Wayfinder:
Knowing where you're going is the easy part
KEY FINDINGS AT A GLANCE
Most insurers know their technology path
Fewer know how to get there at pace.
Strong foundations and optimistic outlook amongst respondents
77%
have a documented technology strategy
73%
have a named executive accountable for delivery
82%
are confident about their technology position three years from now
Stalling execution
29%
report major advances in the past 12 months
74%
take months or longer to integrate a new technology component
87%
have deployed AI somewhere in the business, but only 11% have scaled it across multiple areas
Most insurers have a clear direction and are confident about what comes next. The organisations most confident about their future are distinguished less by strategy and more by two practical capabilities: faster integration and greater trust in their data.
Knowing where you're going is the easy part
Ask a room of Australian insurance leaders whether they have a technology strategy and almost every hand goes up. More than three-quarters of the respondents that participated in this survey have a documented strategy covering the next three years or more, and nearly three-quarters have named an executive accountable for setting and delivering it. On paper, the general insurance sector as a collective knows where it wants technology to take it, has written the route down and put someone in charge of the journey.
Then ask about the progress they’ve made in the past year, and only 29% report major advances, with the rest describing steady incremental progress or less. A sector almost universally equipped with a strategy has for the most part only inched forward. This report illustrates what is happening in that gap between the beginning and the end – the documented strategy and the delivered outcome.
This is Finity’s first Wayfinder Report, built on an independent, anonymous survey of senior general insurance leaders across Australia and New Zealand. The respondents are the people who set technology direction and answer for its delivery: chief executives, and operational, strategy, technology and transformation leads. What they describe is consistent enough across the sample to read as a shared condition rather than a set of isolated difficulties.
For years the strategic question in insurance technology was ‘what’ to do: which technology platforms and what architecture. That question is now largely answered. From our experience, insurers have mostly decided on modular architecture rather than traditional monolith platforms, are focusing on interconnectivity and flexibility, and are actively investing in this transition. With more flexible foundations in place, the next challenge sits downstream — in execution.
Organisations that integrate in weeks or less see their investment pay off 87% of the time; those that take a year or more, just 30%
AI is deployed almost everywhere and scaled almost nowhere.
Leaders see skills and expertise as critical to progress, yet finding and developing them remains a challenge
Each is a facet of the same problem, and none of them is a problem of strategy; it’s the pace and mechanics of turning intent into working technology ecosystems.
This report follows these themes across five chapters, starting with the strategy-to-execution gap itself and why structural accountability hasn't necessarily translated into functional delivery. From there it turns to AI, where the distance between deployment and scale exposes uncertain foundations in architecture and data, then to capability, where need outstrips supply. The fourth chapter discusses why the sector is confident about the next three years despite only modest recent progress. The last asks what separates the confident organisations from the rest, and the answer is, surprisingly, not about strategy but the mechanics of delivery.
CHAPTER 1
The strategy-to-execution gap
Technology projects are rarely only technology projects. Behind every platform migration or AI pilot are decisions about process, ownership and execution that determine whether technology delivers the intended outcome.
Drawing on a targeted survey of senior general insurance leaders across Australia and New Zealand, this report describes a sector that has largely settled the strategy question and is now contending with turning intent into delivered outcomes.
The groundwork is in place, with over three-quarters of respondents (77%) having a documented technology strategy, split between 42% with a full execution roadmap and milestones, and 35% who have documented the strategy without a clear plan. Nearly as many – 73% – have a named executive accountable for technology delivery rather than strategy alone. These are the markers of organisations that have done the thinking, assigned responsibility and committed to a direction.
Alignment with the broader business is more mixed, with 34% describing their technology strategy as fully aligned and a further 42% as mostly aligned, suggesting a discrepancy between where businesses are headed as a whole, and deploying technology with the same vision in mind.
Momentum appears to be slow to build, with only 29% of respondents reporting major advances over the last 12 months, while 53% describe steady incremental progress and 18% report limited progress.
Structural accountability, functional delivery
A named delivery owner is common, major recent progress is not
This is not because accountability is absent: 73% reported a named delivery owner, and this appears to matter. None of the organisations without a named delivery owner reported major progress and, among those with no owner, shared ownership or unclear ownership, only 29% said their technology investment has delivered expected outcomes. What this suggests is that an owner is necessary, but not sufficient on its own. We can infer that often structural accountability is widely in place, but the operating model that allows the named owner to deliver change often is not, and the machinery beneath them frequently does not move at the pace the role implies.
