Artificial intelligence, once the subject of speculative fiction and limited practical deployments, has now become a pillar of business transformation. In today’s enterprise landscape, the role of the Chief Information Officer (CIO) has fundamentally changed, as AI adoption shifts from isolated pilots to integrated, organization-wide strategies. Recent insights from Chris Loake, Group Information Officer at Hiscox, a global insurance provider, not only offer a window into the transformative journey for businesses exploring Microsoft 365 Copilot (M365 Copilot) and generative AI—but also highlight the nuanced realities of CIO strategy, organizational change, and operational value. This feature explores the Hiscox journey, critically analyzing successes and challenges, and aims to provide actionable insights for IT leaders navigating similar paths.
When Chris Loake assumed the role of Group Information Officer at Hiscox in late 2023, the company had already dipped its toes into Microsoft’s AI ecosystem. The insurance industry, characterized by data-heavy processes and stringent regulatory requirements, was well-positioned to experiment with AI. But, as Loake notes, the initial push toward tools like Copilot was driven less by a detailed business plan and more by broad recognition that generative AI could not be ignored. As he put it, the “ChatGPT revolution” forced leadership to proactively pursue AI, lest they fall behind in a rapidly shifting digital environment.
Loake recalls the early days of the internet, making a compelling analogy: "Everybody has to have an AI strategy. It’s like having an internet strategy." His caution, however, is clear. Many early internet strategies offered little practical value, he observes, so an AI strategy must become a living North Star—one that is both aspirational and pragmatic, guiding the development of a truly AI-enabled business rather than serving as empty rhetoric.
For Hiscox, this North Star means flexibility and optionality—advancing a “multi-provider environment” that allows the IT team to blend multiple tools, models, and capabilities in response to evolving needs across organizational units. Loake emphasizes the importance of an architecture and operating model that supports agility, allowing Hiscox to “place some bets and be flexible along the way.” This iterative, bet-hedging mindset reflects the current best practices in AI adoption: avoid vendor lock-in, foster experimentation, and pivot as new opportunities or disruptive threats emerge.
The analogy to email is apt: nearly every modern organization relies on email, but few consider the tool itself a differentiator. The true value, Loake suggests, lies not just in deploying AI, but in designing customer journeys and operational processes that harness AI for empathy, speed, or insight at critical moments. For Hiscox, the “moment of truth” comes during claims processing—a touchpoint where automation alone cannot guarantee customer satisfaction. Empathy and nuanced service remain vital, so the use of AI must be strategic, enhancing rather than replacing the human touch.
Other insurance firms may use AI to optimize for high volume and low cost, focusing on transactional speed above all. In contrast, Hiscox seeks to use AI to support its premium, service-oriented proposition, aligning technology with its brand values. This duality—commoditization versus differentiation—frames the difficult questions CIOs must answer: Where can AI systems offer unique value, and where are they simply catching up to industry benchmarks?
The six-month pilot involved 300 users and quickly illuminated a range of user behaviors and attitudes. Loake identifies three key cohorts:
Conversely, roles with fewer meetings, less documentation demands, and minimal need for information synthesis gained little. This aligns with broader research suggesting that the productivity benefits of generative AI tools are heterogeneous, with creative and collaborative roles at the cutting edge of value realization.
Critically, this uneven value distribution did not dampen enthusiasm for a broader rollout. Instead, it informed a more nuanced, needs-driven adoption model—scaling up intent-based licensing (where users must apply for access and demonstrate purpose), and doubling down on supporting “AI champions” to drive uptake and maximize value.
As Hiscox approached mass deployment—ultimately enrolling over 3,000 employees across fourteen countries—the diversity of outcomes became even more pronounced. Loake notes, “Once you’ve democratized tools like M365 Copilot and given them to everybody, some people will use them loads and get loads of value, but some will get limited value.” The key insight is that universal access alone does not guarantee universal impact; success depends on the interplay of tool, task, and talent.
Loake is adamant that access remains an important symbolic equalizer: “You do want your users to feel like they’ve got the consumer-type tools in their hands, and it’s important for the executives to recognize what’s happening in their business.” This psychological component—ensuring inclusion and a sense of digital parity—can be undervalued in pure ROI metrics but remains critical for morale and long-term cultural transformation.
However, the journey from hype to sustained value is neither easy nor linear. Success hinges on a CIO’s ability to blend vision with pragmatism—to champion AI adoption where it matters most, measure impact honestly, and respond dynamically to new realities.
Loake’s reflections serve as both a blueprint and a cautionary tale. The “AI North Star” must be kept in sight, but CIOs also need the courage to course-correct as evidence emerges. For organizations willing to commit to strategic flexibility, empower grassroots innovation, and confront practical limitations head-on, the dividends can be substantial. For those that treat AI as mere checkbox compliance, the risk is stagnation as competitors harness generative tools to reshape the industry.
Ultimately, developing a CIO strategy for artificial intelligence is not about technology for its own sake—it’s about building a business that can learn, adapt, and win in an AI-powered world.
