top of page
Search

How Deep Should Companies Go Into AI Today?

The Cambiara Group
3 days ago
10 min read

AI is moving quickly from being something companies experiment with to something that is becoming part of how they actually operate. Writing, coding, customer service, fraud detection, inventory management, maintenance and decision-making are just some of the areas where AI is finding its way into everyday business.


But something else is happening at the same time. The enthusiasm is being met with an increasing amount of resistance.


Communities are questioning the impact of data centres on electricity, water and local infrastructure. Employees are wondering what AI means for their jobs and how much monitoring they will be subjected to. Customers are concerned about privacy. Regulators are trying to catch up, while creators are challenging how AI companies use copyrighted material.


In other words, the technology is moving ahead, but public confidence isn't necessarily keeping pace. This raises a more interesting question than whether companies should be using AI.


How deep should they actually go?


Cambiara Group's view is that companies need to go far enough to make a meaningful difference to how they operate, but not so far that they lose control of their business, their people, their data or their ability to make sound decisions.


There is a difference between embracing technology and embracing it without thinking through the consequences. And that difference will become increasingly important.


AI isn't just a technology decision anymore

One of the things that sometimes gets lost in the AI conversation is that, despite being described as a digital technology, AI has a very real physical footprint. The computing power required to support large AI systems has to come from somewhere.


Data centres require electricity, cooling, water, land, fibre connections and ongoing maintenance. In communities where infrastructure is already under pressure, these requirements can create some difficult questions.


Who pays for the additional electricity? Where does the water come from? What does the community get in return? Will the jobs created justify the infrastructure demands? And what happens when private companies benefit from public infrastructure investments?


These aren't anti-technology questions. They're reasonable questions about how business decisions affect the communities in which companies operate.


There is also a people dimension.

Employees are unlikely to embrace AI enthusiastically if they believe the primary objective is to eliminate their jobs, monitor their activities or remove human judgement from decisions. On the other hand, if AI helps eliminate repetitive work, improve safety, reduce errors and give people more time to focus on what they do best, the conversation changes considerably.


The distinction is important.

Technology adoption is rarely just about the technology. It's about what changes for the people expected to use it.


Companies that want to make significant investments in AI need to be able to explain why they're doing it and who benefits. "Because our competitors are doing it" isn't a particularly compelling business strategy.


Going deep doesn't mean buying more AI

There is a tendency to measure AI adoption by the number of tools being introduced, the number of pilots underway or how many employees are experimenting with ChatGPT.

That's not necessarily progress.


Cambiara Group's experience in organizational transformation suggests that introducing more technology doesn't automatically improve an organization. Sometimes it just gives people another system to work around.


The same thing can happen with AI.

Going deep should mean finding meaningful ways to incorporate AI into the work that matters. That might mean improving customer service, reducing operating costs, improving product quality, managing risk or making better decisions.


It should be about business outcomes, not activity.

An organization can have dozens of AI experiments underway and still have very little to show for them. Another organization might have just two or three applications that fundamentally improve how it operates.


The difference is focus.

A practical way to think about AI adoption is in four stages:

Level

What it looks like

Primary benefit

Light use

Employees use public AI tools for drafting, summarizing and brainstorming

Individual productivity

Practical use

Approved AI tools support customer service, reporting, coding, research and document review

Measurable efficiency

Core use

AI becomes part of important business processes such as operations, maintenance, pricing and risk management

Competitive advantage

Frontier use

The company develops specialized models, AI products or significant AI infrastructure

Potentially transformational advantage

Most organizations should be moving beyond light use. Many should be looking seriously at practical applications and selected areas of core use.


But frontier AI isn't for everyone.

A regional manufacturer, for example, doesn't need to build its own large language model to benefit from AI. It might get considerably more value from predicting equipment failures, improving production scheduling, identifying product defects or giving technicians faster access to maintenance information.


A financial institution may have a much stronger case for deeper AI capabilities because fraud detection, risk modelling and regulatory compliance are central to its business.

A hospital has another set of considerations altogether, where patient safety, privacy and accountability must take precedence.


The point is that the appropriate level of AI adoption depends on the business.


Depth should follow business strategy, not technology enthusiasm.


The risk of doing too little


There is a temptation for some organizations to sit back and wait for the AI excitement to settle down. It's an understandable instinct. There is plenty of hype, plenty of competing advice and a lot of uncertainty about which technologies will ultimately deliver.


But doing nothing isn't necessarily the safe option.

