Is Your AI Strategy Just Chasing the Hype? — Aroon Jham, Paychex

Aroon Jham currently leads go-to-market analytics at Paychex, bringing with him a career built across several global organisations, including Thomson Reuters, Dell and Fiserv. Across call centre analytics, marketing analytics and machine-to-machine analytics, go-to-market has become his favourite domain, precisely because its impact on sales can be measured and tested through experimentation. 

Early in his career, Aroon noticed his teams were busy fulfilling report requests without always being able to tie that work to genuine value — a realisation that reshaped how he leads. Today, he trains his teams to act as business consultants rather than order-takers, and applies the same outcome-first discipline to how his organisation approaches AI.

How has your view of analytics evolved from “order-taker” to strategic partner?

Aroon Jham reflects on the early frustration of measuring busyness rather than value, and how that shifted his leadership philosophy.

“In the earlier part of my journey, analytics was considered more of an order-taker function — bring this report, bring that data. We were very busy, but the question that always made me restless was: does busy equal value?

That question began to shape my journey, from measuring analytics like a manufacturing process to treating it as a value-generating process. From a leadership perspective, I train my team not to be order-takers but business consultants. I don’t want reports that go out into the world where nobody knows who is consuming them or what insights they are getting from them. That is passive consumption. Active participation is the way to go.”

What is the biggest misconception leaders have about enterprise AI?

Aroon argues that AI should never be the default answer, and shares how his team designed an AI solution around where sellers already work.

“Organisations are viewing AI as their silver bullet, and that is not the case. If you view AI as ‘I’ve got to use AI,’ then you are doing it wrong. Like any other business challenge, you have to start by asking what the business outcome is that you want to solve — and sometimes AI is not the answer.

At Thomson Reuters, the problem we were solving was that our sellers were spending a lot of time researching their customers before calls. Rather than building a chatbot, we mapped the entire process end to end and met sellers inside Microsoft Teams, where they already worked, so they could ask a question and get everything they needed about a customer in one place. Several AI initiatives fail because leaders don’t foresee the change management required to adopt them.”

Why does successful AI adoption depend on minimising change management?

Drawing on advice from a fellow tech leader, Aroon Jham explains why the best AI tools are the ones users barely notice.

“A friend of mine, a tech leader at another company, told me his advice was to make AI invisible. That way, change management becomes so much easier. The more visible AI is, the more change management is expected, which means the harder it is to adopt.

Change management should not be an afterthought; it should be embedded in the design itself. The easier something is to use, the lower the friction — and ideally, it becomes invisible. That is why the research tool we built at Thomson Reuters lived inside Microsoft Teams, so nobody had to learn a new platform to benefit from it.”

Where should organisations draw the line on AI governance and data exposure?

Aroon describes the tension between giving AI enough context to be useful and exposing too much sensitive information.

“We are tempted to give AI as much context as possible, but you have to be careful about how much you share. That becomes your internal governance. From an enterprise perspective, you have to put guardrails in place so that when someone submits something they should not, the system catches it.

From a design perspective, we believe in minimum exposure: give AI only the minimum information it needs. It is a constant balancing act — how much information should AI have to be effective versus dangerous. There is no golden answer; it is very situational.”

If every executive pursuing an AI transformation initiative paused to ask one question, what should it be?

Aroon Jham offers the standard he holds his own team to before any AI initiative moves forward.

“Can I connect the work that I am doing to an outcome, and is the outcome measurable? If the answer to either of those makes you pause, you should pause, re-scope and re-pivot. You don’t accept the unknown as ‘maybe I am producing value’ — you trace your steps until you get the answer.

I would add a third part to that same line of thinking: does the outcome connect to an actual business goal? Measurement without a goal is still incomplete. If your stakeholder doesn’t know what the goal is, the pressure has to go back on them to define it — you may be technically capable, but you still need a goal in place.”

Join the Conversation: The Ortus Club’s Executive Network

As Aroon’s perspective shows, connecting analytics and AI work to measurable business outcomes is rarely a solo exercise. Leaders at this level rely on peer dialogue to pressure-test whether an initiative is genuinely creating value, or simply keeping a team busy.

His approach to making AI invisible and governance a constant balancing act reflects a broader reality: today’s data and analytics leaders cannot rely on generic playbooks alone. The most effective executives actively seek out peer dialogue to stress-test their thinking against real-world experience.

At The Ortus Club, we host curated executive roundtables that bring together senior leaders facing these exact challenges. Step away from the noise around AI hype and engage in the kind of open, high-value conversations that connect strategy to measurable outcomes.

Frequently Asked Questions

Q: What is the biggest misconception leaders have about enterprise AI?

A: Many leaders treat AI as a silver bullet, applying it by default rather than starting with the business problem. Aroon argues AI should only be used when it is genuinely the right solution to a defined outcome, not adopted for its own sake.

Q: What does it mean to “make AI invisible”?

A: It means embedding AI into the tools and workflows people already use, rather than requiring them to learn a new platform. This reduces the change management burden and makes adoption far easier.

Q: How can organisations balance AI usefulness against data governance risk?

A: By applying a principle of minimum exposure — giving AI only the context it needs to be effective, while maintaining internal and enterprise guardrails to prevent oversharing of sensitive information.

Q: Why does Aroon encourage analytics teams to act as “business consultants” rather than “order-takers”?

A: Because passively fulfilling report requests does not guarantee value. Acting as a consulting partner means challenging stakeholders to define outcomes and goals, and connecting analytics work directly to measurable business impact.

Q: What question should every executive ask before pursuing an AI initiative?

A: Can the work be connected to an outcome, is that outcome measurable, and does it serve a clearly defined business goal? If any part of that is unclear, Aroon recommends pausing before proceeding.