Most employees using AI still treat it as a better search box. They ask for a summary or rewrite, then return to the same process.

New behavioral research from the ActivTrak Productivity Lab puts a hard number on that gap. Its analysis covered 120,620 workers across 1,009 organizations from the fourth quarter of 2025 through the second quarter of 2026.

More than 82% of AI users maintained their usage from one quarter to the next. Yet only 2% reached the most advanced stage in ActivTrak’s model, where AI is embedded consistently in a workflow rather than used for isolated tasks.

The users did record more productive time. ActivTrak reported an average of six hours and 34 minutes a day for AI users, against six hours and 17 minutes for non-users. That is a correlation, not proof that AI caused every additional minute. The more revealing finding is how few people changed the work itself.

One day later, SAP’s finance chief described the same constraint.

Reuters reported on July 23 that SAP CFO Dominik Asam believes AI must move beyond chatbots and coding tools before companies will see broader returns. He said the “lion’s share” of token consumption still goes to those relatively simple uses.

Finance and supply-chain processes are different. An error inside a multistep financial process can carry forward, compound and create a compliance problem. The assurance requirement is higher because the consequence is higher.

Asam’s warning was direct: a general-purpose model will not repair messy legacy data silos. Sending more tokens into an unclear process can raise the cost without improving the result. The most powerful model is not automatically the most useful one.

The practical response is not a company-wide AI program. Choose one recurring process with a clear input, a named owner and an output that can be checked.

Start with a weekly sales report, invoice follow-up, customer-intake summary or inventory exception list. Record the information source, the correct result, the approver and the time or money consumed today.

Then measure the outcome, not the number of prompts. A workflow that saves three verified hours each week has value. Ten people experimenting with a chatbot may be interesting, but activity is not the same as return.

Viktor lives inside Slack and Microsoft Teams. You @mention it in a thread the same way you would ask a colleague. The output — a PDF, a report, a task created in your CRM, an email drafted in Gmail — lands where it should land.

For a weekly sales report, Viktor can collect the approved figures, apply the same structure, flag missing data and place the draft back in Slack for review. For invoice follow-up, it can prepare the messages and wait for approval before anything leaves the business.

That distinction matters. The goal is not more AI use. It is a repeatable piece of work, with known data, a visible result and a human decision at the point where judgment matters.

Viktor operates under SOC 2 controls and supports GDPR and CCPA compliance, with CASA Tier 3 certification. Credentials are isolated in an encrypted vault, contracts prohibit customer data from being used for model training, and approval gates keep sensitive actions under human control in Slack. The full security architecture is explained at viktor.com/security.

You get $100 of free credits to begin. No time limit, no commitment. That's enough to do real work and see what Viktor can actually do before you spend a penny. There's also $50 off your first bill. You must use this exact link to receive both benefits.

Get started with Viktor: https://AIThatDelivers.com.

Disclosure: Some links in this article are affiliate links. If you choose to get started with Viktor using the links provided, I may receive a commission — at no additional cost to you. I only recommend tools I use and believe in.