The Engagement Problem: Designing Automation That Keeps Humans in the Loop

In my previous post, The Automation Paradox, I explored how frictionless financial automation, while designed to make life easier, can sometimes leave users disengaged and unaware of their own money habits. This unintended consequence raises a larger question: How do we ensure that automation enhances, rather than replaces, human judgement? If The Automation Paradox highlighted the risk of passive financial behaviour, this article explores the next step—how we can design automation that keeps people actively engaged.

Every day, millions of people follow turn-by-turn directions from Google Maps. But what makes these apps so effective isn't just their ability to automate navigation—it’s their ability to keep us engaged. They provide real-time traffic updates, suggest alternative routes, and leave the final decision to the driver. In doing so, they embody a crucial principle of automation design: keeping humans meaningfully involved.

This principle is often missing in personal finance and other digital experiences. As financial automation becomes more sophisticated—tracking subscriptions, categorising spending, even optimising savings—we risk creating a generation of passive users. Instead of making better financial decisions, people may simply outsource their thinking, losing touch with their own money habits.

The challenge isn’t just about automation. It’s about engagement. And the best systems won’t just automate tasks—they’ll make people better at them.

The Automation Paradox: Convenience vs. Comprehension

Automation in banking is supposed to make life easier. Features like Monzo’s smart categorisation, Starling’s subscription tracking, and AI-driven savings nudges all promise to help users manage their money effortlessly. And they do—but there’s a catch.

When automation shields users from friction entirely, they stop engaging with their finances. They don’t notice gradual spending shifts, lose track of commitments, and struggle to make informed financial decisions when automation can’t cover for them. The result? A paradox: the very tools designed to help us may be making us worse at managing money.

This isn’t just a finance problem. Across industries, full automation often leads to skill degradation and passive consumption. The solution isn’t to remove automation—but to design it so that it keeps people actively engaged.

Designing Automation for Engagement

From Raw Data to Smart Insights

Fitness apps like Strava don’t just track steps—they analyse patterns, highlight trends, and suggest recovery times. They turn raw data into meaningful insights while keeping users in control.

Financial apps could do the same. Instead of simply listing spending categories, they could highlight behavioural patterns. For example:

• Grocery spending tends to be 20% higher on Wednesdays than Mondays.
• A user has multiple streaming subscriptions, some of which may be unused.

By surfacing actionable insights instead of just data, automation can help users stay engaged without overwhelming them.

Strategic Friction: The Right Kind of Speed Bump

Not all friction is bad. In software development, AI-powered code assistants suggest changes, but developers must review them before implementation. This maintains engagement while reducing tedious manual work.

Financial automation should do the same. Instead of defaulting to auto-renewing every subscription or transferring savings without oversight, systems should introduce friction at decision-critical moments. For example:

• Notifying a user before an energy provider increases rates.
• Prompting a user before they take on a third Buy Now, Pay Later purchase in a month.

The goal isn’t to slow users down arbitrarily but to ensure they remain involved in key financial decisions.

Transparent Automation: Show Your Work

One reason GitHub Copilot works well for developers is that it doesn’t just suggest code—it explains why it made a suggestion. This builds trust and maintains skill development.

Finance apps should adopt the same principle. Instead of simply shifting money into savings or adjusting spending limits behind the scenes, they should provide explanations. For example:

• "We moved £50 into savings because dining-out expenses were lower this month."
• "This transaction was categorised as ‘shopping’ based on past behaviour. Is this correct?"

Transparency keeps users informed and prevents the "black box" problem, where automation works for users but not with them.

Agency-Preserving Defaults: Automation with a Human Touch

Calendar apps that suggest meeting times but require confirmation show how automation can assist without overriding human judgement.

Finance apps should follow suit:

• Instead of automatically cancelling unused subscriptions, they should prompt the user first.
• Instead of sweeping money into investments, they should provide clear scenarios on how different choices impact financial goals.

By maintaining user agency, automation can empower rather than replace decision-making.

The Future of Automation: Augment, Don’t Replace

As AI and automation evolve, the key challenge will be keeping humans in the loop. The risk isn’t just skill erosion—it’s a growing disconnection from decision-making itself.

The best-designed systems will:

• Surface meaningful insights, not just raw data.
• Introduce friction where it matters, not where it doesn’t.
• Make automation transparent, explaining its decisions.
• Preserve human agency by offering choices, not just defaults.

Good automation doesn’t just make life easier. It makes people better at what they do. Just like Google Maps makes you a more informed driver, the next generation of financial tools should make you a better financial decision-maker—not just a passive observer.

In the end, the goal of automation isn’t to remove human judgement. It’s to make it sharper, faster, and more informed.