TL;DR
McKinsey, BCG and Bain are rational choices when a board needs strategic certainty, broad benchmarking and an adviser whose name carries weight. Squirrel AI is built for a different point in the process: a PE fund or portfolio company that needs senior people to choose the right use cases, build them in production and measure the result inside the hold period.
What McKinsey, BCG and Bain are
The three firms commonly grouped as MBB have substantial technology and AI capabilities. McKinsey's AI and data arm is QuantumBlack. BCG combines strategy, technology and product building through BCG X. Bain has a large advanced analytics practice and a well-publicised alliance with OpenAI.
These are global advisory firms with access to senior executives, cross-industry research and large pools of specialist and generalist talent. Their AI engagements are premium and typically begin with a strategy phase: assess the organisation, benchmark its maturity, identify opportunities and set a direction.
That shape makes sense for many buyers. A multinational deciding where AI changes its business model has a different question from a portfolio company trying to remove a manual claims process before the next board meeting.
Where MBB is genuinely strong
MBB brings board and investment committee credibility that few firms can match. Its teams can draw on pattern libraries across sectors, mobilise quickly in several countries and give senior leaders air cover for decisions with material organisational consequences.
The firms are also good at turning an ambiguous question into a structured programme. If a board needs a defensible answer to "what is our AI strategy?", MBB is a rational buy. The brand can help align stakeholders who would otherwise disagree about the starting point, the level of investment or the risks.
That value should not be dismissed merely because the output is often a strategy. A sound direction can prevent a large enterprise from funding disconnected pilots or buying the wrong platform.
Where the model gaps for PE portfolios
The structural issue is that a strategy engagement usually ends where the P&L work begins. The agreed deliverable is a direction, produced by a model priced and staffed for analysis. Shipping production systems inside a mid-market portfolio company is another kind of work.
We regularly meet businesses a quarter after an MBB-grade AI review has landed. The recommendations are rarely wrong. They are simply still a document. Executives describe a familiar output: an intelligent, well-formatted review, presented to the executive team, without enough detail on data readiness, dependencies, sequencing or who will build the systems.
That is not an accusation about any named engagement. It is a consequence of the traditional consultancy model. Much of the day-to-day work is completed by bright generalists early in their careers, with partners reviewing the output. A portfolio company eighteen months from exit often needs the reverse: a few senior people close to the workflow, making commercial and technical decisions together.
The strategy-to-build handoff also weakens accountability. The adviser owns the recommendation, an implementation team owns the system, and management owns the gap between them. We explain that mechanism in Strategy Decks Don't Move EBITDA.
How Squirrel AI is different
An AI value creation partner to PE funds and their portfolio companies. We go into the portcos and build, rather than advising from the sidelines, and everything we build is aimed at the P&L.
Our founders spent almost a decade in investment roles at firms including Apax, 3i and Deutsche Bank before crossing to the operating side. We were investors ourselves before this, so we cover the commercial side as well as the technical.
The delivery model is partner-led. The people who scope the work are the people who build it. We use a trusted offshore engineering bench when extra capacity helps, but client leadership, prioritisation and the key build decisions stay with the senior team.
Our first scoped engagement is typically three to four weeks and ends with a production deliverable, not a report. Every build starts with a number: revenue added, cost removed, time returned or risk reduced. At Healf, the FT's #1 fastest-growing company in Europe, our founding team led AI projects that added more than £10m of annualised revenue to a business running at £100m. The fastest paid back in twelve weeks.
We charge materially less for the same outcome because the model is lighter, not the work. There is no pyramid of junior consultants to absorb the fee and no global brand overhead to fund. More of each pound goes to the people doing the building. Partner-led delivery is the quality argument and the cost argument at once.
Who should choose MBB, and who should choose us
Choose MBB for a large enterprise making a bet-the-company strategic choice, a board that needs independent validation, or a regulated organisation that values global reach and brand cover. It is especially sensible when the central problem is agreement on the direction.
Choose Squirrel AI when the direction is clear enough and the missing piece is execution. We are best suited to PE funds and portfolio companies that know AI matters, have a hold-period clock and want working systems producing a measurable P&L result in weeks.
The choice is not always either-or. A company can use MBB to settle a major strategic question and a focused build partner to convert selected priorities into production. What matters is making the handoff explicit before the strategy begins.
Three questions to ask either of us
- Who exactly will do the work each day, and how much time will the senior people who sold the engagement spend with the management team?
- What will be running in production by week six, using our real data and systems rather than a demonstration environment?
- What number will the work put on the P&L, when will it be measured, and who remains accountable for it after launch?
If those answers point to a small senior team and a defined production outcome, book a 30 minute call. Bring one workflow and its current cost or revenue baseline. That is enough to decide whether there is a build worth doing.