AI is already changing the economics of certain types of work. Tasks that once took hours can sometimes be completed in minutes. Large volumes of information can be reviewed much more quickly. Ideas can be tested, documents analyzed and customer enquiries handled with considerably less effort.


The results aren't always right, and human review remains important. But the productivity implications are real. Organizations that fail to understand these changes may find themselves facing a number of problems.


Competitors will change customer expectations.

If another company can respond more quickly, provide better information or deliver a more personalized service, customers may eventually question why they're still waiting for the old way of doing things.


Good people may become frustrated.

Employees don't necessarily want AI to replace them. But they may wonder why they're still spending hours doing repetitive work that could be handled more efficiently. If organizations prohibit AI without providing alternatives, employees may simply find their own tools and use them without approval.


Operating costs may become harder to justify.

Companies that continue to rely entirely on manual processes may need more time and people to produce the same results as competitors. That may be perfectly reasonable in some businesses, but it needs to be a conscious choice.


Decision-making may fall behind.

AI can help organizations identify patterns in customer feedback, operational reports, market information and maintenance records. Companies that don't develop this capability may find themselves reacting to problems that others saw coming.


And perhaps most importantly, companies that don't develop their own understanding of AI risk becoming overly dependent on vendors.


Buying technology is relatively easy. Understanding what it can and cannot do, how it fits into your business and what you're actually paying for is another matter. The biggest risk isn't that every company that doesn't embrace AI will disappear. It's that some will gradually become less competitive, less attractive to employees and less capable of responding to change.


But going too deep too quickly has its own consequences

Of course, the opposite problem is just as real.. AI can create a considerable amount of organizational risk when companies treat it as though it's infallible. Sensitive information can end up in the wrong hands. AI-generated advice can be inaccurate. Models can introduce bias into decisions. Automated processes can make mistakes at a scale that would have been difficult to achieve manually.


And there is a particular danger in allowing AI-generated information to appear authoritative simply because it is presented confidently. Cambiara Group has seen the consequences of organizations investing in technology before fully understanding the processes they're trying to improve. AI doesn't eliminate that problem. In some cases, it can make it worse.


Before introducing AI into important business activities, companies need to understand:

  • What information is being shared with the technology?

  • Who has access to that information?

  • Are AI-generated results being reviewed before they're used?

  • What happens when the system gets something wrong?

  • Who is accountable for decisions made with AI assistance?

  • What are the actual costs, including integration, training and ongoing support?

  • Are copyright, privacy and record-keeping obligations being addressed?

  • Are people being replaced before the underlying work has been properly redesigned?


The last point deserves particular attention.

Replacing people with technology before understanding the work itself is rarely a recipe for successful transformation. AI should be an opportunity to rethink how work gets done, not simply a faster way of doing things the way they've always been done.


Start with the work, not the technology


When Cambiara Group considers where an organization should begin with AI, the starting point isn't a list of AI products or a presentation about the latest developments.

It's a much simpler question:


Where is the organization losing time, money, quality or customer confidence in the way it currently operates?


That is where the conversation about AI should begin.


Look at the processes that matter to the business. Where are people spending too much time searching for information? Where are handoffs creating delays? Which activities are repetitive? Where are errors occurring? What decisions are being made with incomplete information?


Then consider whether AI can make a meaningful difference. A practical approach involves six steps:


  1. Understand the work. Identify the business processes that have the greatest impact on cost, revenue, quality, safety and customer experience.

  2. Evaluate the opportunities. Look at the potential value of each application alongside the risks involved. Not every opportunity deserves the same priority.

  3. Establish the rules. Decide what information can be shared, which tools are approved and what security and privacy controls are required.

  4. Keep people involved. There are decisions where human judgement must remain central. AI can provide information and recommendations, but accountability still belongs to people.

  5. Measure the results. Did the application actually improve productivity, reduce errors, improve service or lower costs? If it didn't, why is the organization continuing to invest in it?

  6. Prepare people for the change. Training should go beyond teaching employees how to write prompts. People need to understand how to challenge AI, validate its output and recognize when it shouldn't be used.


This last point is particularly important. Organizations can spend considerable amounts of money on technology and surprisingly little on helping people understand how their jobs will change.


That's a familiar transformation problem, and AI isn't going to make it disappear.


There is an environmental cost to consider

There is another dimension to AI adoption that companies can't afford to ignore. Even if an organization doesn't own a data centre, its use of AI contributes to demand for computing infrastructure somewhere.


Every query, model interaction and data processing activity has a resource requirement.


That doesn't mean every application is environmentally irresponsible. In fact, some AI applications can help organizations reduce energy consumption, improve transportation routes, minimize waste and prevent equipment failures.


