By Niel Coetzee, Technical Director at redPanda Software

AI is a foundational technology in the retail sector. Deployed to enhance personalisation, operational efficiency and optimise workflows and planning, AI solutions are rapidly becoming practical, enterprise-scale deployments that allow for deeper human-machine collaboration and customer satisfaction. They’re also proving critical to ensuring retailers remain ahead of evolving customer expectations and accelerating ecommerce growth – global ecommerce retail sales are expected to reach $3.52 trillion this year, accounting for nearly a quarter of all retail sales.

Behind the sleek storefront apps and loyalty platforms, however, lies a fragile reality. Many retailers are weighed down by brittle code, outdated processes and development cycles that can’t keep up with demand. This knocks into reliability and resilience. When core systems are held back by inefficiencies or poor code quality, the impact is immediate. Outages at point of sale, delays in stock updates or sluggish online checkouts translate into frustrated customers and lost revenue.

Download BAO E-MAGAZINE

The high cost of poor code

It’s easy to underestimate the amount of time and budget consumed by preventable errors in code. Manual requirement gathering, repetitive documentation and lengthy review processes slow projects down before a single line of value-adding functionality is even deployed. However, retailers want their projects delivered in months at a high quality so they don’t risk falling behind competitors that can launch their new loyalty features, adapt their pricing models and respond swiftly to operational disruptions.

It’s a balancing act. Retailers need the speed of service delivery, but they can’t risk that this comes at the cost of reliability. They’re running in a race they didn’t choose with horses they didn’t buy against competitors who want to win the whole track. Projects that stretched over 18 months are expected in three to six, and they have to deliver new features, stabilise old systems and do it without adding costs or risks. And without compromising on quality of delivery.

In retail, mistakes, poor quality and bottlenecks cost more than time. They erode customer loyalty and trust and result in lost sales. The irony isn’t lost on most – at the very moment retailers need to be fast, their systems make them slow.

A new layer in the development chain

The good news is that AI is changing the narrative. Chatbots can distil hours of discussion into structured requirement documents which provide product teams with a head start. In design, AI can generate architecture diagrams and build estimation tools in days compared to the usual weeks. And, in development, automated code reviewing tools are capable of inspecting every line, flagging risks and suggesting fixes, while MCP (model code protocol) is changing how different AI tools work together to achieve specific results.

The best part is that they don’t tire and they maintain the momentum needed by retailers to compete effectively in the current market. These tools don’t replace development teams, they instead remove friction, reduce repetition and help companies build stronger code foundations.

Human-machine collaboration

For retailers, these developments mean that projects land faster and features reach customers sooner. The systems that run stores and ecommerce platforms are more resilient, and code quality becomes inherent. The AI in tandem with developers and decision-makers smooths over the rocky complexity that comes with most development cycles by reducing effort and errors.

Staying up to date with the AI trends as they evolve and grow is essential, ensuring that tools add value to clients beyond simplifying manual efforts. The goal is to embed AI carefully throughout the product development lifecycle and to validate outputs with human oversight, ensuring that security and ethics remain non-negotiable. In practice, this means fewer defects in production and more predictable delivery with systems capable of scaling reliably across every store node.

AI’s role in software delivery in retail is still maturing with different tools serving different functions. Today, the smartest approach is to combine them. Tomorrow, emerging standards will allow AI agents to chain tasks seamlessly from requirement to deployment. As AI evolves, retailers need to rely on trusted partners capable of strengthening their systems, so they are in a position to innovate and adapt quickly.

AI is already changing how retail systems are built and maintained. The tools will evolve, but the direction is clearly towards software that adapts quickly, scales reliably and is delivered at pace. For leaders, the decision is whether to tackle that future now or be forced to catch up later. Retail resilience, and retail growth, will depend on the strength of the code beneath the surface.

When an employee leaves an organisation, most leaders focus on succession, handovers and HR paperwork. But behind the scenes, another risk often goes unchecked: the “shadow employee”. Retaining access to company systems long after they’ve left, these ex-staff members pose a serious cybersecurity threat that can lead to data breaches, financial loss and reputational damage – even if everyone parted ways with smiles, hugs and pizza.

“The shadow employee phenomenon is more common than many realise, particularly in organisations with high staff turnover or fragmented and cloud based systems,” asserts Anna Collard, SVP Content Strategy and Evangelist at KnowBe4 Africa.

She says it often goes undetected because access management tends to focus more on onboarding than offboarding. “When IT and HR operate in silos or access isn’t centrally tracked, it’s easy for credentials, third-party accounts or shadow IT tools to be overlooked,” Collard comments. “It shouldn’t be seen as just a technical issue; it’s a human one, too, where attention to digital hygiene and processes are lacking.”

Risks of rogue access

The threat of shadow employees was brought into sharp focus in 2023 when a US company suffered a major data leak traced back to a former IT consultant whose access to internal drives was never revoked. The incident exposed client information and resulted in a six-figure (dollar denominated, no less) settlement on top of contract losses.

Download BAO E-MAGAZINE

“The risks are serious and multifaceted,” states Collard. “They encompass operational risk, reputational risk and financial risk.” In terms of operational risks, she explains that outdated access rights can disrupt workflows, expose sensitive information or allow unauthorised changes to systems – even inadvertently.

