The CEO Who Returned from a Conference Convinced He Needed AI

The CEO of a Caribbean insurance company returned from an industry conference in Miami with a conviction and a problem. The conviction was that artificial intelligence would transform the insurance industry and that his company needed to adopt it or risk obsolescence. The problem was that he had no idea where to start.

The conference had been dominated by presentations from global insurers deploying AI at scale: machine learning models that assessed risk and priced policies in milliseconds, natural language processing systems that read and classified claims documents automatically, chatbots that handled customer enquiries without human intervention, and predictive analytics platforms that identified fraud patterns across millions of transactions. The presenters spoke of data lakes, neural networks, large language models, and algorithmic underwriting. The CEO had taken careful notes and returned to Kingston energised and anxious in equal measure.

His company employed 140 people. The IT department consisted of three staff: a manager who maintained the company’s core insurance system, a network administrator, and a help desk technician. The company’s data resided in a fifteen-year-old policy administration system, a separate claims management system that did not integrate with the policy system, an accounting package, and approximately 200 spreadsheets maintained by various departments. The annual IT budget was US$280,000, of which approximately US$220,000 was consumed by licence fees, hardware maintenance, and internet connectivity. The discretionary technology investment budget was approximately US$60,000.

The CEO convened a meeting with his executive team and announced that the company needed an AI strategy. The CFO asked what the expected return on investment would be. The chief operating officer asked which processes AI would improve. The IT manager asked where the data would come from. The head of underwriting asked whether AI would replace her team. Nobody asked the question that should have been asked first: what specific business problems are we trying to solve, and is AI the right tool to solve them?

Six months later, the company had spent US$45,000 on an AI proof of concept that attempted to build a machine learning model to predict claims frequency using the company’s historical data. The proof of concept failed — not because the technology was flawed, but because the company’s data was insufficient: inconsistent coding of claims categories across years, missing data fields, no linkage between policy characteristics and claims outcomes, and a dataset that was too small and too noisy for machine learning to extract meaningful patterns. The US$45,000 had produced a PowerPoint presentation documenting why the project had failed and a set of recommendations for data quality improvements that would cost an estimated US$180,000 and take eighteen months to implement.

The CEO’s frustration was understandable: “Every conference I attend tells me AI is transforming our industry. We tried it. It didn’t work. Are we too small for AI?”

The answer was no. The company was not too small for AI. But it had approached AI in the wrong sequence, attempting to deploy sophisticated machine learning before establishing the data foundations, the process clarity, and the organisational readiness that AI requires. The company’s mistake was not unique — it is the mistake that Dawgen Global observes most frequently among Caribbean enterprises that attempt to adopt AI without a practical framework for doing so.

AI Is Not One Thing

The first misconception that Caribbean mid-market enterprises must overcome is that AI is a single, monolithic technology that either applies to the enterprise or does not. In reality, artificial intelligence encompasses a spectrum of technologies with vastly different complexity, data requirements, cost, and applicability to mid-market enterprises.

Process Automation (Low Complexity): Robotic process automation and rule-based automation are the entry point for most mid-market enterprises. These technologies automate repetitive, structured tasks — data entry, invoice processing, report generation, email routing, document formatting, and reconciliation — without requiring the large datasets or sophisticated algorithms that machine learning demands. A Caribbean insurance company that automates its premium notice generation, its policy renewal reminders, and its monthly regulatory reporting extracts immediate value from automation without building a data lake or training a machine learning model. Process automation is achievable with modest budgets, existing data, and commercially available tools.

Intelligent Document Processing (Medium Complexity): AI-powered document processing uses optical character recognition, natural language processing, and classification algorithms to read, understand, and route documents. For Caribbean enterprises that process high volumes of paper or PDF documents — insurance claims, loan applications, customs declarations, invoices, contracts — intelligent document processing can dramatically reduce manual handling time and error rates. The technology has matured significantly and is available as cloud-based services that do not require the enterprise to build or train its own models.

Conversational AI and Customer Service Automation (Medium Complexity): AI-powered chatbots and virtual assistants can handle routine customer enquiries, process simple transactions, provide account information, and route complex requests to human agents. For Caribbean enterprises with customer service operations — banks, insurance companies, telecommunications providers, utilities — conversational AI reduces response times, extends service availability beyond business hours, and frees human agents to handle the complex interactions that require judgement and empathy. Modern conversational AI platforms, including those powered by large language models, can be deployed with relatively modest configuration effort and can be trained on the enterprise’s specific products, policies, and procedures.