But companies should understand the costs rather than assume they don't exist. A few questions are worth asking:

  • Are we using a large model when a smaller one would do?

  • Does this application really need real-time AI?

  • How much information are we storing, and for how long?

  • Are we duplicating information across multiple platforms?

  • Do our technology providers report energy consumption and environmental impacts?

  • Are we using AI for activities that create little or no meaningful business value?


This is particularly relevant in Canada, where energy availability, electricity generation, water resources and infrastructure capacity vary considerably between provinces and municipalities. A project that makes economic sense in one location may have a very different impact somewhere else.


Companies need to recognize that responsible technology adoption includes understanding its broader consequences.


Not every company needs to become an AI company


This is one of the more important distinctions in the whole discussion. There is a tendency to talk about AI as though every organization needs to become an AI leader. Cambiara Group doesn't share that view.


Some companies need to become better users of AI. Others need to adapt AI to their own business. A relatively small number need to build their own AI capabilities and products.

These three positions can be described as:


AI Users

These organizations adopt established tools to improve everyday business activities. This is likely the right starting point for many small and mid-sized businesses. The focus should be on safe adoption, employee training and practical results.

AI Shapers

These companies go further by integrating AI with their own information, processes and customer experiences. They may have complex operations, valuable internal data or specific regulatory requirements. Their focus is on integration, governance and measurable performance.

AI Builders

These organizations develop AI products, train specialized models or operate significant AI infrastructure. They need the financial resources, technical expertise and risk management capabilities to support that investment.


The mistake is trying to be something you're not.


A logistics business may need to become an AI shaper because routing, scheduling and asset utilization are central to its competitiveness. A local services company may gain enormous value from becoming a disciplined AI user. A software company may need builder-level capabilities because AI could fundamentally change its products. There is no prize for spending more money on AI than the business needs to.


Governance shouldn't become another organizational obstacle

The word governance can sometimes make people's eyes glaze over. It often sounds like another layer of approvals, policies and committees designed to slow things down.

But good governance should do the opposite.


It should give people the confidence to use AI appropriately without having to guess where the boundaries are.


Companies need to establish some basic ground rules. Which tools can employees use? What information must never be entered into public AI platforms? Which applications require security, privacy or legal review? Who owns the technology? How are errors reported? And which decisions must always have a human being involved?


Not every application needs the same level of control.


Using AI to brainstorm ideas or summarize non-sensitive information is very different from using it to assess a job applicant, make a lending decision or provide medical advice. The level of oversight should reflect the potential consequences.


This doesn't have to become a complicated exercise. A practical, risk-based approach is often enough to get started. The important thing is that employees understand the rules and management takes responsibility for enforcing them.


So, how deep should companies go?


The answer isn't to be cautious about everything, nor is it to move as quickly as possible simply because the technology is available.


Cambiara Group advocates disciplined adoption.


Over the next 12 to 24 months, most organizations should be doing a few things:

  • Moving beyond random experimentation.

  • Identifying a manageable number of high-value applications.

  • Establishing clear rules around data, privacy and security.

  • Giving employees the skills to use AI effectively.

  • Integrating AI into selected business processes.

  • Measuring the actual business benefits alongside the costs and risks.

  • Understanding the environmental and community implications.

  • Reviewing their approach as the technology, regulations and public expectations evolve.


The risk of not going deep enough is that an organization gradually loses ground in productivity, cost, customer expectations and decision-making. The risk of going too deep, too quickly, is spending money without creating value, exposing the organization to unnecessary risks and potentially undermining the trust of the very people needed to make the transformation successful.


There is a balance to be found.


And finding it requires more than a technology strategy. It requires an understanding of the business, its people, its operating environment and the outcomes it is trying to achieve.


AI is no longer simply a technology decision. It is an organizational decision.


The companies that get this right won't necessarily be the ones that use the most AI. They'll be the ones that understand where AI makes a difference, where it doesn't, and how to make the two work together.


Cambiara Group - November 2026


 
 
 

Comments


CambiaraGroup Logo Round - August 14 2026
Cambiara Group gratefully acknowledges that our community connects across the traditional, ancestral, and unceded territories of First Nations, Inuit, and Métis Peoples throughout what is now known as Canada. We recognize and honour the enduring relationships Indigenous Peoples have with these lands, waters, cultures, languages, and communities, and we celebrate the wisdom, strength, and stories they continue to share

@ Cambiara Group 2024

linkedin
GoogleLogo.png
FacebookLogo.jpg
bottom of page