Regarding reputational risk, a data breach caused by a former staff member can erode customer trust and damage brand credibility. “Ex-employees with active credentials can intentionally or unintentionally cause data breaches, leak sensitive information, manipulate internal systems or impersonate staff,” she says.

“In some cases, disgruntled employees may delete or sabotage critical data,” she elaborates. “Even if there’s no malicious intent, the mere presence of active credentials outside of an organisation’s control creates vulnerabilities that threat actors can exploit, especially through credential stuffing or phishing.”

The last risk to organisations involves financial risk. “Rogue access can result in regulatory fines, legal costs and lost revenue,” she says. The reason why these security breaches occur is that many organisations treat offboarding as an almost “optional HR thing”, not a cybersecurity event. “They fail to conduct thorough access audits or delay revoking credentials across all systems, especially cloud platforms, collaboration tools and unmanaged software-as-a-service (SaaS) applications,” argues Collard.

Why robust offboarding is key

To close the loop and reduce the shadow employee threat, organisations must build strong offboarding processes that bridge HR and cybersecurity. “It starts with a shared mindset: offboarding must be seen as a collaborative security process, not just an admin task,” she comments.

Another important step is to automate deprovisioning to revoke access in real-time. “Integrating identity and access management (IAM) tools and involving security or risk teams in offboarding governance can also help,” she says. Other action items include performing regular access reviews to identify dormant or unauthorised accounts and educating managers to close the gap on shadow IT.

“Make line managers accountable for flagging all tools and systems used by exiting staff and track unofficial tools in your access control system,” she recommends. The HRM Report also noted that “Shadow AI” use is a growing concern across Africa, with 46% of organisations still developing formal AI policies while staff increasingly use generative AI from work networks without checks on credentials or information sharing. This lack of governance around new technologies further underscores the need for robust offboarding processes that account for all forms of access, not just traditional systems.

In conclusion, Collard maintains that former employees shouldn’t keep the digital keys to your organisation’s kingdom. “As the workplace becomes more hybrid and decentralised, organisations must rethink offboarding as a critical component of cybersecurity hygiene,” she emphasises.

By Rory Bosman, Chief Sales & Marketing Officer at Ecentric

In 2024, Black Friday stopped being an imported promo day in South Africa and started dictating terms. It was a supercharged, concentrated surge of shopping that reset expectations and rewired retail playbooks. According to the Ecentric Payment Systems’ Black Friday Index 2024, the four-day Black Friday to Cyber Monday window punched well above its weight with online transactions jumping from 7.9% to 10.3% – a 30.4% increase – while in-store revenue more than doubled to 11.1%.

The shape of the event has changed from a vague outline of some stores offering discounts to a firm discount buster with extraordinary sales volumes. Globally, Black Friday has become an almost month-long experience, something that has already started in South Africa, with Black November bringing promotions throughout. But these tend to be the start of the tail that leads to the weekend itself, where, as the report found, shopping has become concentrated over four days, whereas in the past, shopping would continue well into December. For the first time, the peak moved from early December, and the strongest returns were felt over the weekend.

Download BAO E-MAGAZINE

This was a global trend as well. Shopify merchants alone processed more than $11.5 billion over the Black Friday weekend at a 24% year-on-year increase, and according to Adobe Analytics, Black Friday sales were up to $10.8 billion from $9.8 billion in 2023. Shoppers are saving their big-ticket purchases for that late November window and holding back on discretionary spending until the deals drop.

Yet the numbers only tell a part of the story. What made 2024 distinctive wasn’t just the scale and timing of spending, but the ways in which consumers chose to shop. Mobile devices have become the most popular storefront with millions of consumers transacting on their phones while on the move. At the same time, physical stores showed they could still pull in the crowds if they offered something beyond discounts. Festive atmospheres, interactive demonstrations and exclusive in-store promotions gave shoppers reasons to queue.

This resurgence of in-store retail in South Africa is particularly interesting. Ecentric’s Index showed that in-store revenue over the Black Friday weekend more than doubled its share of the holiday total. It suggests that while online convenience is still a winner, South Africans are enjoying the energy of a shopping trip when the experience feels worthwhile.

The rules of engagement are slowly changing. Mobile has to be rapid and easy with payment systems that can handle the extraordinary loads. Customers don’t want to lose deals because a payment system fails or websites can’t handle the load. This does leave a bad taste, and customers are vocal about disappointing experiences or the perception of false advertising. In-store experiences have to be equally smooth and satisfying. Sure, there will be queues, but let these be accompanied by working systems and easy access to stock and unexpected experiences.

Black Friday in 2025 is going to be as much about consumer delight as it is going to be about deals and discounts.

Customers also don’t see Black Friday as a novelty anymore. It has to work hard for its money. Flexible payment options are expected to become more popular, so price-sensitive consumers can benefit from the deals but stay within their budgets. Personalisation, powered by AI, is also going to get sharper and more effective with tailored promotions sent to devices in real time.

This year, intelligent tech, smart payments, faster checkouts and bigger deals are going to lead the way, but they will be supported by immersive experiences, deeper personalisation, and more excitement as retailers build on last year to create something completely transformative.