AI tools deliver more when you design for humans in the loop
AI adoption isn’t the advantage anymore— workflow design is
Yes, most large enterprises are running AI in production. But from copilots to conversational interfaces, adoption, even at scale, isn't enough to deliver deep ROI.
Why? Because workflows need to be reimagined and redesigned, end to end, to reflect new processes and standards built around human oversight, accountability, and measurable outcomes.
The question leaders need to ask now isn’t, “What can we automate, and with what tools?”
It’s, “How should we redesign the work we do, so people and artificial intelligence succeed together?”
How soon can your teams achieve human+AI fluency?
45% of enterprises are creating new AI-related roles, but 93% of leaders report workforce barriers limiting AI value. Explore Slalom’s 2026 AI Research Report.
Human-centered design for AI-powered advantage
Human-centered design (HCD), also called “human+AI by design,” means redesigning workflows so two things happen:
- Humans and AI can each do what they do best (strategy and efficiency, respectively).
- Humans and AI work better together over time (compounding value).
Instead of adding AI on top of existing processes, HCD starts by designing workflows and talent strategies with people and processes, not technology. It sits at the heart of human-in-the-loop advantage because HCD considers how people think, decide, and create. Then it operationalizes roles and workflows around key decision points, so human judgment, creativity, and accountability shake hands with AI’s speed, scale, and pattern recognition.
Four forces make this shift urgent, and all of them call for explicit leadership choices:
- Technology: AI is moving from pilots to enterprise-wide platforms, so what guardrails exist?
- Customer: Customer expectations continue to rise, so what are our customer or patient experience standards?
- Workforce: The evolution of work demands new skills, so how do organizations reskill at scale?
- Regulation: Organizations must scale AI responsibly, so what human-oversight checkpoints exist?
Industry spotlight #1: Human-centered design solutions for financial services
TP ICAP, a world-leading provider of financial markets infrastructure and data, connects institutional buyers and sellers in the world’s financial, energy, and commodities markets, so they can transact with confidence.
The company’s first three human+AI solutions to reach production are projected to generate millions in value:
- 5,004% prospect screening solution ROI
- 350% sales outreach solution ROI
- 400% test automation solution ROI
Industry spotlight #2: Human-centered design solutions for software engineering
Ocuco are pioneers in software engineering for the optical retail and eyecare industry focused on crafting digital solutions that redefine what’s possible for optometrists, optical labs, and retail chains. The organization redesigned workflows and adopted a change management program that empowered Ocuco developers to fully adopt GitHub Copilot.
The company achieved significant gains in just their first quarter:
- 30% increase in story points delivered per developer
- 233% increase in time savings per week
- 62% of user report better quality code post-training
- 19% reduction in new defects per sprint
Industry spotlight #3: Human-centered design solutions for media & entertainment
Keywords Studios is trusted by the world’s leading video game and entertainment companies to work alongside them during the full game development cycle and bring immersive content to life. Its customers face fierce competition fueled by cutting-edge technology, gamer expectations, and a recent gaming surge.
By leveling up its data and AI strategy, Keywords Studios is changing the game:
- 3 months from prototype to launch of AI sentiment analysis tool, Lens
- 90% reduction in time to uncover customer insights and generate reports
What are the tradeoffs between deploying AI tools with and without humans in the loop?
New evidence is emerging that over-use of generative AI (GenAI) is accelerating cognitive deterioration and may compromise critical thinking and decision making. Firms that keep humans in the loop sharpen capability and resilience. Firms that hand judgment to machines become brittle and dependent, widening the gap every quarter.
Meanwhile, MIT Media Lab reports that 95% of investments in enterprise GenAI pilots do not deliver value or move into production. In other words, 95% of GenAI pilots produce zero returns.
Human+AI design risks vs. upsides
|
Risk if design is ignored |
Upside if design is prioritized |
Tradeoff |
|---|---|---|
|
Capability erosion
Over-delegating decisions to AI weakens human judgment and reduces capacity over time. |
Expansion
Designing workflows that keep humans engaged in judgment helps AI personalize intelligently |
Decay vs. growth |
|
Cost drag
Higher expenses from exception handling, audit failures, and manual oversight don’t offset gains. |
Productivity
Purposeful automation shifts effort from routine tasks to human talent, unlocking higher-value reasoning and synthesis. |
Burdens vs. efficiency |
|
Reputation risk
Unexplained decisions erode customer confidence and drive employee resistance; rebuilding trust slows momentum and gets expensive. |
Cognitive strengthening
Responsible scaling with interaction patterns that stimulate reflection and refine mental models make individuals and teams think better. |
Shatter vs. sharpen |
How should leaders think about designing human+AI workflows?
