Erick Ngwiri has spent 17 years helping organizations turn complex, scattered information into decisions they can act on. Within that time he has seen why most organisations still get data wrong, and what it takes to fix it, one clear problem at a time.
In an era of rapid digital transformation, Erick Ngwiri believes the future belongs to organizations that prioritize data and technology.
“How companies operate will be remarkably transformed by AI and data. So, the way we respond to this shift will determine what the next wave of organisations will look like,” he says.
Sadly, Erick notes that most organisations are still not able to effectively use the data they have in decision making. He faults several factors.
Top of his list is that most organisations can’t clearly define the problem they are trying to solve and how insights from data can help early on.
“Decisions generally fall into two categories: those guided by a mix of data and gut sense, and those that must be entirely data-driven to determine the next step. Too many organizations rely heavily on gut instinct, or bring data in long after critical decisions have already been made,” he says.
The second obstacle is the inability to effectively translate data collected.
“You lose people when you start talking about metrics. You have to translate those numbers into language and insights people can actually understand, articulate, and use,” he says.
Beyond poor problem definition and weak translation, Erick points to another common trap of gathering too much data.
“Companies try to collect everything to build a treasure trove of information, but then they can’t make sense of it quickly or reliably. AI has only accelerated this problem by making vast amounts of information instantly accessible, resulting in information overload,” he says.
As a solution, Erick urges teams to name only a handful of things that genuinely matter for a given problem before adding anything else.
"When you focus on a very specific set of things, it is a lot easier to come to a decision once you build those things, and then you can add additional sets of information," he says.
Data meets the real world
Having served in disaster response with, Erick draws on firsthand experience to show how data challenges unfold during real crises.
“If a cyclone or tsunami hits the Kenyan coast, a mechanism is activated where the government leads the response alongside the United Nations Office for the Coordination of Humanitarian Affairs (OCHA). Multiple UN, local, and international organizations then step in to assist. The biggest problem, however, is that we lack a coordinated way to measure the extent of the damage," he says.
Erick witnessed a solution to this exact problem during a major disaster in Southern Africa."
“One project I worked on was when Cyclone Idai hit the port city of Beira, Mozambique, in March 2019. I was part of a UN team deployed to support the government’s disaster agency, INGC, with information management. We built map-based platforms that tracked the cyclone’s path, layered wind speeds, and mapped structural damage to buildings, schools, health centers, and roads. Across the hundreds of agencies on the ground, it became one of the key tools used to determine where to go next and who does what,” he says
Another challenge during real-time crises is that responders rarely have an accurate picture of how events are unfolding
“The best source of real-time ground information often comes from local radio stations and local news outlets. We are currently piloting AI to synthesize these varied sources into concise, draft situation reports. After a human review, responders can immediately use these reports to plan operational deployments,” he says.
AI readiness
While Erick advocates for the use of AI in data, he also warns that organisations must meet certain prerequisites to safely and effectively adopt AI. The first is data itself.
"AI works off data. You need to have data you can trust, that's structured in the right way, that's usable, and that actually is the right kind of information to feed into the AI. Focusing on data governance, trust, and quality is what ultimately determines if you are ready,” he says.
Erick also advises organisations to invest in data and AI literacy even if it’s on a basic level.
“Data literacy is essential so employees understand what these tools mean for their specific roles. You don’t need to be a data scientist, but as routine tasks become automated, you must understand what the data is telling you to stay effective,” he says.
The third and perhaps biggest pillar of AI readiness according to Erick, is deciding what role AI should play in an organisation
“You have to be clear about what you are asking these tools to do and where they fit into your operations, including what AI can handle and what it simply can’t. That way, rather than imposing a sweeping change overnight, you are bringing your team along on the journey,” he says.