Descriptive and Diagnostic Analytics (Medium Complexity): Before an enterprise can predict the future with AI, it must understand the present and the past. Descriptive analytics — dashboards, visualisations, and reports that show what has happened — and diagnostic analytics — analysis that explains why it happened — are the analytical foundation on which predictive AI is built. Caribbean enterprises that invest in business intelligence platforms, data visualisation tools, and the data governance that makes analytics reliable create the foundation for the predictive and prescriptive AI capabilities that follow.

Predictive Analytics and Machine Learning (Higher Complexity): Machine learning models that predict outcomes — claims frequency, customer churn, demand forecasting, credit risk, fraud probability — require clean, structured, historically consistent data in sufficient volume for the algorithms to identify meaningful patterns. This is where the Caribbean insurance CEO’s proof of concept failed: the data was not ready. Predictive AI is powerful but demands investment in data infrastructure, data quality, and data governance before the models can deliver reliable results. For most Caribbean mid-market enterprises, predictive AI is a Year 2 or Year 3 capability, not a starting point.

Generative AI and Large Language Models (Emerging): The emergence of generative AI — large language models capable of drafting text, summarising documents, generating code, answering questions, and assisting with analysis — has created new opportunities for Caribbean enterprises of every size. Unlike traditional machine learning, generative AI does not require the enterprise’s own training data: it leverages pre-trained models accessed through APIs or commercial platforms. Caribbean enterprises can deploy generative AI to assist with report drafting, email composition, policy document analysis, customer communication, research, and internal knowledge management with minimal infrastructure investment. Generative AI is the most immediately accessible advanced AI capability for Caribbean mid-market enterprises.

The Practical AI Adoption Sequence for Caribbean Enterprises

Dawgen Global has developed a practical AI adoption sequence designed specifically for Caribbean mid-market enterprises — enterprises with limited IT resources, constrained budgets, and data environments that are imperfect but improvable. The sequence moves from immediate value to progressive capability, ensuring that each stage builds the foundation for the next.

Stage 1 — Automate the Obvious (Months 1–6): Identify the five to ten most time-consuming, repetitive, rule-based processes in the enterprise and automate them. These are the processes where staff spend hours on tasks that follow predictable patterns: generating reports from data that already exists in systems, sending notifications based on dates or triggers, formatting and distributing documents, reconciling data between systems, and processing routine transactions. The tools for this stage are commercially available, affordable, and do not require AI expertise to implement. The return on investment is immediate and measurable — hours saved, errors reduced, and staff capacity freed for higher-value work.

Stage 2 — Deploy Generative AI for Knowledge Work (Months 3–9): Introduce generative AI tools to support knowledge workers across the enterprise. Equip staff with access to AI assistants that can help draft correspondence, summarise lengthy documents, research regulatory requirements, generate first drafts of reports, and answer internal knowledge queries. Establish clear usage policies — what data can and cannot be shared with AI tools, what outputs require human review, and how AI-assisted work should be documented. This stage requires minimal infrastructure investment but delivers significant productivity improvements across professional and administrative functions.

Stage 3 — Build the Data Foundation (Months 6–18): Invest in the data infrastructure that advanced AI requires. This means cleaning and standardising existing data, establishing data governance practices, integrating systems that currently operate in silos, and beginning to collect the structured, consistent data that predictive analytics will need. This stage is the most important and the most frequently skipped — enterprises that jump from automation to machine learning without building the data foundation repeat the Caribbean insurance CEO’s expensive mistake.

Stage 4 — Implement Business Intelligence (Months 12–24): Deploy dashboards and analytics that give management real-time visibility into the enterprise’s performance. Replace the monthly PDF reports with interactive dashboards that show current sales trends, customer behaviour, operational metrics, financial performance, and the key performance indicators that drive business decisions. Business intelligence is not AI in the narrow sense, but it is the analytical capability that makes predictive AI possible — and it delivers substantial value in its own right.

Stage 5 — Deploy Predictive AI (Months 18–36): With clean data, integrated systems, and a functioning analytics capability, the enterprise is ready for predictive AI. Start with one or two high-value use cases where the data is strongest and the business impact is clearest: demand forecasting, customer churn prediction, credit risk assessment, or fraud detection. Deploy models iteratively — start simple, measure performance, refine, and expand. Predictive AI is a continuous improvement capability, not a one-time implementation.

What Caribbean Enterprises Get Wrong About AI

Starting with the Technology Rather Than the Problem: The most common mistake is selecting an AI technology and then searching for a problem to apply it to. Effective AI adoption starts with the business problem: what decision needs to be improved, what process needs to be faster, what customer experience needs to be enhanced, what risk needs to be better managed? The technology follows the problem, not the reverse.

