Shabana Patel of Medihelp spoke to ITWeb TV about the situation that many organisations find themselves in – a great deal of executive excitement about AI, but struggling to translate vision into measurable results. This episode was recorded at the ITWeb Data Insights Summit 2026.
#ITWebdatainsights #Data #ai
Organisations are hyped up about AI and many are eager to apply the technology in their operations, but the reality is they struggle to turn their AI vision into tangible outcomes.
This is according to Shabana Patel, who heads up data strategy and AI at Medihelp.
Speaking to ITWeb TV on the sidelines of the ITWeb Digital Insights Summit 2026, hosted recently at The Forum in Bryanston, Patel said before any action is taken to adopt and apply AI, there are important factors to consider.
“You know the saying – when all you have is a hammer, everything looks like a nail! Sometimes AI is seen as a one-size-fits-all solution and we need to be thoughtful in our application, because if we apply AI within situations where it is not economically viable or operationally appropriate, then the cost and complexity will exceed the value.
“So, if you think about a niche edge case that affects only 1% of exposure, is an AI solution possible? Yes, it’s technically possible, but your costs are going to exceed your savings.”
Patel stressed the importance of AI literacy, which is generally defined as the skills and knowledge required to understand and effectively use AI.
“Gartner predicts that if we focus on AI literacy for executives, then by next year, those organisations will have a 20% higher financial performance than the organisations that don’t.”
Patel said that this literacy empowers executives to ask tough questions about the AI solution to critically evaluate whether the technology is viable or not.
“Rather than simply slapping an AI solution onto any given problem or challenge, decision-makers ask and answer questions like what is the associated risk? Is there bias? Is there explainability? Can we measure this? What is the cost?”
Shabana Patel, head of data strategy and AI at Medihelp. (Photograph by Lesley Moyo)
Accountability and ownership
“AI is often seen as a project. There are different milestones… the data and analytics teams own a piece, the IT team owns a piece, the business owns a piece, but who takes full ownership of this outcome?”
Patel believes there needs to be a mindset shift to viewing AI as a product, not a project. “We have one owner who is accountable for this full lifecycle and typically what that would mean is that this owner’s performance will be tracked based on how this AI tool is being implemented, executed and adopted in the business.”
Patel said every AI use case must have an owner, someone who is responsible for the AI solution end-to-end.
“It’s difficult to achieve in practice; we need governance that enables this. You need to understand who the decision owners are and what is the risk that is acceptable in that context. You must also build on the assumed trust [and] not just that we assume trust, we must design trust.”
This speaks to the influence of a company’s culture on its AI readiness. Patel said when a user or frontline staff cannot understand the rationale behind an AI recommendation, they don’t trust it.
“Part of the culture would be this AI literacy drive, but also things like explainability and interpretability. What goes a long way here is something we call a fusion role. This is a hybrid person who can speak tech and speak business. They are really like force multipliers because they explain the technical aspects behind an AI recommendation to the business.
“If you keep it transparent and you have these refinement loops with your users, it is very easy for them to then adopt the recommendation you are suggesting,” Patel added.
The Medihelp executive believes this fusion role will grow in significance as businesses enforce their AI strategies and stakeholders continue to seek clarity around the technology.
Patel said while research like the McKinsey State of AI 2025 Survey does list several negative consequences of AI use, businesses are more aware of these things and are mitigating risks more aggressively.
“That definitely means better things for AI. We need to be aware of its potential and aware of its risks, and if we mitigate these risks, we get to a better picture.”
