Agentic AI is a phenomenon of language in today’s IT landscape. Definitions vary from vendor to vendor. Everyone says they do it. Does agentic mean the human isn’t involved (meaning the system is completely autonomous)? What about AI Agents? If I build an AI Agent in CoPilot, is that considered agentic?

There seems to be more questions than answers. Yet, like any transformation or change to technology, there are certain things that will remain true. There are certain ingredients that make sure the recipe you’re cooking up is a recipe for success.

The recipe is quite simple. It only takes three ingredients. These ingredients will make any transformation palatable, repeatable and impactful.

1. Employee Impact

Think of employees as the yeast in a freshly baked loaf of bread. Without them, nothing rises. Served customers enable an organization to thrive. Served customers cannot exist where the presence of served employees is absent.

When it comes to transformation, especially with AI, organizations must make their employees the focal point. In practice, this can take various flavors.

Sometimes, it means bringing non-technical employees in to test. It could be holding side-by-sides to understand the day-to-day of various departments (never a bad idea even if no AI initiative is underway!). Perhaps it’s using survey and employee sentiment data to understand how employees feel about the AI training you offer, the AI they use daily or what AI capability is on their wish list for the future.

While there could be an entire book written on how to focus on employees during transformation, the central idea is that the voice of the employee matters. The work they do matters.

AI for AI sake is a recipe for disaster. AI for employee impact’s sake is where culture can be shaped, engagement enabled and better outcomes live.

2. Customer Impact

There is a reason employee impact comes before customer impact. You cannot serve customers well if you do not first serve your employees well. No customer experience AI can replace the lasting impact human to human connection can have on your customers.

So, now that the employee impact is squared away, what does this ingredient of customer impact look like?

It would be inaccurate and inauthentic to claim that there is no AI that can impact customers without human dialogue. Serving our customers well is about meeting them where they are, removing friction where friction need not be and ensuring their experience leaves them in a better place than where they started.

As you navigate AI transformation, every vendor will sell you “Agentic” capabilities. To ensure that customer impact is a part of the recipe, you must understand the impact to your customer in practice; not theory.

This means pushing suppliers of AI (internal or external) to go deeper and show you the proof. Many times, the proof is shown in one or two examples (not proven over time and simply not proven for your customer’s challenge). This matters because if you chose to say yes, purchase that solution, without seeing it fully baked, you risk technical debt and spend that yield little fruit.

In the art of Generative AI prompting, the more specific you can be, the more context you provide, the deeper you go into your situation, the better the generated response will be. The same can be said for the implementation of AI for your customers.

Know the friction for your customers. Let that be the driving force in your AI transformation. As Nate Brown says, “Make the quick parts quick and the slow parts meaningful.”

3. Security and Governance

Technically the first ingredient is a pair. Governance is to security as sugar is to a freshly baked cookie. Sure, you can bake a cookie without sugar, but the taste would be repulsive.

Information security is more than just a check box. It is about having the right guardrails in place to protect the customers and employees your organization serves.

Here's what’s non-negotiable:

• Having AI policies in place that inform the decision-making process and what is acceptable
• Knowing the security posture of technology partners providing AI or AI components within their product
• Ensuring you understand what is happening with your customers data

Information Security teams are in the business of creating friction. This is a necessary evil to protect our customers and organizations from the negative impact that follows a data breech, trade secrets being leaked or having to notify a customer that their information was compromised.

Governance is the vehicle that helps ease the friction that Information Security necessarily creates. Make sure the documented policies are followed, communication is clear and the right people have a seat at the table when discussions and decisions are being made.

Agentic AI promises transformative change, but true success comes from blending employee impact, customer impact and security/governance into a holistic strategy.

Employees provide the foundation, customers bring purpose and security and governance ensures everything holds together. Just as a master baker never skips an essential ingredient, leaders and IT professionals should weave these three elements into every AI transformation effort. This is how you bake change that lasts.