TL;DR

Strategy&, the consulting arm PwC has owned since 2013, is acquiring Datasparq. The target is small — roughly 20 people in London — but its client list runs to GSK, GXO and easyJet, and the buyer wants the engineering capability pointed at deal work.

A small team, a specific purpose

Datasparq works across AI strategy, data platform building and applied delivery, for clients in supply chain, professional services and private equity. Its people will join Strategy& on completion, which puts data engineers and scientists inside a 4,500-strong consultancy spread over 75-plus offices.

The intended use is narrow and revealing. PwC wants the capability at every stage of a transaction, from diligence beforehand through to transformation afterwards. The firm says it serves over 90% of the world’s biggest PE investors and advises on the majority of private equity transactions globally, so the acquisition is less about entering a market than about defending a position inside one.

Nicola Preedy, who heads deals at PwC in the UK, put the commercial logic in terms of speed: in private equity, execution pace and the capacity to realise an investment remain the differentiators. Founder John Wyllie said the attraction was the surrounding breadth — sector depth and functional expertise his firm could not have assembled alone.

The pattern

This is the third Big Four AI move Resultsense has covered inside a fortnight. KPMG and EY took a £456m contract to train the civil service; Bain announced an Anthropic partnership days after its own Whitehall deal. The consistent shape is firms buying or partnering for capability rather than building it, and pointing it at clients who will pay for advisory hours.

Looking forward

For UK AI consultancies the read-through is straightforward: a 20-person specialist with blue-chip logos is now an acquisition target for a Big Four firm, which resets what independence is worth against a trade sale.

For buyers of these services, the question is whether capability acquired this way survives integration. Twenty engineers inside a 365,000-person network is a rounding error, and the value depends entirely on whether they keep doing the work that made them worth buying.