Investment outcomes tell a compatible story, with just over half of respondents (53%) saying their technology investment met or exceeded expectations (with 47% meeting expectations and 6% exceeding expectations). 31% said it had partially met expectations, 8% said it was too early to say, and 8% felt it had not met expectations. When asked about their challenges to delivery, leaders name budget (29%), talent and skills gaps (16%), vendor delivery (16%) and the pace of change in the market (15%). This suggests that disappointing investment outcomes may be less about failed technology and more about a gap between what was scoped and what was delivered.
"In our experience, when an insurer feels a technology project doesn’t deliver to expectations, there is usually a combination of common factors: overly optimistic expectations and/or objectives that shifted mid-delivery. This is intensified where the timeframes are long: mistakes are harder to undo, time and effort sunk is greater and frustration is higher."
Marcello Negro, Director and Head of Product Management, Finity
How long it takes to integrate a new technology component
Almost three-quarters of respondents (74%) said it takes months or longer to integrate a new technology component, with 37% taking months and a further 37% taking more than a year
“We find that protracted timeframes for implementation tend to inflate expectations for delivery and perceived payoff. When expectations are set too high, the pressure to deliver can feed on itself – driving scope creep, shifting objectives and an ever-expanding delivery burden.
Edward Owen, Principal and Head of Product Development & Innovation, Finity
Integration speed tracks closely with perceived return on investment. Among organisations that integrate a new component in weeks or faster, the large majority say their technology investment met or exceeded expectations. Among those taking months, that share falls to a little over half. Among those taking more than a year, it drops to less than a third. The pattern runs in one direction across all three groups: the faster an organisation can integrate, the more likely its investment delivered to expectations.
Faster integration, better returns
Share of organisations whose technology investment met or exceeded expectations, by integration speed
Beyond the connection
"Why integration takes as long as it does is the part most often misread as a purely technical issue. Often system size and complexity drive time to implement, yes, but there are usually other downstream contributing factors. The timeline is set as much by the surrounding work as by the connection itself: agreeing new processes, establishing data ownership, redefining governance, and building confidence in unfamiliar ways of working – all at scale. However, when a modular approach is taken, transformation happens in chunks, not only reducing the complexity and effort, but minimising the operational disruption experienced at any one time.
For example, one of our clients recently replaced their rating engine – they had the new algorithm live and in-market in eight weeks. The insurer could decouple the old component and connect the new one through the existing interface. Their architecture made the integration a contained piece of work rather than a system-wide one, and without it, the same change is measured in quarters, not weeks."
Edward Owen, Principal and Head of Product Development & Innovation, Finity
The survey results highlight the difference between architecture approaches. Among organisations running a modular, connected architecture, half can integrate a new component in weeks or faster, and only one in five takes more than a year. Among those on siloed or monolithic systems, almost none integrate that quickly, and half take more than a year.
A near industry-wide move away from monolithic platforms is evidenced by a strong relationship between strategic advancement and architecture type. 60% of those with no strategy are using monolithic architecture, of whom 80% were uncertain about their future technological positioning.
It is clear that the architecture an organisation runs is closely tied to the speed at which it can change, and the speed at which it can change is closely tied to satisfaction with outcomes. That chain, from how systems are built, to how fast they can be extended, to return on investment, is a consistent theme across the survey. The way organisations source technology reinforces the reading. A third (34%) buy platforms and build or augment on top of them, 24% co-develop with external partners, 16% run a mix that varies by initiative, 15% build primarily in-house and 11% buy off-the-shelf. Taken together, the sector favours tailored solutions over either extreme, neither fully bespoke nor purely out-of-the-box.
"The preference towards tailored solutions makes sense to me: you avoid the effort and risk involved with building a system, without losing the ability to configure your environment to your business’s own specific needs and ways of working. However, it raises the stakes on integration and partnership, because a tailored, multi-component architecture only delivers at pace when the pieces connect cleanly and the partners building them understand the domain."
Marcello Negro, Director and Head of Product Management, Finity
What separates intent from outcome is the speed and mechanics of execution, and integration speed is where that separation is most visible. It is also, as the later chapters show, one of the clearest markers of which organisations are confident for good reason.
"We can more rapidly adapt our processes to make the most of enterprise technology than we can adapt technology to meet our legacy processes…every aspect of our operation is impacted by the reliance on our technology vendor to make system changes to implement any kind of operational change."
In response, to "In your own words, what would most change your confidence?"
CHAPTER 2
AI maturity and the data foundation beneath it
Across insurers today, AI is deployed almost everywhere, scaled almost nowhere. Nearly nine in ten organisations (87%) have deployed AI or machine learning somewhere in the business, however, only 11% have scaled it across multiple areas.
Most organisations sit in the middle of that distribution. Alongside the 11% that scaled across multiple areas, 31% have scaled in one or two, 44% are piloting without having scaled, and 15% are still exploring options before piloting. The 58% who have piloted or explored without scaling are largely testing it for live applications and have not yet found a route from a working pilot to production across the business.
AI shows up most in operational areas and least in the ones where an error carries legal, regulatory or disciplinary weight, a pattern that roughly equates to ‘the cost of getting it wrong’. A wrong answer in back-office automation is a productivity cost; a wrong answer in a compliance or claim decision can mean a regulatory breach or an unfair outcome for a customer.
Where AI is deployed today
Deployed almost everywhere; still shallow in oversight roles
Biggest barrier to scaling AI
Integration leads; budget and vendor maturity tie for last
When asked to name the biggest barrier to scaling AI, leaders put integration with existing systems first at 32%, followed by talent at 23%. Data quality (13%), organisational culture (10%), regulatory uncertainty (8%), budget (7%) and vendor maturity (7%) completed the list. Budget, the most-named challenge in executing technology strategy overall at 29%, ties for last place as an AI-specific barrier, level with vendor maturity at 7%. The thing executives are most likely to name as their predominant limitation is not what constrains deploying AI at scale.
When technology transformation slows, it is often underpinned by business case and budget constraints. However, this survey suggests that when it comes to AI, the obstacle sits further along the value chain with the systems themselves. The biggest impediment to scaling AI is technology that doesn’t connect cleanly enough to move a model from one part of the business to another.
The architecture landscape in insurance supports this narrative. Just under half of organisations (48%) operate with some degree of systems siloing, 29% with limited connectivity between systems and 19% on a single monolithic platform, while 13% are actively moving from monolithic to modular. Almost the same proportion, 44%, are piloting AI without being able to scale it. A model that works in claims does not travel to underwriting on its own; without connective architecture beneath it, each new deployment is closer to a fresh build than an extension. Getting past this stalled scaling means fixing the architecture first, which relies on integration that is time, labour and disruption-intensive for these organisations.
While data quality was listed third as a barrier to scaling AI, organisations that trust their data are more than twice as likely to have scaled AI somewhere in the business than those that don't, 52% against 23%.
In practice: Turning information into insight
One of the less visible challenges in scaling AI is that many of the most valuable signals in insurance sit outside structured datasets. Claims notes, legal decisions, medical reports and free-text descriptions often contain insights that influence reserving, risk assessment and operational decision-making, but extracting those insights manually can be difficult to scale.
Through its work with insurers and compensation schemes, Finity has seen growing interest in applying AI to convert these sources into structured, analysable information. In one example, AI was used to analyse approximately 3,000 workers' compensation appeal and review decisions, extracting injury characteristics, severity indicators, evidence signals and outcome patterns into a consistent dataset. Prior to this, individual judgments could take around 20 minutes each to review manually.
In another application, AI-derived insights from hundreds of case notes per participant were combined with structured claims and payment data and analysed using Finity's Dynamic Clustering methodology, enabling decision-makers to identify emerging risk patterns and participant groups with materially different future care outcomes. The analysis revealed drivers of outcomes that were not visible through structured data alone, including housing instability, rehabilitation disengagement and behavioural issues.
These examples highlight a broader theme reflected in this research. The value of AI is rarely in the technology itself. It comes from turning information that was previously difficult to access or interpret into trusted insights that can be embedded into operational models, governance and decision-making. As organisations look to scale AI, the challenge often lies less in deploying models and more in integrating those insights into day-to-day workflows.
"Data quality is a prerequisite to moving AI beyond pilots into production. Without trust, pilots can’t eventuate to more than a tech demo. AI models are not magic, and waiting for stronger frontier models is rarely the solution – AI will only produce reliable results when fed reliable data."
Dylan Neenan, Head of AI Solutions, Finity
Dylan Neenan, Head of AI Solutions, Finity
The barriers to scaling AI are whether the systems underneath can carry a pilot into production, and whether the data feeding them can be trusted.
CHAPTER 3
Building capability for what's next
While budget is what insurance leaders selected as their main challenge to executing technology strategy, at 29%, when asked what single factor would most increase progress over the next 12 months, the survey revealed a slightly different insight. At 24%, talent and capability was chosen first, just ahead of budget at 21%.
From constraint to enabler
Budget leads the list of challenges; talent leads the list of what would most help
Capability is the lever insurers name most often for progress, and a lack of it is a constraint they hit frequently trying to deliver, with talent and skills gaps coming in as the joint-second challenge in executing strategy (16%). Talent is also the second-biggest barrier to scaling AI (23%).
Just under half of respondents (45%) name technology investment as their primary response to workforce and capability pressure, using automation and tooling to reduce reliance on individual expertise. Alongside that, 34% are hiring and building graduate pipelines, 34% are turning to external partners and advisers, and 31% are running structured knowledge-transfer programmes.
The paradox is that organisations are turning to technology to relieve capability pressure, but talent is the second-largest barrier to scaling that same technology. It would suggest that organisations short on capability turn to technology to compensate, then find they need capability to deploy that technology at scale. The solution and the constraint draw on the same resource, which means capability-building must run alongside the technology work rather than after it.
More than a quarter of organisations (27%) have not addressed workforce and capability pressures in any systematic way, suggesting that intent and action have not yet met. However, more than half (52%) are concerned about knowledge loss through retirement over the next five years, split between 39% somewhat concerned and 13% very concerned, while 42% say it is not a current priority. That is interesting when considering the sector's demographic trajectory: the Insurance Council of Australia expects almost 30% of the current workforce to reach or exceed retirement age by 2030.¹ The survey shows that the sector recognises knowledge loss as a growing long-term challenge but doesn’t generally view it as an immediate crisis. What we can’t tell from the data is whether we’re seeing measured concern and good mitigation or a looming 'retirement cliff'.
"An implementation partner that knows insurance."
In response to "In your own words, what would most change your confidence?"
The vendor relationship comes up twice in survey responses and, while it wasn’t first in either, it is notable, with vendor delivery a joint-second in execution challenge (16%), and better vendor or partner support the third-most named accelerator of progress (19%).
“In our experience, insurers find their vendor relationships to be both a frustration and a source of aspiration. Our clients tell us that they value and need external partners, but find that partners who do not understand insurance well enough or cannot deliver at the pace required are a source of disappointment and frustration. The combination of technology expertise and genuine domain knowledge is essential for successful project delivery. We also find that insurers are comforted when their vendors show true understanding of their business, governance and processes, because along with the opportunities represented by AI, comes an insurance-sector-specific set of risks."
Marcello Negro, Director and Head of Product Management, Finity
Past technology transitions were forgiving of knowledge loss, because the old system kept running alongside the new one and experienced people were still in place to catch what the new system got wrong. The move to AI may not be forgiving in the same way. As AI takes on more of the building itself, the underwriting logic, the pricing rules and the first-pass decisions, the experienced staff who would once have written and checked that work are the same staff retiring out of the sector.
The risk is not only in the loss of knowledge, it is that the work which used to build judgement, the years of writing rules and catching errors, is precisely the work being automated, so the path by which a junior becomes senior enough to supervise the system starts to disappear. An organisation can buy a model that writes underwriting rules; however, it cannot as easily buy the accumulated judgement to know when those rules are wrong, and on current trends, it may struggle to grow that judgement internally. That is not a problem the survey can measure, but it highlights an important question about capability: beyond headcount, how is expert judgement retained when the judgement is what gets automated?
As organisations build the capacity to reach their strategic technology goals, how can they keep hold of the judgement that makes technology safe to rely on? Capability is held up as an imperative for insurers, and its scarcity the constraint they continue to face. Over the next few years insurers will be required to find balance in this tension and futureproof their knowledge, judgement and governance, while benefiting from the advances offered by technology, specifically AI.
CHAPTER 4
Confidence and outlook
The most striking single number in the survey is not about architecture or AI, but about belief. More than four in five organisations (82%) are confident or very confident that they will be better positioned technologically in three years than they are today, split between 45% confident and 37% very confident. Only 5% are not confident. While the sector is faced with a range of challenges, as described in the earlier chapters, there is a robust purported level of optimism about what comes next.
High confidence, modest recent progress
The report's central tension in two numbers
While 29% made major advances in the last 12 months, the majority, 53%, describe steady incremental progress, and 18% report limited progress. Though the sector has mostly inched forward over the past year, there is overwhelming confidence about the next three years.
“Optimism about a three-year horizon can be entirely reasonable, even when the last year was incremental, particularly where the groundwork has been laid and the results are still to come. Go back over Finity’s articles from the past few years and you’ll see that we’ve always advised a judicious approach to technology transformation, especially transformation involving AI. There are many shiny new things out there, but not all will deliver meaningful value, and taking the time to really explore, test and evaluate can save future pain – even when there is a lot of noise suggesting innovation is accelerating without you.
“Many of our clients have taken this approach: budgets have been approved, strategies documented, major programmes started, or in some cases, their funded programmes are about to deliver, so I believe this confidence is well-founded. If the confidence is based on less concrete footing, and you can’t articulate what will rapidly move you forward, it might be worth asking what or who can give you the confidence"
Marcello Negro, Director and Head of Product Management, Finity
The open-text responses give the confidence question further colour. Asked what would most change their confidence, leaders point overwhelmingly to execution rather than resources:
“We need project management, rejection of scope creep, better evaluation of the meaning of MVP and prioritisation and utilisation of SMEs.”
"Redefined strategy, with regular true benefit realisation & reporting.”
“Clear program governance and good learning and development programs on new technology builds.”
These concerns do not suggest a lack of strategy or budget. They are the concerns of organisations that are in the thick of project delivery and are watching to see whether their strategy and budget will yield the desired returns. Open-text responses also repeatedly tie optimism to one programme or decision, and therefore also bear the weight of that initiative's risk, and slips if the programme does. That connects back to the investment-outcomes picture from the opening chapter, where 39% of organisations saw their technology investment only partially meet expectations or miss altogether. Some of the gap between three-year confidence and recent progress may trace to a similar root: expectations set optimistically or scoped against objectives that moved during delivery. That pattern holds here too, as the organisations that integrate quickly are also the ones whose investments met expectations, while the slowest integrators are the ones most likely to have been disappointed. Confidence that rests on a delivery engine already turning is on firmer ground than confidence waiting on a pace that has not yet materialised.
CHAPTER 5
What separates the confident from the rest
Every chapter so far has articulated the same themes: strategy is almost ubiquitously in place; delivery is the dividing factor.
AI is deployed almost everywhere and scaled almost nowhere. Capability is the lever most reached for and the resource most often short. However, rather than being separate findings, perhaps they are one.
When you split the confident organisations from the rest, what differs between them is having the mechanics for delivery in place, and showing progress and momentum in delivering.
What separates the confident from the rest
The plan barely divides the two groups; pace and data trust divide them sharply
73%
of low-confidence organisations take more than a year to integrate a new component.
73%
of confident organisations trust the data they rely on.
27%
of uncertain organisations trust their data.
Low-confidence organisations are far more likely to take more than a year to integrate a new component, 73% against 29%. On data trust the gap is wider still, with 73% of confident organisations trusting their data, against 27% of the uncertain ones. These two dimensions divide the sector by outlook more cleanly than anything else in the data, and neither is a matter of strategy.
Confident organisations are somewhat more likely to have a documented strategy, 80% against 64%, but that difference is not decisive, and it is dwarfed by the integration and data-trust gaps. Nearly two-thirds of the least confident organisations have a documented strategy. Having a high level, documented set of objectives and a roadmap is common, and it is not what distinguishes the organisations that believe they will arrive from the ones that doubt it.
When we look at the other factors measured in the survey, we see similar sentiment. On investment outcomes, 55% of high-confidence organisations report that their technology investment met expectations, however 45% of the low-confidence group said their investment did not meet expectations at all (a further 36% called it only partially met). There is a difference between delivery that has built a track record and delivery that has repeatedly disappointed. On AI, we see the pattern repeated; across the whole sample 15% are still exploring rather than piloting, but among low-confidence organisations that figure is 45%, three times the rate. Organisations that cannot move on technology broadly cannot move on AI either.
Confidence in the three-year horizon tracks to the ability to deliver, not to the presence of a strategy. The organisations pulling ahead are the ones that integrate quickly and trust the data they build on. A strategy can be written in a quarter, whereas architecture that integrates in weeks, and data clean enough to build on, are the product of years of decisions. Time spent building solid foundations up front results in years of innovation and growth at pace.
It seems the sector has largely finished choosing directions and is now competing on execution, and execution increasingly means connecting systems and changing the processes and governance around them at the same pace as the technology itself.
Marcello Negro, Director and Head of Product Management, Finity
