Info Tech Inc

Photo By:  Immo Wegmann

For several years, using AI at work was fairly easy to understand. An employee opened a chatbot, typed a question and received an answer. IT teams could decide which tools employees were allowed to use and who had access to them.

But today, agentic AI is changing that.

Tools such as Claude Code and Codex can do much more than answer questions or suggest code. They can work with files, use software tools, run code and complete multi-step tasks with varying levels of human supervision among other things. The technology is moving AI from something that simply responds to people toward something that can carry out work on their behalf.

As these tools spread through companies, a new problem is emerging: companies may have a harder time understanding exactly how AI is being used.

The issue is no longer just how many employees use AI. It is what those employees are asking AI to do. The change is especially clear in and among software development. Imagine a developer asking an AI agent to fix a bug. The agent could inspect the relevant code, make changes, run tests and respond to problems it encounters along the way. One request from an employee can lead to a series of actions by the AI.

The employee made one request. The AI performed many steps. That difference creates a new challenge for company security teams.

Traditional security systems are largely built around questions such as: Who is this person? What systems can they access? What permissions do they have? With AI agents, companies also need to ask: What can the AI actually do with those permissions?

That question becomes more important as companies connect agents to internal data, software and business systems. AI agents are increasingly being used for tasks beyond software development, including research, reporting and other business processes.

Harsha Kumar, CEO of NewRocket, has focused on similar issues as companies bring AI into enterprise workflows. His work has included discussions around AI governance, data and the systems needed to make AI useful inside large organizations.

The shift also makes AI adoption harder to measure. A company might say that a certain percentage of its employees use AI. But that number does not necessarily show how much work those employees are delegating to agents.

One employee might use an AI assistant for a few questions each day. Another might use several agents to work through complex tasks over several hours.

The amount of work being done by AI can therefore vary significantly from one employee to another.

The cost can vary too.

Agentic systems can use more computing resources when they perform longer tasks or use multiple tools. That makes AI spending different from traditional software spending, where companies often buy a fixed number of licenses for a predictable price.

Companies are already paying closer attention to this problem. JPMorgan, for example, has introduced a sandboxed environment for some engineers using AI and has placed spending limits on certain users. The goal is to give employees access to powerful AI while creating additional controls around how the technology is used.

There is another challenge which is, oversight.

One approach is to require employees to approve important actions. But as agents perform more steps, asking a human to approve every action may become difficult to scale.

The alternative is to put stronger controls around the agent itself.

That can mean limiting which systems it can access, giving it only the permissions required for a particular task, recording its actions and placing it inside a controlled environment.

This points to a broader change in how companies may need to think about AI governance.

Instead of simply asking how many employees have access to an AI tool, companies may increasingly need to understand how many agents are operating, what systems they can reach, what permissions they have and how much work they are completing without direct human involvement.

That is a much harder problem to measure.

It also changes the meaning of AI adoption.

A company can know exactly how many employees have an AI license. It is much harder to know what those employees are building with the technology, what workflows they are creating and how much authority they are giving to AI agents.

That unpredictability could become one of the defining challenges of enterprise AI.

Agentic AI promises to make companies more efficient by taking on work that once required human effort. But as these systems become part of everyday business processes, they could also make the organization itself harder to understand.

The biggest question for businesses may no longer be:

“How much are our employees actually using AI?”

The question instead may be:

“Do we know what our AI is doing?”

Photo By:  Immo Wegmann

For several years, using AI at work was fairly easy to understand. An employee opened a chatbot, typed a question and received an answer. IT teams could decide which tools employees were allowed to use and who had access to them.

But today, agentic AI is changing that.

Tools such as Claude Code and Codex can do much more than answer questions or suggest code. They can work with files, use software tools, run code and complete multi-step tasks with varying levels of human supervision among other things. The technology is moving AI from something that simply responds to people toward something that can carry out work on their behalf.

As these tools spread through companies, a new problem is emerging: companies may have a harder time understanding exactly how AI is being used.

The issue is no longer just how many employees use AI. It is what those employees are asking AI to do. The change is especially clear in and among software development. Imagine a developer asking an AI agent to fix a bug. The agent could inspect the relevant code, make changes, run tests and respond to problems it encounters along the way. One request from an employee can lead to a series of actions by the AI.

The employee made one request. The AI performed many steps. That difference creates a new challenge for company security teams.

Traditional security systems are largely built around questions such as: Who is this person? What systems can they access? What permissions do they have? With AI agents, companies also need to ask: What can the AI actually do with those permissions?

That question becomes more important as companies connect agents to internal data, software and business systems. AI agents are increasingly being used for tasks beyond software development, including research, reporting and other business processes.

Harsha Kumar, CEO of NewRocket, has focused on similar issues as companies bring AI into enterprise workflows. His work has included discussions around AI governance, data and the systems needed to make AI useful inside large organizations.

The shift also makes AI adoption harder to measure. A company might say that a certain percentage of its employees use AI. But that number does not necessarily show how much work those employees are delegating to agents.

One employee might use an AI assistant for a few questions each day. Another might use several agents to work through complex tasks over several hours.

The amount of work being done by AI can therefore vary significantly from one employee to another.

The cost can vary too.

Agentic systems can use more computing resources when they perform longer tasks or use multiple tools. That makes AI spending different from traditional software spending, where companies often buy a fixed number of licenses for a predictable price.

Companies are already paying closer attention to this problem. JPMorgan, for example, has introduced a sandboxed environment for some engineers using AI and has placed spending limits on certain users. The goal is to give employees access to powerful AI while creating additional controls around how the technology is used.

There is another challenge which is, oversight.

