Angela Rayner has just resigned. The Deputy Prime Minister's defence about her stamp duty affairs—"I followed professional advice"—finally collapsed this morning when the ethics committee concluded she'd breached the ministerial code. Her position had become untenable after admitting she underpaid stamp duty on an £800,000 flat in Hove, which she claimed was based on legal advice.

In her resignation letter, Rayner said she "deeply regrets" not seeking additional specialist tax advice and takes "full responsibility for this error". But here's what's telling: for weeks, she'd insisted the fault lay with her advisers. The conveyancing firm who handled the purchase stated they never gave her tax advice and always direct clients to tax experts for such matters. They were, in their words, being made "scapegoats."

It's a familiar refrain. Last year, the CEO of a well-known UK retailer spent £2 million on a rebrand after their brand consultancy convinced them they needed to "premiumise" to compete with online players. New logo, new stores, new positioning—the works. Customers hated it. Sales dropped 30% in six months. When hauled before shareholders, the CEO's explanation was predictable: "We followed expert agency advice based on extensive market research."

The agency still got paid. The CEO got sacked. Funny how that works.

It reminded me of a meeting last week where a company director blamed her adviser for a botched repair decision that's now costing residents dearly. Different scale, same dodge.

Here's the uncomfortable truth: advice is input, not absolution. When you sign the form, reject the repair, or approve the budget, the decision—and its consequences—belong to you.

The accountability dodge

We've built an entire culture around outsourcing responsibility. Politicians blame advisers. Directors blame consultants. Managers blame AI recommendations. It's become so common that "I was acting on advice" has become the corporate equivalent of "the dog ate my homework."

This isn't new, but it's accelerating. As decision-making gets more complex and advisers multiply—human experts, AI models, data analytics—the temptation to hide behind their recommendations grows stronger.

The problem? Uncertainty doesn't disappear just because you've hired someone clever. It just gets repackaged.

Why advice goes wrong (and it's still your fault)

Every adviser comes with limitations, whether they're charging £500 an hour or running on GPT-4:

Human advisers optimise for their incentives, not yours. The tax adviser minimises tax exposure but might miss reputational risk. The consultant who recommended against repairs? Perhaps they were thinking about this year's budget, not next year's liability. Rayner's conveyancers? They calculated stamp duty based on what she told them—they weren't qualified to advise on the trust implications.

AI advisers sound supremely confident while hallucinating facts. They'll quote non-existent regulations with the certainty of a High Court judge. They optimise for statistical likelihood, not your specific context.

Both types share a fatal flaw: they're not accountable for outcomes. You are.

The trust-but-verify framework

Here's the approach I've developed over thirty years of making decisions (and sometimes getting them wrong):

Start by clarifying what decision you're actually making. Not what advice you're seeking—what choice you're committing to. Then surface the assumptions. Every piece of advice rests on assumptions about your situation, your goals, your constraints. Make them explicit.

Get a second opinion, but make it adversarial. Don't ask another adviser to agree; ask them to poke holes. If you're using AI, run the same question through a different model and note the discrepancies. Rayner might still be Deputy PM if she'd got that second specialist opinion she now "deeply regrets" not seeking.

Document everything, but keep it simple. One paragraph: what you decided, what advice you received, what checks you performed, and why you chose what you did. This isn't about covering your backside—it's about thinking clearly.

Making AI work without losing your mind (or accountability)

AI has made getting advice almost frictionless, which makes it more dangerous, not less. Here's how I use it responsibly:

Never paste confidential data into public models. Obvious, but you'd be amazed how many don't follow this.

Always demand sources. I use this prompt: "Provide evidence for your recommendation with specific sources and links. Where uncertain, say 'I'm not certain about this.'"

Cross-check everything that matters. Different models, different prompts, different days. AI responses can vary wildly based on how you ask.

Most importantly, remember that AI doesn't know your context. It doesn't know your risk tolerance, your company politics, or that specific clause in your contract. That's your job.

The director's dilemma

Back to that director who rejected the repair works based on an adviser's recommendation. A responsible approach would have been straightforward: request the survey data and moisture readings. Get a second opinion from another qualified surveyor. Document the decision with clear rationale. Set a review date—perhaps after the next heavy rainfall.

Instead, she pointed to an email from an adviser and washed her hands of it. Now residents are dealing with water damage that could have been prevented.

The adviser might have been wrong. But the decision—and the accountability—was hers.

When to raise the bar

Some decisions demand extra scrutiny. If it's high-impact, hard to reverse, or legally sensitive—like, say, paying property tax when you're the Housing Minister in the middle of a housing crisis—then you need to slow down. Get multiple opinions. Make verification mandatory.

For lower-stakes decisions that are easily reversed, move faster with small experiments. But never forget: the accountability gradient doesn't change with the stakes. You own every decision, big or small.

The leadership standard

Leaders aren't judged on whether they have advisers, but on how they use them. Human or AI, advice should sharpen your thinking—not replace it.

Rayner learned this the hard way. In her resignation letter to Starmer, she accepted she "did not meet the highest standards" despite the ethics adviser concluding she acted "in good faith and with honesty throughout." Good faith isn't enough when you're making decisions that affect others.

"I just followed advice" may explain your process. It doesn't excuse your outcome. The decision was yours. The consequences are yours. Own them.

Because if you won't take responsibility for your decisions, why should anyone trust you to make them?