AI should improve worker clarity, reduce cognitive load, and create calmer, more meaningful workflows.

Posted Dec 8 2025

Credit: Roman Samborskyi / Shutterstock

Most AI projects inside organizations still begin with the same question: What can we automate? It’s a reasonable instinct, but in practice it often produces systems that make the human experience of work more fragmented or stressful.

However, there is a slightly more nuanced way to identify how AI can help us in our work life, and ultimately make work feel better.

Start with the Human Experience, and the Systems Behind It

When planning AI initiatives, companies often map processes as a sequence of steps. But people experience work in moments. The emotional “dips” in those moments of frustration, waiting, uncertainty, constant context switching that often correspond to deeper system-level issues: scattered data sources, unclear source-of-truth, poorly designed handoffs, or information retrieval overhead.

In my consulting work, I encourage organizations to map what I call human value journeys, e.g., customer journeys, employee journeys, insights and decision making journeys. They surface questions such as:

• Where does information live, and how hard is it to retrieve?
• Where do people re-enter or reconcile data across systems?
• Where do they switch tools unnecessarily?
• Where does the cognitive load outweigh the task itself?

These friction points often reveal more value than any workflow diagram because they point to both human and technical bottlenecks.

A client of mine had account managers manually monitoring online reputation for dozens of brands. Every week, they opened multiple review platforms, collected data by hand, and compiled sentiment snapshots. This was a classic case of high cognitive load + low-value data retrieval.

Initially, the organization thought of “a report generator” to make sentiment insights more consistent across accounts. But mapping the team’s value journey showed the real issue: the information retrieval loop was the bottleneck.

The solution was a basic agent built on top of lightweight scrapers, combining retrieval, normalization, and summarization. It aggregated data into a consistent structure, generated a sentiment overview, and highlighted anomalies. Humans still reviewed and interpreted the results, but they started from a clean context, not from scratch. The ultimate goal was reducing system friction so that human judgment became easier and more effective.

A Simple Framework for Human-Centered AI Strategy

Organizations can adopt this approach with a straightforward shift:

1. Map the moments that matter

Identify emotional friction and connect it to underlying system behavior: information gaps, coordination delays, or high-context tasks unsupported by current tools. By identifying these moments, organizations can pinpoint where AI can make the biggest difference for people, not just where automation is convenient.

2. Diagnose the technical root causes

Once friction points are clear, analyze the underlying system issues. Is the issue caused by poor data architecture, lack of unified retrieval, missing (meta)data, or repetitive cognitive overhead?

Most “people problems” turn out to be systems problems in disguise, and understanding this distinction allows AI strategies to target the source rather than the symptom.

3. Apply AI to remove friction

AI is most effective when used as a socio-technical layer that:

• retrieves and consolidates information
• summarizes or interprets context
• standardizes repetitive inputs
• smooths handoffs in multi-system workflows
• reduces cognitive effort.

Focus AI strategy on enhancing these capabilities for people, and help them focus on judgment, creativity, and high-value tasks rather than tedious busywork.

The Point

AI shouldn’t only make organizations faster. It should improve clarity, reduce cognitive load, and create calmer, more meaningful workflows for the people using the systems.

When AI is applied as a friction-removal layer rather than a replacement layer, organizations see higher trust, stronger adoption, and better outcomes.

If we start with how work feels, we’ll build systems that work better for everyone.

***Alessio Bilato *is the founder of Value Iteration, helping online entrepreneurs and emerging tech ventures design AI-driven products and services, and set up the data and system infrastructure to bring them to life.

Submit an Article to CACM

CACM welcomes unsolicited submissions on topics of relevance and value to the computing community.

You Just Read

AI Should Make Work Feel Better

© 2025 Copyright held by the owner/author(s).