AI Management Survey Highlights Gaps in Business Review Workflows

The short answer: the tools arrived before the process

A survey overview published by monday.com, conducted by Nielsen, asked 1,701 directors and senior managers across seven countries how AI is changing their jobs. Most respondents use AI daily and believe in its value, yet only 29% describe their organization's AI infrastructure as centralized and fully optimized (monday.com, n.d.). The responses suggest a gap between individual adoption and organizational support.

TechCrunch reported on the survey on October 21, 2026, adding detail from an interview with Michal Miller Levi, head of market research and customer insights at monday.com. She described managers who find AI useful in their own work while organizational processes struggle to keep pace (Iyer, 2026). For business owners, that distinction raises a practical question: who checks the output, and how does an approved result reach the right business record?

What the survey actually measured, and what it did not

Every figure in this article comes from a sponsor-funded self-report, not a controlled experiment or an independently replicated productivity measurement. Respondents were directors and senior managers in seven countries, and the public overview does not disclose sampling methods, fieldwork dates, question wording, or confidence intervals. The full report is offered through an email form, which was not submitted for this article. TechCrunch's coverage is separate reporting of the same survey, not independent replication of its results.

With that in mind, the numbers remain useful as signals about how managers describe their experience. The overview reports that 72% face an employer mandate or strong encouragement to use AI, while 29% describe centralized, fully optimized infrastructure. It also reports that 94% are enthusiastic about finding AI optimization opportunities, but 59% consider the work of finding and testing valuable uses more demanding than the time it returns (monday.com, n.d.). These responses suggest that enthusiasm and perceived implementation burden can coexist; they do not establish how much time any company gains or loses.

Where the extra checking work comes from

The survey and interview suggest that businesses should examine the review work surrounding AI output. A polished draft, quote, or customer reply still needs a person to check its substance; a workflow may need to reserve time for that check. The overview also describes roughly half of respondents using both employer-approved and unapproved tools in a typical week (monday.com, n.d.). Where that pattern applies, outputs could arrive through untracked channels, use inconsistent formats, or lack a recorded approval. These are possible operational risks, not outcomes measured by the survey.

For a small business, warning signs worth investigating include scattered outputs, unclear approval ownership, and customer records that differ from what was sent. The survey does not measure these outcomes. Our practical recommendation is to examine the review path before adding another tool: identify who checks each draft, what evidence they need, and where the approved result belongs.

What this means for your business

Detailed In Design could build an intake-to-quote workflow for a business facing exactly this pattern. Customer requests and source documents land in one place. A draft summary or quote input is prepared, then routed to a named human reviewer rather than to whoever happens to notice it. The reviewer's changes and approval are recorded, and only after approval does the system update the authorized customer-tracking record. The expected benefit is accountability, not automation for its own sake: fewer silent handoffs, a traceable decision trail, and a reviewer who knows which items are theirs.

The limits matter as much as the design. Such a system requires client permission, accessible customer-tracking and document interfaces, suitable data permissions and retention rules, clearly assigned review ownership, acceptance tests, and escalation and rollback procedures. Human review still consumes time and can miss errors, and integration and logging do not guarantee accuracy, fairness, security, or savings. This is proposed custom software, not a shipped product, a monday.com capability, or a vendor partnership, and nothing here measures a return on investment.

If your team has AI tools but no agreed place where their output gets checked before it reaches a customer, start with the smallest version of that path: one intake point, one named reviewer, one record of what was approved and when.

Iyer, R. (2026, October 21). AI promises productivity, but for managers, it's mostly a headache for now. TechCrunch. https://techcrunch.com/2026/10/21/ai-promises-productivity-but-for-managers-its-mostly-a-headache-for-now-says-new-survey/

monday.com. (n.d.). AI at work 2026: Squinting in the sun [Survey overview]. https://monday.com/gp/ai-report-2026