InMarketing Today

Barclays reported second-quarter pre-tax profits of £3.3bn, up almost a third, on revenues of £8.3bn — more than £1bn ahead of last year and ahead of analysts' estimates. Investment banking revenues rose a fifth, with equities trading income up 45 per cent and advisory and underwriting fees up 32 per cent. Chief executive CS Venkatakrishnan upgraded the 2026 group revenue target by around £500mn to £31.5bn. The shares fell 4.4 per cent in early trading.

Barclays has spent years telling investors it is pivoting away from the investment bank towards UK retail and corporate banking. This quarter did the opposite: the investment bank carried it, the three UK businesses missed on net interest income as deposit competition bit, and the private bank and wealth division saw return on tangible equity fall to 26.9 per cent from 31.9 per cent a year earlier. Jefferies analyst Jonathan Pierce called it "a slightly messy set of numbers". When results contradict the positioning, the market prices the contradiction rather than the profit.

Set your last three quarters of results against the story you tell about your business. If the division you are quietly de-emphasising keeps rescuing the numbers, either change the story or change the business — carrying both is what produces a sell-off on good news. And treat Venkatakrishnan's "watch this space" on the wealth unit as the promissory note it is: audiences remember the ones that go unredeemed.

X Money is now live for US Premium and Premium Plus subscribers. It offers peer-to-peer payments and transfers, a deposit account and a physical Visa debit card for point-of-sale and ATM use, with X advertising up to six per cent APY, three per cent cashback and pay arriving up to two days early. Elon Musk set out the ambition at an all-hands call in October 2023: "When I say payments, I actually mean someone's entire financial life... you won't need a bank account." He said then it would blow his mind if the product were not shipped within roughly a year. It has arrived around two years late.

On paper the offer is competitive; six per cent buys attention. But a deposit account is a trust product, not an attention product, and the two are bought with different currencies. X has extraordinary distribution and an unusually polarised brand — and distribution is not permission. That the headline rate is the loudest element of the launch tells you which of the two X is relying on.

If you are launching a money product off the back of an existing audience, test whether that audience will trust you with a salary rather than merely a follow — the two behaviours are not correlated, and the research to separate them is cheap. Rate-led acquisition in deposits also selects for the customers most likely to leave when the rate moves. Report retention alongside sign-ups from day one, or you will be celebrating a number that reverses.

Margaret Lee set out how she built an AI-first marketing team, and the useful part is the discipline rather than the tooling. Marketers spend nearly twice as much time with AI tools as other functions, on TMetric's data, while McKinsey's State of AI research finds adoption rising without the work itself changing. Her rule is that every training course is paired with a real project starting the same week. Her scorecard targets 70 per cent weekly active usage in the pilot team, a 60 per cent cut in time spent on the top three routine tasks, campaign reporting turnaround halved, and content volume up fivefold without added headcount.

One line carries the piece: if a process is not clear, AI will not fix it — it will amplify what is already wrong. That turns the question from "where can we use AI?" into "which bottlenecks does AI solve?" In a regulated firm the point lands twice. An unclear review process, automated, is an unclear review process running at speed, with the compliance exposure scaled to match. Every AI-assisted output still leaves the building with your name on it.

Map your recurring processes before you buy another seat. Take the three that consume the most hours, write down the current state, the target state and the name of the person who owns the change. Then measure the output — reporting turnaround, error rates, review cycles — and ignore prompts run and training hours completed. Vanity metrics will let a team feel productive for a full quarter while nothing actually changes.