Trust Is a Design Choice: Leading AI Where It Actually Matters

I started the year hopeful, intentional, and optimistic! I had a vision: to connect my experiences in design, planning, enablement, and delivery with the curiosity that AI hype was creating among the professionals.

What I saw was familiar:

• Leaders in denial that behaviors had to change

• Teams overwhelmed by how fast tools were multiplying

• Skeptics saying, “This too shall pass”

• Laggards frozen by fear of new technology

My reaction was simple: explain, simplify, and give people a compelling reason to change how they think about how they work or what they use but to get them to ask why should I use AI?

That intention shaped everything that followed in 2025.  The content of the workshops, the discussions I participated in, the keynote and speaking engagements. I am grateful to all those organizations and sponsors who helped me drive my vision to success last year.

Thought Leadership Wave 1: AI Is About Leadership and Decisions

From the years I have managed or led cross- functional global AI initiatives, I note:

AI initiatives fail when they focus on solutions instead of decision systems.

Even when we are using predictive AI approaches the challenges of real time decisions being fed by automated processes can be a challenge. With fragmented decisioning and non-streamlined AI applications and assets, what struck me as bigger risks were that:

• Accuracy alone should not be the goal, decision quality should.

• Automation doesn’t remove accountability; it concentrates it beyond that step.

• Leadership intent must come before AI strategy. “Drive thru” strategies lack the holistic vision that AI needs.

TIP 🧭

🌟If you can’t name the owner of an AI-driven decision, you’re not ready to automate it.

🌟 “AI isn’t the easy button—it just hides complexity behind a nicer interface.

🌟” This thinking became the backbone of my AI for Business Leaders workshop (1).

Thought Leadership Wave 2: Data Supply Chain Enables AI

Working with Colin Coleman (2), I added a new lens to dismantle the hype. Defining the elephant in the AI hype that was not being directly addressed in 2024:

“Can we trust what flows into and out of AI systems?”

We argued:

• AI is only as resilient as its data supply chain

• Most organizations govern data after value has already leaked

• Generated data creates new cost, risk, and ownership questions

From raw data → semantic layers → decision systems → human consumption—every step has a cost:

• Bad data and quality of data handled in the supply chain

• Delayed decisions, siloed or fragmented handoffs of data

• Short-sighted choices, lack of feedback, monitoring and lineage transparency

TIP 🧭

Trust isn’t built at the model design. It’s built in the data path design. Larger and layered an organization in retail, manufacturing, financial services, healthcare etc., the trust starts eroding as you move further from the source data.

🌟"Hallucinations aren't corrected at the output stage; they're addressed within the data supply chain."

Thought Leadership Wave 3: Trust Must Be Designed, Not Declared

At the McLeod User conference (3) for early AI adopters in trucking, logistics, and supply chains, I saw something powerful:

Organizations trusted AI more when they could see how it worked, especially when it supported their real-time decisions. As I learnt about the complexity of the trucking logistics, from shippers, wholesales, distribution agents, to the end customer and the existence of information-fragmentation in scheduling logistics. The sophistication of the digital metrics collected from the trucks alone (100-200 sensor readings about the truck, the payload, driver, driving conditions etc.) must be managed against the shippers and trucker agency availability.  As an AI solution architect, I could call it out as a large optimization problem that we can solution for. The challenge is to provide this for the early adopters in a more consumable and relatable way.  This can happen as we move from excel reports to excel reports fed by AI assets, from simple static visualization dashboards to dynamic real time reporting and insight recommendations.

That’s when it became clear to me that whatever form of AI we adopt and make accessible: we are still missing a key decision factor. It is not just about accessing the expertise; it is about knowing how to use it.

Responsible AI must be operational—not aspirational. Trust must survive operational stress test.

Decision leaders should not just ask, “Is it impressive?” They should be asking:

• Should we ship it, as is?

• Ship it with guardrails and monitoring thresholds?

• Or block it?

TIP🧭: 

Governance isn’t about slowing things down. It’s about knowing when speed is earned.

🌟“Governance only matters when something goes wrong.”

Thought Leadership Wave 4: Human-in-the-Loop Is Strategic Design

Over Q2, I kept hearing - “How do we structure this?”

My answer: Bring humans back—not as safety nets, but as strategic designers.

I recommended a RADAR framework in 2024 as an individual AI risk management action plan. NIST RMF, OECD FAIR principles, and EU AI Act provide playbooks and guidelines for global governments and organizations to MAP, Measure, Manage, and Govern Risks. But there is no guidance for individual professionals. That is left to the organizations to enforce top down.

Human-in-the-loop works when:

• Ambiguity is managed first closest to its emergence

• Escalation paths are intentional and customized to business structures

• Automation earns autonomy only with transparency

The use case of Klarna showed us what happens when humans are removed too early: hidden risk explodes.

TIP 🧭: 

AI Leaders need to be aware that: Full automation is a business decision—not a technical destiny. 

🌟“Humans manage ambiguity. Machines manage repetition.”

This shaped my work on the RADAR framework and sessions on scaling GenAI—from exploration, to operations, to scale. Carlos Alejandro Gorricho and I built a GenAI value model to guide investment decisions to support AI leadership. (4)

Thought Leadership Wave 5: AI Fluency at the Edge

Publishing AI Leadership Compass: Lead with Clarity, 7 Moves that Power AI Transformation (5) was my way of making AI fluency accessible:

• A shortcut for rookies, early AI adopters

• A reset for experienced leaders

• A reminder not to abandon what already works in leadership

We see,

• Innovation happens at the edge.

• Governance lives at the core.

The bridge between them is fluency.

TIP🧭: 

You don’t need more AI experts. You need better questions from everyone.

🌟“Collective intelligence beats individual brilliance every time.”

Where I Landed

By the end of the year—after studying AI risk, governance, and regulation—I’m more confident than ever in this: AI success is not about models. It’s about

o   decision ownership,

o   trust under uncertainty, and

o   leadership that doesn’t disappear when automation shows up.

Final Thought🤔

That’s when AI actually matters.

 Thanks to the opportunities to build the foundations around AI Fluency:

#AI+IMGlobalSummit , #MPact, #UC2025, #McLeodSoftware, #EDW+DGIQ25, #AlignAI, #DataConnect25, #EDMC, #FabricTour, #EFX, #SQLweek25, #WEOP,  #Thisdot, #TheGeorgiaGrowth&SuccessSummit, #ElevateHer, #Molynyckle, #DECU, #CBTS, #CDA, #ElevateAI, #WomeninData, #DAMA, #INFORMS, #TAG.

References:

• (1) May 2025, AI for Business Leaders: Empower your leadership with Actionable Strategies, Facilitated by Priya Sarathy and Beverly Wright.

• (2)April 2025, Workshop AI+IM Conference Atlanta, Data Management and Information Management: Pillars of AI strategy Execution, Dr. Priya Sarathy and Dr Colin Coleman.

• (3) May 2025, Keynote speaker at Mpact Trucking summit: Intelligence Amplified: The Convergence of BI, Data Science and AI

• (4) May 2025, Workshop on Data Strategy for Scaling Your AI Program, DGIQ+EDW 2025 conference, Priya Sarathy & Carlos Gorricho

• (5) AI Leadership Compass- Lead with Clarity: 7 Moves That Power AI Transformation (On Amazon)