Start with discovery, not deployment. Know where humans must stay in control (e.g., judgment, context, ethics) and where AI should accelerate patterns and surface signals. Also, prioritize designing new human+AI collaborative workflows where friction is highest and human oversight is essential.
Here are three tips for how to design AI workflows with humans in the loop:
- Map how decisions move between humans and systems.
- Prototype how AI challenges assumptions and sharpens thinking.
- Set escalation rules for ambiguity, risk, and exceptions.
When designed well, AI becomes a thinking partner that enhances competitive processes, strengthens judgment, sharpens skills, and accelerates learning. It also changes what you measure.
Here’s a framework for taking action:
|
The first 30 days |
The first 60 days |
By Day 90 |
First quarter financial close |
|---|---|---|---|
|
Set the vision
CEOs must make it clear that AI is for augmentation, not replacement, and take actions that build trust across the organization. |
Lead workflow redesign
CIOs & CTOs need to drive rethinking before buying tools and map milestones where AI adds scale and where people lead the way. |
Shape the skill plan
CHROs need to shape a reskilling plan to support new workflows given the broader skills scarcity. |
Set KPIs and value baseline
CFOs must demand outcomes—not activity—and track them against a clear baseline to prove real value. |
KPIs for human+AI collaboration and workflows with humans in the loop
When AI becomes a teammate instead of a tool, business models evolve. Workforces adapt and pivot from execution to oversight, with more consideration for compelling creativity and judgment. Executives need to ask three key questions, and keep asking them quarter by quarter:
- Are redesigned workflows scaling?
- Are employees adopting AI with trust?
- Are customer experiences improving?
Expect your first KPI to show movement and reportable business value within one to two quarters. Expect clear trend lines within nine months.
AI-enabled workflow KPIs
|
KPI |
What it shows |
Executive lens |
|---|---|---|
|
% of workflows redesigned |
Depth of AI-human integration |
CIO/CTO/LoB Leader
Are we stuck in “pilot purgatory” or are we scaling effectively? |
|
Hours of productivity gained Target cadence: Quarterly |
Tangible time savings |
CEO/COO/CFO/LoB
Are we unlocking capacity? |
|
Employee adoption and trust
Target 60-80% active use in 6 months for prioritized roles |
Workforce trust and readiness |
CHRO/LoB Leader
Are people using it confidently? |
|
Customer trust and satisfaction |
Quality of experience |
CXO/LoB
Are we building loyalty? |
|
Critical cycle time reduction |
Speed of critical processes |
CEO/CIO/LoB
Are we faster to market? |
|
% savings reinvested into growth |
Turning efficiency into value |
CEO/CFO/LoB
Are we fueling innovation? |
When human+AI workflows are designed well, AI scales what matters
When AI is designed into work—not layered on top—it creates a design dividend that compounds over time. Teams build stronger judgment, learn faster, and improve outcomes as humans and AI reinforce each other. When design is ignored, the opposite happens: Capability erodes, dependence grows, and early gains slough off into long-term drag and negative business value.
Design workflows for humans+AI to win together
FAQs
Because most organizations are layering AI on top of existing workflows instead of redesigning the work itself. Without new roles, decision points, and accountability, AI speeds up broken processes rather than fixing them. Real ROI comes when workflows are redesigned so humans and AI succeed together.
Human+AI by design means intentionally redesigning workflows so humans remain accountable for judgment while AI accelerates speed, scale, and pattern recognition. Instead of asking what can be automated, leaders ask how work should change so people and AI each do what they do best. Over time, this creates compounding value rather than one-off efficiency gains.
Human-in-the-loop design keeps people actively involved at critical decision points—especially where context, ethics, and accountability matter. This prevents over-delegating judgment to machines and protects long-term capability. Organizations that maintain human oversight build resilience and trust as AI scales.
Start with discovery, not deployment. Identify where humans must stay in control and where AI should accelerate patterns or surface signals. Prioritize redesigning workflows where friction is highest and oversight is most critical before scaling tools.
Leaders should measure redesigned workflows, not just tool usage. Key indicators include productivity gained, employee trust and adoption, customer experience improvements, and cycle-time reduction. Expect early movement within one to two quarters and clear trend lines within nine months.