Underestimating the Data Requirement: Machine learning requires data — clean, structured, historically consistent data in sufficient volume. Caribbean enterprises that attempt machine learning with small, messy, inconsistent datasets produce models that do not perform reliably. The data foundation must be built before the models can be trained.

Expecting AI to Replace Judgement: AI augments human judgement; it does not replace it. A machine learning model that predicts claims frequency does not decide whether to underwrite a risk. A chatbot that handles routine enquiries does not resolve a complex customer complaint. AI handles the repetitive, data-intensive, pattern-recognition tasks that humans do slowly and inconsistently, freeing humans to focus on the judgement-intensive, relationship-dependent, creative tasks that AI cannot do.

Treating AI as a Project Rather Than a Capability: AI adoption is not a project with a start date and an end date. It is a capability that the enterprise builds incrementally, maintains continuously, and improves iteratively. The enterprise that implements one AI solution and then stops has not transformed — it has completed a project. Transformation requires sustained investment in data, technology, skills, and organisational change.

Ignoring Change Management: AI changes how people work. Employees who have performed tasks manually for years may resist automation. Managers who have made decisions based on experience may distrust data-driven recommendations. Customers who have interacted with humans may be sceptical of chatbots. Successful AI adoption requires deliberate change management: communication, training, involvement of affected staff in the design process, and leadership commitment to supporting the transition.

Dawgen Global’s AI Adoption Advisory

Dawgen Global’s AI Adoption Advisory is designed specifically for Caribbean mid-market enterprises that want to capture the value of AI without the costly missteps that the fictional insurance CEO experienced.

AI Readiness Assessment: Dawgen Global evaluates the enterprise’s current data environment, technology infrastructure, process maturity, and organisational readiness for AI adoption. The assessment identifies the highest-impact AI opportunities, the data gaps that must be addressed, and the practical sequence for adoption. The deliverable is a prioritised AI adoption roadmap with realistic timelines, investment requirements, and expected returns.

Process Automation Design and Implementation: Dawgen Global identifies, designs, and implements process automations that deliver immediate value. Our automation advisory covers process selection, tool evaluation, implementation, testing, and the change management that ensures adoption.

Generative AI Strategy and Governance: Dawgen Global helps enterprises develop strategies for deploying generative AI across knowledge work functions, including usage policies, data governance frameworks, vendor evaluation, and the governance structures that ensure AI is used responsibly and effectively.

Data Foundation Advisory: Dawgen Global designs and supports the implementation of the data infrastructure — data governance, data quality, systems integration, and analytics architecture — that advanced AI requires. Our data advisory recognises that Caribbean enterprises operate with imperfect data environments and designs practical, achievable data improvement programmes.

Predictive AI Use Case Development: For enterprises that have built the data foundation, Dawgen Global supports the identification, development, and deployment of predictive AI use cases. Our approach is iterative: start with the highest-value use case, deploy a minimum viable model, measure performance, refine, and expand.

AI for Every Caribbean Enterprise

The fictional insurance CEO’s question — “Are we too small for AI?” — has a clear answer: no Caribbean enterprise is too small for AI. But every Caribbean enterprise must adopt AI in the right sequence, at the right pace, and with the right expectations. The enterprise that begins with process automation and generative AI, builds its data foundation, deploys business intelligence, and then progresses to predictive analytics will capture real value at every stage. The enterprise that skips stages — that attempts machine learning without data readiness, that deploys chatbots without process clarity, that invests in AI without change management — will repeat the expensive failures that give AI a reputation it does not deserve.

AI is not magic. It is a set of tools that, applied to the right problems in the right sequence with the right data, create measurable business value. Caribbean mid-market enterprises have an advantage that large enterprises often lack: the agility to adopt new technologies quickly, the proximity between leadership and operations that enables rapid decision-making, and the scale at which the impact of even modest AI improvements is immediately visible. The enterprises that use these advantages — that approach AI practically rather than aspirationally — will discover that AI is not a technology reserved for enterprises with unlimited budgets and dedicated data science teams. It is a capability that every Caribbean enterprise can build.

Begin Your AI Journey

Dawgen Global invites Caribbean enterprises to take the first practical step toward AI adoption. Our AI Readiness Assessment evaluates where your enterprise stands today, identifies the highest-impact opportunities, and delivers a phased adoption roadmap designed for Caribbean mid-market realities.

Request a proposal for Dawgen Global’s AI Readiness Assessment and Adoption Advisory. Email [email protected] or visit www.dawgen.global to begin the conversation.

DAWGEN GLOBAL | Big Firm Capabilities. Caribbean Understanding.

Request a proposal for Dawgen Global’s AI Readiness Assessment and Adoption Advisory.

Email: [email protected]

Web: www.dawgen.global

About Dawgen Global