ACN Market Brief: AI Is Changing What Organizations Need From Us
October 12, 2026
There’s an uncomfortable gap in the AI conversation. People are producing things faster, yet organizations still struggle to explain what improved. Meanwhile, consultants, coaches, and team leaders are being asked to help make the investment worthwhile.
This week’s market developments point toward some practical opportunities for our community. Buyers want clearer evidence of value. AI agents are raising questions about responsibility and decision making. Learning needs to become part of everyday work.
For those of us working in organizational agility, these are familiar problems taking new forms.
Faster Work Still Has to Become Better Results
EY’s October survey of 1,200 CEOs found that 50 percent see AI as their biggest driver of material productivity gains. Yet 23 percent struggle to turn those gains into measurable financial outcomes. Read EY’s findings.
That gap deserves attention. Saving someone an hour doesn’t automatically improve delivery. The work might still wait three days for approval. A team might generate more options while taking longer to choose one. Faster production can even create a larger review burden for someone downstream.
BCG’s October 2 analysis adds another complication: AI spending is spreading across the business, making its costs and returns harder to understand. Read BCG’s analysis.
For our community, the useful move is to follow one piece of work from beginning to end. Where did AI help? Where did the saved time go? Did the customer receive something sooner or better? Did errors, rework, or costs change?
These questions belong in team conversations and leadership reviews. They also give coaches and practitioners a concrete way to contribute. Helping an organization understand its flow of work may prove more valuable than helping it count AI users.
Consulting Demand Is Returning Unevenly
Recent financial results suggest that organizations are still buying outside help, although the recovery remains uneven.
Accenture reported $22.17 billion in quarterly bookings on October 1, including $9.4 billion in consulting bookings. TCS reported annualized AI revenue of $3.1 billion on October 8, while its overall constant currency revenue grew just 0.5 percent from the previous quarter. EY reported 4.7 percent annual global revenue growth and substantial investment in employee development.
Sources: Accenture results, TCS results, and EY results.
These figures describe large firms. They don’t tell us that independent consulting or Agile coaching demand has recovered. They do suggest that spending is available when buyers see a connection to a pressing business need.
My reading is that practitioners will benefit from making that connection much clearer. “I help organizations become more Agile” leaves a buyer with considerable interpretation to do. “I help teams reduce delays between decisions and delivery” gives them something they can recognize.
For people reconsidering their careers, this matters too. A familiar job title may attract fewer openings while the underlying problems remain. Look for where your experience helps with stalled adoption, unclear ownership, slow decisions, excessive coordination, or learning that never spreads beyond one team. Then explain that contribution with examples from your work.
AI Agents Need Clear Boundaries and Accountable People
A new vendor sponsored survey of more than 540 organizations reported that 57.3 percent were piloting or running AI agents, while only 8.4 percent believed their data was reliable enough for production. Just 18 percent reported formal AI accountability frameworks. Read the report coverage.
Those percentages are directional evidence, rather than a definitive picture of every enterprise. Still, they raise a useful question: are organizations giving agents responsibilities before agreeing on the conditions under which they should act?
An agent that drafts a summary presents one kind of challenge. An agent that changes a customer record, approves a transaction, or triggers work across several systems presents another.
Teams need to know who owns the result, what the agent can access, when it must seek approval, and how someone can interrupt or reverse its actions. “A human will review it” is incomplete unless that person has the time, information, and authority to perform a meaningful review.
This is work our community can help facilitate. Bring the people doing the work together with those responsible for the technology and its consequences. Walk through an actual scenario, including what happens when the result is wrong.
The model Human led → AI assisted or executed → Team verified is a useful starting point. Its value comes from making each responsibility explicit.
Learning and Review Have to Continue After Deployment
Employees are already finding their own ways to use AI. An October 7 Reuters analysis argues that organizations should assume some of that use happens outside approved tools and respond with practical guidance, suitable alternatives, and training relevant to people’s roles. Read the analysis.
At the same time, vendors are building continuous evaluation and improvement into their agent platforms. Kore.ai’s new Autoloop offering, for example, evaluates agents against task completion, accuracy, safety, and cost, then reverifies proposed changes. This is a product announcement, but it signals growing attention to what happens after deployment. Read the announcement coverage.
For leaders and coaches, the implication is practical. A launch and a completed course won’t tell you whether people are using AI well six weeks later. Work changes. Models change. People discover shortcuts, encounter failures, and quietly develop habits nobody else sees.
Communities of practice can help make those discoveries visible. Give people room to bring a real task, show what they tried, and discuss the result. Include the attempts that failed. Those often reveal more than a polished demonstration.
What Deserves Action
There’s no need to wait for a complete AI strategy to begin useful work. Pick a current workflow and bring the people involved into the conversation.
Establish a baseline. Record how long the work takes, where it waits, and what makes an outcome useful.
Clarify responsibility. Agree on what AI can do, what requires verification, and who remains accountable.
Create time to learn. Let people practice with real work and share successful approaches and failures.
Review the whole result. Examine quality, delivery time, rework, cost, and customer impact together.
Our community has spent years helping people learn through experience, make decisions with better evidence, and adapt how they work. The current market gives us fresh reasons to apply those skills.
Start with a problem people recognize. Make a small change. See whether it helped. That’s still a worthwhile contribution, whatever technology sits in the middle.
Keep exploring
Is AI Really Making Software Cheaper? — A community discussion about AI costs, software delivery, and the work behind the promises.
Is AI Making Organizations More Effective, or Just Busier? — A closer look at the gap between faster activity and better organizational outcomes.
Agile Coaching Is Declining. That Doesn’t Mean You Are — A perspective for coaches and practitioners reconsidering their place in a changing market.
ACN is a community conversation. Bring your experience, questions, or curiosity to a future live session. Listening is welcome, too.
Help keep the conversation going
If ACN brings value to your work, consider supporting our work. Your contribution helps us keep these community conversations going. Sharing an episode with a colleague helps, too.
Comments