Source: Computer Weekly Interview: Developing a CIO strategy for artificial intelligence | Computer Weekly
The AI Imperative: A New Strategic Paradigm for the CIO
When Chris Loake assumed the role of Group Information Officer at Hiscox in late 2023, the company had already dipped its toes into Microsoft’s AI ecosystem. The insurance industry, characterized by data-heavy processes and stringent regulatory requirements, was well-positioned to experiment with AI. But, as Loake notes, the initial push toward tools like Copilot was driven less by a detailed business plan and more by broad recognition that generative AI could not be ignored. As he put it, the “ChatGPT revolution” forced leadership to proactively pursue AI, lest they fall behind in a rapidly shifting digital environment.Loake recalls the early days of the internet, making a compelling analogy: "Everybody has to have an AI strategy. It’s like having an internet strategy." His caution, however, is clear. Many early internet strategies offered little practical value, he observes, so an AI strategy must become a living North Star—one that is both aspirational and pragmatic, guiding the development of a truly AI-enabled business rather than serving as empty rhetoric.
For Hiscox, this North Star means flexibility and optionality—advancing a “multi-provider environment” that allows the IT team to blend multiple tools, models, and capabilities in response to evolving needs across organizational units. Loake emphasizes the importance of an architecture and operating model that supports agility, allowing Hiscox to “place some bets and be flexible along the way.” This iterative, bet-hedging mindset reflects the current best practices in AI adoption: avoid vendor lock-in, foster experimentation, and pivot as new opportunities or disruptive threats emerge.
Commodity vs Differentiator: Where AI Delivers Value
A crucial theme running through Loake’s account is the commoditization of off-the-shelf AI systems. As more organizations deploy similar AI functionalities, the risk is that such tools become as undifferentiated as email—necessary, but not a direct driver of competitive advantage.The analogy to email is apt: nearly every modern organization relies on email, but few consider the tool itself a differentiator. The true value, Loake suggests, lies not just in deploying AI, but in designing customer journeys and operational processes that harness AI for empathy, speed, or insight at critical moments. For Hiscox, the “moment of truth” comes during claims processing—a touchpoint where automation alone cannot guarantee customer satisfaction. Empathy and nuanced service remain vital, so the use of AI must be strategic, enhancing rather than replacing the human touch.
Other insurance firms may use AI to optimize for high volume and low cost, focusing on transactional speed above all. In contrast, Hiscox seeks to use AI to support its premium, service-oriented proposition, aligning technology with its brand values. This duality—commoditization versus differentiation—frames the difficult questions CIOs must answer: Where can AI systems offer unique value, and where are they simply catching up to industry benchmarks?
The M365 Copilot Pilot: Anatomy of an Enterprise AI Rollout
Hiscox’s adoption of Microsoft 365 Copilot offers a detailed case study in measured AI deployment. Loake shares that the pilot began with just a few licenses, distributed across different business units. Importantly, licenses were not assigned top-down, but instead made available for users to apply for—an approach designed to encourage buy-in and ensure that those with genuine interest and clear use cases would engage most fully.The six-month pilot involved 300 users and quickly illuminated a range of user behaviors and attitudes. Loake identifies three key cohorts:
- Problem Solvers: Users who would deploy Copilot to solve a specific challenge and, if the tool failed, move on to other solutions without lingering frustration.
- Curious Innovators: Highly motivated individuals interested in exploring AI’s potential, willing to “chip away” at new features, and serving as self-appointed “AI champions.”
- Skeptics: Users for whom Copilot proved marginally useful or even distracting, with little interest in further experimentation.
Where Copilot Excelled—and Where it Fell Flat
A crucial finding from the Hiscox pilot was that Copilot’s benefits were not evenly distributed across all job functions. Certain roles—such as those tasked with creating new documents from scratch, consolidating large volumes of information, or participating in frequent meetings—reported the greatest productivity gains. Copilot’s ability to generate summaries, action items, and synthesis documents proved indispensable for these users.Conversely, roles with fewer meetings, less documentation demands, and minimal need for information synthesis gained little. This aligns with broader research suggesting that the productivity benefits of generative AI tools are heterogeneous, with creative and collaborative roles at the cutting edge of value realization.
Quantifying the Impact: Measured Productivity Gains
Quantitative data from the pilot phase provides substance to anecdotal success stories. According to Loake, approximately 15% of users reported saving an hour a day—an enormous potential efficiency gain in wage and opportunity cost terms. Twenty percent logged savings of 30 minutes per day, while another fifth cut 10 to 15 minutes from daily workloads. A quarter of participants saw only marginal time savings, a result consistent with the distribution of productive use across job types and user engagement levels.Critically, this uneven value distribution did not dampen enthusiasm for a broader rollout. Instead, it informed a more nuanced, needs-driven adoption model—scaling up intent-based licensing (where users must apply for access and demonstrate purpose), and doubling down on supporting “AI champions” to drive uptake and maximize value.