One approach is to require employees to approve important actions. But as agents perform more steps, asking a human to approve every action may become difficult to scale.

The alternative is to put stronger controls around the agent itself.

That can mean limiting which systems it can access, giving it only the permissions required for a particular task, recording its actions and placing it inside a controlled environment.

This points to a broader change in how companies may need to think about AI governance.

Instead of simply asking how many employees have access to an AI tool, companies may increasingly need to understand how many agents are operating, what systems they can reach, what permissions they have and how much work they are completing without direct human involvement.

That is a much harder problem to measure.

It also changes the meaning of AI adoption.

A company can know exactly how many employees have an AI license. It is much harder to know what those employees are building with the technology, what workflows they are creating and how much authority they are giving to AI agents.

That unpredictability could become one of the defining challenges of enterprise AI.

Agentic AI promises to make companies more efficient by taking on work that once required human effort. But as these systems become part of everyday business processes, they could also make the organization itself harder to understand.

The biggest question for businesses may no longer be:

“How much are our employees actually using AI?”

The question instead may be:

“Do we know what our AI is doing?”

Finding the Cause Behind Retail Results

Photo By: charlesdeluvio

Retailers and consumer brands have access to more data than ever before. They can track purchases, customer behavior, advertising, pricing, promotions, product reviews and market trends. Yet having more data does not necessarily answer one of the most important questions in business: What actually caused the outcome?

That is where causal inference comes in. Causal inference is a set of statistical and research methods used to determine whether one factor actually causes a change in another. It goes beyond identifying a relationship between two things. A correlation can show that two events happen together, but it does not necessarily mean that one caused the other.

Consider a retailer that offers a 20% discount on a product and then sees sales increase by 30%. It might be tempting to conclude that the discount caused the increase. However, other factors could have played a role. The product may have received additional advertising, a competitor may have raised its price, or demand may have naturally increased because of a holiday. Looking only at sales before and after the discount does not reveal which factor caused the increase.

From Correlation to Causation in Retail

One of the central ideas behind causal inference is the counterfactual, which asks what would have happened under a different condition. If a retailer gives a customer a discount and that customer makes a purchase, the retailer can observe what happened after the discount. It cannot observe what the same customer would have done at that exact time without the discount. Researchers therefore use experiments and statistical methods to estimate this missing outcome.

Randomized controlled experiments are one of the strongest approaches for studying cause and effect. A retailer could randomly offer a discount to some customers while giving similar customers no discount. If the customers who received the discount purchase more frequently or spend more, the retailer has stronger evidence that the discount caused the difference.

Applying Causal Inference to Business Decisions

Retailers can apply this approach across many parts of their business. Marketing teams can test whether an email campaign increases purchases. Pricing teams can study whether a price reduction increases demand. Merchandising teams can examine whether changing product placement affects sales. Digital teams can test whether a website change causes more customers to complete a purchase.

Retailers cannot always run randomized experiments. Some decisions may be too expensive, take too long or be impractical to test. In those situations, researchers can use observational data and statistical techniques such as propensity score methods and difference-in-differences. These approaches can help estimate causal effects when their underlying assumptions are appropriate.

The Challenges of Causal Inference in Retail

One of the biggest challenges in observational data is confounding. A confounding variable is a factor that influences both the variable being studied and the outcome. If researchers do not account for it, they can mistake an association for a causal relationship.

For example, suppose a retailer discovers that members of its loyalty program spend more than nonmembers. It might conclude that the loyalty program caused customers to spend more. However, customers who choose to join a loyalty program may already shop more frequently. Their existing shopping behavior could explain some or all of the difference.

The same issue appears in advertising. If a retailer spends $1 million on advertising and sales increase afterward, the increase alone does not establish that the advertising generated $1 million in additional sales. Some customers who saw the advertisements may have purchased anyway, while other factors may have contributed to the increase.

For retailers, this distinction between correlation and causation can have significant financial consequences. A dashboard might show that several marketing channels performed well during a period when sales increased. That does not necessarily mean every channel caused incremental revenue. Causal analysis attempts to separate the revenue that would have occurred anyway from the revenue attributable to a particular decision.

This is an area where companies such as Kapnova are applying causal inference to consumer business decisions. Kapnova describes its platform as a decision engine built specifically for consumer brands, combining causal inference, econometrics, simulation and optimization. The company is led by CEO and co-founder James Sun. Kapnova is an agentic revenue and profit optimization system. AI finds the opportunities. Math determines the answer.

Kapnova’s platform uses causal modeling and econometric methods to evaluate decisions involving areas such as pricing, promotions, marketing, inventory and demand. It also uses Monte Carlo simulations to explore thousands of possible outcomes and quantify uncertainty. The platform can incorporate factors including consumer sentiment, reviews, competitors and broader market signals into its analysis.

The approach is particularly relevant to retail because companies rarely want to know only what happened in the past. They want to understand what could happen if they change a price, shift an advertising budget, launch a promotion or make another business decision.

Causal inference does not guarantee that every business question can be answered with certainty. The quality of a causal conclusion depends on the data, research design and assumptions used to produce it. Observational studies require careful consideration of potential confounding, selection effects and other sources of bias.

For retailers and consumer brands, however, causal inference provides a way to move beyond simply observing what happened. It allows companies to investigate why something happened, estimate the impact of a particular decision and evaluate what could happen if they make a different choice.

Ultimately, causal inference comes down to a simple question: If a company changes something, what difference will that change actually make? For businesses making decisions about prices, promotions, advertising, inventory and customer relationships, answering that question can be far more valuable than simply knowing which numbers moved together.