TL;DR
JMAN Group combines genuine private equity specialisation with the scale to deliver data programmes across a large portfolio. Squirrel AI is smaller by design: partner-led, build-first and suited to judgement-heavy AI work where the team must choose priorities with management, ship quickly and own a P&L result.
What JMAN Group is
JMAN Group was founded in 2010 and has offices in London, New York and Chennai. It has a team of more than 600 people, works with more than 150 private equity funds and 400 portfolio companies, and is backed by Baird Capital.
The firm positions itself around translating data into value for PE funds and their portfolio companies. Its model combines a management consultancy commercial mindset with technology delivery. Services span core reporting, data and AI advisory, value creation, diligence and exit preparation.
JMAN also has a substantial delivery operation in Chennai and a well-known graduate recruitment engine. This gives the firm the capacity to serve many accounts and run sizeable programmes at once.
Where JMAN is genuinely strong
Real PE specialisation is rare and worth respecting. JMAN understands that a fund, its operating team and a portfolio company can have different information needs. Its experience across hundreds of portcos also gives it a useful view of recurring data problems around ownership, reporting and exit readiness.
Scale is a genuine asset when the brief is repeatable. A fund standardising reporting across 20 portfolio companies needs delivery capacity, consistent methods and enough data-engineering muscle to handle several implementations in parallel. JMAN has been built for that sort of programme.
Its price point also sits below MBB, while still offering an established firm, broad service coverage and senior engagement. For a buyer that wants one provider across diligence, data foundations and portfolio reporting, that breadth can simplify procurement and governance.
Where the model gaps for some PE portfolios
The trade-off is structural. Scale in consulting is built through a staffing pyramid: larger delivery teams weighted towards early-career analysts and offshore engineers, with senior oversight spread across accounts. That model works well for clearly specified, repeatable data work.
It becomes harder when the work depends on concentrated judgement. Choosing which three builds matter most to this portfolio company's exit story is not a standard engineering task. Neither is redesigning a workflow with a sceptical management team, or making an agentic system stick in a business that has struggled to adopt new software.
As implementation shifts from traditional data engineering towards agentic AI, senior people need to spend more time inside the workflow. Models behave less like fixed software rules. The delivery team must understand exceptions, commercial risk, staff behaviour and where human approval belongs. Buyers should therefore ask who will actually sit with management and who has authority to change the scope when the evidence changes.
We compete against and follow larger data and AI firms regularly. Portco executives often describe the same pattern after an advisory phase: an intelligent, well-formatted review has reached the executive team, but the practical questions about data readiness, dependencies, priorities and ownership remain open. That observation is about how multi-stage consultancy programmes work, not a claim about a particular JMAN engagement.
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.
We are small by design. Our founders spent almost a decade in investment roles at firms including Apax, 3i and Deutsche Bank before moving to the operating side. We were investors ourselves before this, so we cover the commercial side as well as the technical.
The people who scope an engagement also build it. A trusted offshore engineering bench adds capacity where needed, but never owns commercial judgement or client leadership. That keeps the people making prioritisation decisions close to the system, the users and the P&L baseline.
The first scoped engagement usually takes three to four weeks and produces something live. For a Big-6 accountancy backed by a major PE fund, we built a payroll automation in days that saves 20 to 30 days of team time each year. We also built a credentials database for its go-to-market teams. These are contained systems, but each attaches to a clear operating constraint and a measurable result.
We charge materially less than larger firms because there is no pyramid to fund. The work is not lighter. The company structure is. More of each pound pays for the people building, and the senior team is the delivery layer rather than a review layer.
Who should choose JMAN, and who should choose us
Choose JMAN if a fund needs to standardise data and reporting across a large portfolio, or if a portfolio company has a well-specified data-engineering programme that requires substantial capacity. Its PE experience, established methods and scale are meaningful advantages for that brief.
Choose Squirrel AI when the work begins with a harder commercial question. We fit funds and portcos that want senior, commercially fluent builders embedded with management, choosing initiatives by P&L impact and the time left in the hold period. We are also a better fit when the first result needs to be live in weeks and management does not want separate strategy and implementation teams.
Three questions to ask either of us
- Who exactly will do the work, and which people will sit with our management team when priorities or workflow assumptions change?
- What will be running in production by week six, and what must our team or data estate provide before that can happen?
- What number will the build put on the P&L, how will we establish the baseline, and who will measure the result?
If your priority is judgement-heavy AI implementation rather than a portfolio-wide reporting rollout, book a 30 minute call. We can take one candidate workflow and test whether its likely P&L effect warrants a three to four week build.