Scaling Up: Lessons from Organization-Wide Deployment
Following the successful pilot, Hiscox expanded Copilot access to 1,000 users, again focusing on voluntary application to encourage intrinsic motivation. The emergence of “AI champions” became even more important at scale. These champions, embedded in each department, functioned as internal trainers and helpers, running workshops and hosting drop-in sessions to ease colleagues’ AI adoption.As Hiscox approached mass deployment—ultimately enrolling over 3,000 employees across fourteen countries—the diversity of outcomes became even more pronounced. Loake notes, “Once you’ve democratized tools like M365 Copilot and given them to everybody, some people will use them loads and get loads of value, but some will get limited value.” The key insight is that universal access alone does not guarantee universal impact; success depends on the interplay of tool, task, and talent.
Loake is adamant that access remains an important symbolic equalizer: “You do want your users to feel like they’ve got the consumer-type tools in their hands, and it’s important for the executives to recognize what’s happening in their business.” This psychological component—ensuring inclusion and a sense of digital parity—can be undervalued in pure ROI metrics but remains critical for morale and long-term cultural transformation.
A Critical Analysis: What’s Working, and Where are the Risks?
The Hiscox case study reveals both notable strengths and potential pitfalls in contemporary CIO strategies for AI.Strengths
- Flexible, Agile Architecture: By refusing to overcommit to a single provider or model, Hiscox can continually optimize its AI approach as new solutions emerge.
- User-Centric Adoption: Encouraging self-selection rather than mandating deployment improves initial buy-in and helps surface early champions.
- Internal Champions and Community: Creating a distributed network of AI ambassadors drives adoption, enhances peer support, and amplifies organizational learning.
- Data-Driven Iteration: Careful measurement of productivity gains informs targeted scaling and supports resource allocation decisions.
- Strategy with Substance: Rather than relying on generic “AI for the sake of AI” rhetoric, Hiscox is explicit about the need to link AI deployment to business differentiators—like improved claims handling and executive understanding.
Risks and Limitations
- Commoditization Reduces Differentiation: As off-the-shelf AI tools become widely adopted, their power as competitive weapons may wane. Organizations must continually ask where AI truly adds unique value, and where it simply levels the playing field.
- Uneven Value Distribution: Not all roles benefit equally. Blanket rollouts can result in wasted licenses or, worse, employee disengagement if everyone is expected to use a tool without clear rationale.
- Adoption Plateau: Early enthusiasm may eventually flatten, especially if ongoing support and training are not prioritized or if user expectations outstrip tool capabilities.
- Security, Privacy, and Compliance: AI deployment in regulated industries such as insurance raises ongoing concerns about data protection, explainability, and regulatory alignment. While these issues were not deeply explored in the initial Hiscox pilot, they remain evergreen areas of risk.
- Change Fatigue: As digital transformation accelerates, user fatigue and resistance pose real threats to continued innovation, underscoring the need for thoughtful change management and clear communication.
Practical Takeaways for CIOs: Building a Robust AI Strategy
For CIOs seeking to replicate Hiscox’s successful adoption—or learn from its roadblocks—the following action steps emerge:1. Articulate a Living AI Strategy
A credible AI strategy is neither static nor vague. It should:- Emphasize optionality and agility (avoid vendor lock-in)
- Prioritize pilots that tie directly to value-driving use cases
- Set clear decision points for scale-up or pivot
2. Champion User-Led Innovation
- Allow interested employees to opt in to early pilots
- Recruit and empower “AI champions” to lead internal knowledge propagation
- Foster a living knowledge community (via Teams, Yammer, or other platforms)
3. Target Value-Add Areas
- Identify job roles and workflows where AI naturally multiplies productivity (creation, synthesis, meeting-heavy functions)
- Avoid forcing adoption in roles where value is unclear, but remain open to organic experimentation
4. Measure, Iterate, Repeat
- Collect robust data on usage patterns, time saving, and qualitative feedback
- Use pilot insights to refine rollout plans, support structures, and training investments
- Recognize that some users or cohorts will gain more than others, and use this data to inform ongoing investments and incentives
5. Manage Risk Proactively
- Stand up clear protocols for data protection, privacy, and compliance from day one
- Collaborate with vendors and internal risk management to stress-test AI frameworks regularly
- Communicate openly about known limitations and future uncertainties
The Outlook: AI Strategy at an Inflection Point
The Hiscox experience mirrors a broader inflection point for enterprise technology leaders. AI is no longer a backroom experiment or a niche productivity hack; it is a foundational capability, analogous to the early internet in its potential to redefine the future of work.However, the journey from hype to sustained value is neither easy nor linear. Success hinges on a CIO’s ability to blend vision with pragmatism—to champion AI adoption where it matters most, measure impact honestly, and respond dynamically to new realities.
Loake’s reflections serve as both a blueprint and a cautionary tale. The “AI North Star” must be kept in sight, but CIOs also need the courage to course-correct as evidence emerges. For organizations willing to commit to strategic flexibility, empower grassroots innovation, and confront practical limitations head-on, the dividends can be substantial. For those that treat AI as mere checkbox compliance, the risk is stagnation as competitors harness generative tools to reshape the industry.
Ultimately, developing a CIO strategy for artificial intelligence is not about technology for its own sake—it’s about building a business that can learn, adapt, and win in an AI-powered world.
Source: Computer Weekly Interview: Developing a CIO strategy for artificial intelligence | Computer Weekly
Last edited:
