By 2035, no one in life insurance still talks about “running the model”.
That phrase belongs to an earlier era; a time when modelling was an event rather than a capability, when results arrived hours or days after questions were asked, and when insight lagged behind decision-making.
In 2035, modelling is simply there. Always on. Always available. And quietly shaping almost every material decision a life insurer makes.
The Death of the Monthly Run
The most obvious change is the disappearance of the monthly close as a modelling focal point.
Regulatory reporting still exists, of course, but it no longer defines the rhythm of actuarial work. Models update continuously as experience emerges, assumptions evolve, and strategies change. Capital positions, profitability measures, and risk metrics are observable states, not retrospective calculations.
The idea that senior management would wait three weeks to understand the impact of a decision feels as strange as waiting for a printed balance sheet. In 2035, if you ask “What happens if we reinsure this block?”, the answer arrives while the question is still being discussed.
From Models as Artefacts to Models as Environments
In 2025, most modelling conversations still revolve around artefacts: model files, assumption sets, scenarios, and outputs.
By 2035, models are no longer things you open. They are environments you enter.
Actuaries don’t “configure runs”. They explore spaces. They nudge assumptions and watch responses ripple across capital, cashflow, risk, and shareholder value in near real time. The boundaries between pricing, valuation, ALM, and capital modelling blur, not because regulation changed, but because computation finally stopped being the constraint.
The model doesn’t just answer questions. It pushes back.
AI as the First Actuary You Talk To
Every modeller in 2035 works with an AI counterpart, not as a decision-maker, but as a constant analytical companion.
Before an actuary touches the model, AI has already identified which assumptions historically mattered in similar situations, highlighted regions of non-linear behaviour, flagged scenarios that produced counterintuitive outcomes, and suggested where judgement rather than computation is needed most.
AI doesn’t replace actuarial expertise. It filters reality so that expertise is applied where it can actually add value.
Extreme Transparency Becomes Non-Negotiable
One surprise of the 2035 world is how little tolerance there is for black boxes.
When models run continuously and inform real decisions in real time, opacity becomes operational risk. Regulators, boards, and capital providers demand to understand not just outputs, but causality.
This forces a profound shift in modelling architecture. Fully opaque vendor engines fade. Fully bespoke internal builds struggle to scale. The winning approach is structured transparency: platforms that enforce discipline and consistency while keeping actuarial logic visible, explainable, and adaptable.
This is where the long arc bends toward environments like Mo.net, not because they were fashionable, but because they were built around ownership of thinking rather than ownership of machinery.
Scenario Explosion and the End of Point Estimates
In 2035, point estimates are treated with suspicion.
Decision-makers expect distributions, response surfaces, and stress maps as a matter of course. Thousands and sometimes millions of scenarios are explored automatically, not because they look impressive, but because the marginal cost of exploration has collapsed.
Uncertainty stops being something to explain away and becomes something to work with.
What Has to Break to Get Here
This future does not arrive gently. Several deeply ingrained assumptions have to fracture along the way.
First, the idea that modelling speed is a “nice to have” must die. As long as slow models are tolerated, organisations will continue to schedule insight rather than demand it. Real-time modelling forces uncomfortable questions about why decisions were ever delayed in the first place.
Second, the myth that control requires opacity has to collapse. For decades, many firms equated locked-down vendor platforms with safety. In a continuous-modelling world, opacity becomes risk. Control comes from understanding, not concealment.
Third, the boundary between “production” and “strategic” modelling must erode. In 2035, strategies are not tested once and then operationalised, they are continuously shaped by feedback from live models. This breaks traditional governance structures that assume exploration is episodic and contained.
Fourth, the role of the actuary has to be redefined. Actuaries who define their value by model construction rather than interpretation will struggle. Those who lean into judgement, explanation, and strategic framing will thrive.
Finally, organisations must accept that their bottleneck is no longer computation, but decision-making itself. When models move faster than committees, the friction shifts upward. This is not a technology problem. It is a cultural one.
Looking Back from 2035
From the vantage point of 2035, the debates of the 2020s feel oddly constrained. Arguments about vendor platforms versus open source seem incomplete. Questions about whether AI will replace actuaries sound naïve. The real shift was never about tools but about tempo.
Once modelling stopped being slow, everything else had to change. And the firms that adapted earliest were not the ones with the biggest budgets or the most complex systems.
They were the ones willing to let a few comfortable assumptions break, before reality did it for them.
Spot the Difference
| Now | 2035 |
| Models are run | Models are there |
| Insight is scheduled | Insight is immediate |
| Monthly close drives behaviour | Decisions drive behaviour |
| Actuaries wait for results | Results respond to actuaries |
| Scenarios are chosen carefully | Scenarios are cheap and abundant |
| Single answers are defended | Ranges are explored |
| Speed is negotiated | Speed is assumed |
| Black boxes feel “safe” | Black boxes feel reckless |
| AI is optional | AI is unavoidable |
| Models slow organisations down | Models expose where organisations slow themselves down |