Lean AI Consultancy Boxtree Launches Into a Market Built on Broken Promises

The licence problem nobody wants to say out loud

There is a running joke in UK boardrooms in 2026: every chief executive is "implementing AI", and almost none of them can tell you the return. Max Pardo-Roques, the founder of newly launched Birmingham consultancy Boxtree, has stopped laughing at it.

"Look closer and they've bought ten or fifteen Copilot licences, a few people are writing better emails and the marketing copy has sharpened up," Pardo-Roques told Consultancy.uk this week. "That's individual productivity, not transformation. Nobody has touched the processes where the money actually sits."

Boxtree Consulting, based in Birmingham's Jewellery Quarter, applies Lean methodology (the improvement philosophy derived from the Toyota Production System) to identify the manual processes AI can now automate, from invoice checking and data entry to scheduling and compliance. The firm builds solutions inside the systems clients already run, rather than bolting on a new platform. The model comes with a guarantee: every engagement is cost-neutral within its first year, targeting a five-to-one return on investment.

Pardo-Roques brings more than 15 years in operational excellence and transformation consultancy, most recently as commercial director at Develop Consulting. His client base spanned manufacturing, FMCG, airports, logistics and professional services: exactly the sectors where administrative waste compounds fastest and where AI pilots go to die.

Why most AI projects fail before they start

Boxtree's launch is uncomfortable reading for a consulting sector that has sold AI strategy hard for three years. The numbers justify the discomfort. Nearly 85% of AI projects fail to escape the pilot stage, and the share reaching production has actually declined, from 32% in 2024 to an estimated 25% in 2026, according to analysis by AI consultancy Helium42. The Management Consultancies Association found that 77% of UK consulting firms have integrated AI into their own systems, yet fewer than a third of their clients report positive ROI.

Boxtree's diagnostic method addresses this directly. A Lean process map first puts a pound value on the waste inside a client's operation. A business case is then built per opportunity, and projects proceed only where first-year returns provably exceed the cost of the work. The sequencing matters. Most of the sector, as Pardo-Roques puts it, "starts with the technology and goes hunting for a use case."

That critique lands harder when you look at market structure data. AI Business Dispatch analysis of Companies House and IPO data (as of July 2026) shows 2,312 new SIC 70.22 management consultancy companies registered in 2026-Q3 alone, a figure down 81.9% versus the prior period as the post-pandemic formation boom corrects sharply. Among all active SIC 70.22 companies, 99.7% hold no Class 45 trademark, meaning the overwhelming majority of the cohort has no registered brand protection in legal or advisory services. UK Class 45 trademark filings sit at just 420 in Q3 2026, down 73.5% on the prior period. The consultancy market is contracting and consolidating, yet almost nobody in it is investing in the brand infrastructure that signals long-term intent.

Boxtree is a deliberate counterpoint to that pattern: unglamorous, process-first, guarantee-backed.

Demand is real; the skills to meet it are not

The timing of Boxtree's launch coincides with a structural problem on the supply side. Demand for AI agent expertise increased by 5,800% over the past year across Malt's global freelancer network, while the number of AI engineers on the platform grew 229%. The skill set is real and growing, but nowhere near keeping pace with demand. Consultancy.uk reported this week that AI agent skills among the UK workforce are growing at roughly half the rate of the demand for them.

For small and mid-size businesses, that gap is the practical reality. They cannot compete on salary with the large enterprises absorbing specialist AI talent. They cannot afford Big Four retainers. So they buy SaaS licences, call it a strategy, and watch utilisation hover at the individual productivity level Pardo-Roques describes.

The Malt data also points to why boutique operators like Boxtree exist: major UK organisations are increasingly turning to independent AI specialists and fractional executives to bridge the skills gap, specifically because the permanent hire market is broken for this skill set.

The compliance backdrop tightens

One process category Boxtree will find itself working through, particularly for professional services clients, is recruitment and HR automation. The ICO's March 2026 report, drawn from evidence across more than 30 UK employers, found that most are non-compliant with automated decision-making rules when using AI to screen and rank candidates. The Data (Use and Access) Act 2025, effective from 5 February 2026, replaced the near-total prohibition on significant automated decisions with a permissive but safeguard-led regime. Final ICO guidance is expected this summer.

The compliance burden is not going away. It is exactly the kind of process complexity (legal basis, human review requirements, candidate rights) that a Lean diagnostic is well suited to map before an AI tool is switched on. That is an unadvertised use case for Boxtree's model, but a real one.

So far, the Jewellery Quarter address is the only glamorous thing about the firm. That may be the point.