Seen Here: Professor Yudhvir Seetharam – Chief Analytics Officer for FNB Commercial Photo Credit: Supplied
Businesses must rethink their pursuit of speed and efficiency in an increasingly automated world, with deliberate human oversight emerging as a critical safeguard against financial, reputational and strategic risks.
In an opinion piece published on 14 October 2026, Professor Yudhvir Seetharam, Chief Analytics Officer at FNB Commercial, argues that while artificial intelligence (AI) and data-driven systems have transformed business operations, removing friction from every process is not necessarily the best approach.
Seetharam says automation is particularly effective when dealing with low-risk, repetitive decisions where outcomes are predictable, processes are well understood and errors can be corrected quickly.
However, he warns that decisions with significant implications for customers, revenue, reputation or long-term business strategy require a different approach — one that allows time for scrutiny, contextual assessment and human judgement.
The answer, he argues, is “friction by design”.
The principle calls for the level of scrutiny applied to a decision to be proportionate to its potential risk. Rather than introducing unnecessary bureaucracy, the approach aims to ensure that important decisions receive the attention they deserve before action is taken.
For small businesses, this responsibility often rests directly with the owner. While AI can help gather information, analyse trends, compare options and generate recommendations, accountability for major decisions ultimately remains with the person running the business.
As organisations expand, decision-making becomes more complex, with management teams, specialist departments, governance structures and boards providing additional perspectives and challenging assumptions.
Seetharam believes AI should play a similar supporting role — contributing valuable insights without becoming the final authority.
Although AI systems can analyse vast amounts of information, their recommendations may not fully account for ethical considerations, reputational consequences or circumstances that fall outside the available data.
Technology providers also carry business risks
The growing reliance on external AI platforms and technology providers introduces another layer of responsibility for businesses.
Seetharam notes that organisations may not develop every AI tool they use, but they remain accountable for how these technologies affect their customers.
This means technology providers form part of a business’s broader risk profile. Customers are unlikely to separate the actions of a technology supplier from those of the organisation that chose to use its services.
Businesses therefore need to evaluate their technology partners carefully and consider whether they are comfortable attaching their reputation to the providers they trust.
The principle extends beyond selecting reliable technology. It also requires organisations to maintain appropriate oversight of how automated systems operate and the consequences they produce.
Why customer complaints matter
According to Seetharam, friction by design should extend beyond the initial decision-making process to include the ongoing review of automated outcomes.
Businesses often assess performance through aggregate indicators such as error rates and complaint volumes. While these measures are useful, they can conceal problems experienced by smaller groups of customers.
A process may appear successful across the majority of transactions while still producing unfair or unsatisfactory outcomes for certain individuals.
Individual complaints can therefore provide valuable insight into flaws in decision-making logic, unintended consequences and shortcomings that broader performance indicators may fail to reveal.
They may also highlight systemic problems affecting customers who have not formally complained.
For businesses increasingly dependent on automated systems, paying attention to these exceptions is essential to identifying weaknesses and improving customer experiences.
Governance must evolve with AI
Seetharam further argues that businesses cannot afford to treat their governance frameworks as permanent solutions in a rapidly changing technological environment.
As AI capabilities develop and adoption expands, organisations must continually reassess which decisions require additional scrutiny and how automated outcomes should be monitored.
Risk frameworks that were appropriate six months ago may no longer adequately reflect emerging technologies, evolving customer expectations or the scale at which automated systems now operate.
This calls for continuous evaluation, clear accountability and a willingness to adjust safeguards as circumstances change.
Ultimately, Seetharam does not frame the challenge as a choice between human judgement and automation. Instead, he argues that effective businesses must understand where each approach adds the greatest value.
As automation becomes more sophisticated, competitive advantage may depend less on how much friction a business eliminates and more on whether it knows when to retain it.
In an age of automation, the ability to pause, question and exercise sound judgement could become one of the most valuable tools in effective business risk management.
