from overnight email, meeting and chat transcripts, working documents, open tasks, and prior action items. An AI agent can assemble it automatically and have it ready each morning, with an optional end-of-day version that reviews progress and flags what still needs attention.
Do AI workflows replace employees?
The aim described here is to remove monotonous and administrative work so people can spend time on more valuable tasks, not to replace them. Workflows also make execution more consistent and help people buried in scattered tasks stay focused and perform.
Who should build AI workflows, and do you need engineers?
You generally do not need an engineering team. A team leader can set up most workflows using available tools and data connectors. The people who best understand which inefficiencies to target are often line-level employees and middle managers, so involving them improves what gets built.
Which business processes should you build AI workflows around first?
Start with the repetitive, high-frequency rhythms where a quick win is visible fast, such as the daily brief, weekly planning, and meeting preparation, and with administrative work that moves information between systems. Save the harder, more detailed automations for after the early wins prove the value.
How do AI workflows connect strategy to performance?
By embedding AI in the recurring cadence of execution, the plan, the daily focus, and the one-to-one and team conversations, workflows keep the goals that matter in view week after week. That turns AI from a set of disconnected conversations into a repeatable link between what a company intends and what it does, the heart of continuous performance management and strategy execution.
Have questions about putting AI to work in your organization? Reach out at inspiresoftware.com to schedule a conversation about how Inspire connects strategy to performance across the organization.
Related reading from this series
This is the eighth article in Inspire’s executive series on how AI helps leaders develop and execute a winning strategy. The earlier installments:
1. How AI Helps Executive Leaders Develop a Winning Strategy
2. How AI Helps Executive Leaders Visualize and Communicate Strategy
3. How AI Helps Executive Leaders Turn Strategy into Action Through OKRs
4. How Using AI for Planning Toward OKRs Can Help Organizations Turn Strategy into Action
5. Your OKRs, Wherever Your AI Works
6. The Execution Engine: Where Strategy Actually Gets Done
7. How AI Turns Continuous Performance Data into More Meaningful Business Reviews
8. Why Your Organization Needs AI Workflows, Not Just AI Agents (this article)
Resources
McKinsey & Company. The State of AI in 2026: On the Road to ROI. August 2026. Source of the AI-adoption and EBIT-impact figures and the workflow-redesign finding cited above. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
MIT NANDA (Project NANDA). The GenAI Divide: State of AI in Business 2025. Aditya Challapally, Chris Pease, Ramesh Raskar, and Pradyumna Chari. Source of the finding that about 95 percent of organizations see no measurable return from generative AI, with 5 percent of integrated pilots capturing value. https://nanda.media.mit.edu/ai_report_2025.pdf
NBC News. Former Google CEO Eric Schmidt booed during graduation speech about AI. May 2026. Source of the University of Arizona commencement reference. https://www.nbcnews.com/tech/tech-news/former-google-ceo-booed-graduation-speech-ai-rcna345585
About the Author: Jason Diamond Arnold (Extended)
Jason Diamond Arnold is the Director of Leadership Solutions and a Leadership, OKR, and Performance Coach at Inspire Software. Inspire is a strategy execution and performance management platform that helps organizations align goals, execute strategy, and improve leadership performance across teams.
With more than 25 years of experience in leadership development, organizational performance, and strategy execution, Jason works with executives, managers, and teams to translate leadership theory into practical systems that drive measurable business results. Through Inspire Software’s OKR framework, coaching programs, and leadership development tools, he helps organizations strengthen strategic alignment, improve employee engagement, and build high-performance cultures.
Jason’s work bridges behavioral science, leadership development, and performance technology, helping organizations move from strategy planning to consistent execution. He has coached leaders across industries including technology, retail, and professional sports, helping teams improve accountability, strategic focus, and measurable performance outcomes.
Experience and Background
Before joining Inspire Software, Jason worked as a product manager and consultant, collaborating with major organizations including Apple, Sephora, the NBA, and Verizon. His work has focused on helping organizations align leadership practices with measurable performance systems.
Jason is currently pursuing a PhD in Leadership at the University of San Diego and is a candidate for certification through the International Coaching Federation (ICF). He also serves as a Lecturer at the University of San Diego School of Leadership, where he teaches and researches modern leadership frameworks and organizational development.
With more than 1,000 hours of coaching and consulting experience, Jason specializes in helping organizations implement leadership systems that support strategic alignment, goal management, and sustainable performance improvement.
Areas of Expertise
- Jason specializes in leadership development and strategy execution, including:
- OKR Implementation and Coaching: Helping organizations implement Objectives and Key Results (OKRs) to align teams around measurable strategic goals.
- Leadership Development and Organizational Coaching: Supporting leaders in building effective leadership practices that drive accountability, engagement, and performance.
- Strategic Alignment and Performance Management: Helping organizations connect leadership behaviors, team goals, and measurable performance outcomes.
- Managerial Leadership and Self-Leadership: Equipping managers and employees with the tools to align individual performance with company strategy.
- Performance Excellence and Behavioral Leadership Science: Using research-based leadership frameworks to help organizations improve team performance and operational efficiency.
Coaching Impact
- Jason has helped organizations improve leadership performance and operational outcomes across industries.
- Examples of his work include:
- Global Retail Organization: Led leadership alignment and strategic restructuring initiatives that improved operational efficiency by 30 percent.
- Professional Sports Organization: Coached leadership teams to align career development with strategic goals, improving team collaboration and performance by 25 percent.
- Technology and Software Teams: Implemented coaching and leadership alignment programs that increased engagement and improved goal achievement metrics by 40 percent.
Leadership and Thought Leadership at Inspire Software
As Director of Leadership Solutions at Inspire Software, Jason helps shape how leadership development integrates with strategy execution technology.
He has contributed more than 100 thought leadership articles, research projects, eBooks, and podcasts focused on leadership, OKRs, and performance management. His work helps organizations combine leadership development with modern strategy execution tools to create cultures of accountability, alignment, and sustained performance.
At Inspire Software, Jason’s work focuses on aligning leadership theory, behavioral science, and technology to help organizations build high-performing teams and execute strategy more effectively.
