Hi Reader

Read time: 4 mins

This week’s stories made me smile for the same reason they made me wince – they all show how easily good intentions trip over themselves. AI meant to save time creates fake quotes. Earn-outs designed to align incentives end in arguments. Micromanagement masquerades as care. Business, it seems, is just people doing their best not to make the same mistake twice.

Enjoy!

When the audit needs an audit

Not checking the accuracy of generative AI outputs can be costly – just ask Deloitte Australia, which is refunding part of an AU $440,000 fee after admitting it used AI to produce a report containing non-existent references and a fabricated court quote. The report examined an IT system that automates welfare penalties – a politically charged area given the fallout from the country’s Robodebt scandal, in which thousands of welfare recipients were wrongly pursued for debts generated by faulty algorithms – where there is little margin for error. For Deloitte, it’s an expensive embarrassment. The firm, like many others, has championed the potential of generative AI while stressing the importance of human oversight. Get the story here. It’s paywalled but you can access it with Archive.

The selling price depends on tomorrow: earn-outs demystified

The hardest part of selling a business isn’t always letting go – it’s staying long enough to hit the numbers that decide your final payout. Earn-outs link part of the sale price to future performance, bridging the gap between what a buyer is willing to pay now and what a founder believes the company is worth. They sound fair in theory but can become fraught in practice – especially when definitions of “performance” blur or incentives diverge. In a jittery market where valuations are hard to fix, earn-outs are increasingly being used to keep deals alive and test how much founders truly believe in their own growth story. Find out more here.

The transparency trap: why admitting AI use can shrink trust

A new study finds that telling people you’ve used AI makes them trust you less – even when the work is just as good. Across 13 experiments, researchers found the same pattern in settings from hiring to investing to creative work: disclosure triggers doubt about legitimacy and effort. The effect holds whether the disclosure is voluntary or mandatory, and even if you stress that the output was checked by a human. Being exposed later is worse, but honesty still carries a cost. It looks to me like the challenge is to make AI use visible without making it seem like a shortcut. Read the report here.

AI prompt of the week: inclusive hiring

Most founders genuinely want diverse teams but unconsciously rely on biased processes – narrow job descriptions, homogenous networks and subjective interviews that favour familiar backgrounds. Research shows diverse teams outperform on innovation and decision-making but achieving this requires intentionally designing bias out of your hiring system, not just good intentions.

Help me audit and redesign my hiring process to attract and fairly evaluate diverse candidates. My current process: [describe stages]. Identify where bias likely enters – job descriptions, sourcing channels, interview questions, evaluation criteria. Design specific interventions for each stage: inclusive language, expanded sourcing strategies, structured interviews and objective scoring systems that assess actual job performance predictors while reducing bias.

The cautious CMO: brand building in the age of risk

Marketing leaders are moving fast – but watching their step. Censuswide’s Voice of the UK CMO 2025 report shows that nearly nine in ten CMOs now use generative AI and most say it has improved creativity and performance. Yet consumer sentiment lags: people are twice as likely to feel uneasy about AI as excited, and fewer than four in ten are comfortable with brands using it for social content. The same split appears around brand activism, where half of consumers expect brands to take a stand and half do not. Against this backdrop, CMOs are doubling down on brand awareness and research, using data not just to prove marketing’s worth but to navigate a more cautious, divided market. Read the report here.

A repository of excellent advertising

The Big Ad is an archive celebrating standout creative work from around the world – from billboards to lo-fi viral videos. What’s striking isn’t the production value but the clarity: simple ideas, executed with wit and precision, that cut through the noise of overproduced content. In an age of algorithmic sameness, this inspirational site is a reminder that originality still starts with a timeless human insight, not a passing trend. Discover some of the world’s best ads here.

How AI agents are rewiring B2B sales

BCG’s new analysis of “AI agents” in B2B sales shows how far automation has crept into what used to be pure human territory. These aren’t chatbots – they’re autonomous systems that can qualify leads, handle pricing and even close smaller deals. For business leaders, the point isn’t to replace people, but to rethink what people are best at. The future sales team might look like a partnership between humans who build trust and AI agents who never sleep, forget or fudge a follow-up. The challenge now is designing AI into the sales process without stripping out the human signals that build trust. Check out the report here.

Build for humans – and the machines reading over their shoulders

MarTech’s Marc Sirkin argues that the web we’ve optimised for two decades has quietly died. AI systems now consume and reinterpret content at scale and most brand sites are unreadable to them – too visual, too wordy, too unstructured. The risk isn’t invisibility but misrepresentation: if an LLM can’t parse your data, it fills the gaps with guesses or competitor information. The fix is what Sirkin calls the “dual web”: one version rich and human, another stripped to pure signal – structured data, schema markup and plain-text clarity. In the age of machine readers, whoever builds the cleaner data layer controls the narrative. Get the story here.

The real cost of micromanagement

Tightly controlling management looks like care from the surface, but underneath it signals mistrust, creates bottlenecks and drains initiative. The higher your standards, the more you need to step back – not to lower the bar, but to let people reach it their own way.

​

Drop me a line

Thanks for reading and feel free to get in touch. I’m always curious to know what’s resonating with you – or if there’s a trend or story I should know about. Hit reply or message me if something here sparked a thought.

Cheers!
Adam

Newsletter Exclusive Discount!

Use code 'GM20' to get 20% off your first fix with JustFix

​
​Unsubscribe · LinkedIn