OpenAI Signs Multiyear Enterprise Deals With Accenture, BCG, Capgemini, and McKinsey

OpenAI announced multiyear partnerships with Accenture, Boston Consulting Group, Capgemini, and McKinsey & Co. to help enterprise customers define their AI strategy and move AI agents into real production workflows. The deals formalise a delivery channel that has been quietly running for eighteen months and reshape the buying motion at the top of the market.

Scale at the top tier

The numbers behind the announcement are striking. Accenture's $3 billion investment in its Data & AI practice aims to double the AI workforce to 80,000 people, and the firm has been clear that AI Refinery is now a branded platform offering rather than an internal tool. BCG reported that 25% of its $14.4 billion in 2025 revenue — roughly $3.6 billion — came directly from AI-related consulting work. EY completed a $1.4 billion AI investment, building EY.ai as an internal large-language-model platform and rolling out EYQ to thousands of audit and tax staff. Deloitte's PairD tool has gone from 25% to nearly 75% adoption among UK audit staff in twelve months.

The Big-4-vs-boutique split is sharpening

At the top of the market — Accenture, Deloitte, McKinsey/QuantumBlack, PwC, KPMG, EY — engagements typically start around $300K, scale into the multi-millions for full programs, and command day rates of $350-$700. The pitch is scale, change-management muscle, and the ability to staff cross-functional teams quickly across geographies.

Below them, a fast-growing layer of boutique and specialist firms is winning a different kind of work — focused pilots in the $50K-$150K range, production deployments from $250K to $1M+, with senior practitioners actually doing the work rather than handing off to a tiered delivery pyramid. The trade-off is the trade-off it has always been: less change-management depth, more delivery sharpness, faster cycle times, and a much shorter distance between the partner you met during sales and the engineer who ships your code.

What buyers should ask

• Who actually staffs the work? The team you meet during pitch is rarely the team that delivers. Ask to interview the engagement lead and at least one technical practitioner before signing.
• What's the engagement model? Bodies-on-seats versus productized pilot versus outcome-based pricing all have different incentives and very different exit dynamics.
• What's the IP capture path? Half the value walks out the door each evening unless the SOW explicitly covers documentation, runbooks, and knowledge transfer.
• What does success look like in writing? SOWs that produce slides instead of working systems are the most common failure mode in this category, and the easiest to avoid by demanding measurable acceptance criteria up front.

For practitioners on the consulting side, the OpenAI partnership announcements are a useful market signal: the model providers are now treating the consulting channel as core to enterprise distribution, which means partner enablement, productized offerings, and named relationships are about to matter more than they did a year ago.