Expertise

AI strategy and adoption

The value of AI does not come from the tool. It comes from redesigned processes, sound governance and real ownership by the teams. So every engagement opens with a close look at how you operate: where can generative AI create value for you, in what order, and with what safeguards in place?

• AI maturity assessment. A clear read on current usage, data and skills, and on the use cases with the most potential.
• A roadmap prioritised by return on investment. A sequenced rollout plan driven by expected return and the effort required, not by what happens to be fashionable.
• A governance and usage framework for generative AI. Usage rules, data confidentiality, and human sign-off on outputs.
• Executive alignment. A shared ambition and a common vocabulary across the leadership team.
• Guidance on APIs and tools. A considered choice of models (Claude, GPT), platforms (Cloudflare) and the programming interfaces your use cases call for.
• Prototyping. Working prototypes, so decisions rest on evidence rather than promises.

Corporate AI training

AI is not learned in theory; it is put into practice. These programmes teach teams to use generative AI in their day-to-day work and to build agents that automate their own tasks. Every session is built on the participants' real cases, the agents are assembled in the session itself, and everyone leaves with working tools. By the end, every participant can work with an LLM and generative AI, and two in three can also deploy agentic solutions such as Cowork, by Claude.

• Role-based programmes. Executives and business teams: customer service, sales and marketing, human resources, legal, IT, operations, finance.
• Participants' real cases. Exercises draw on your own documents, processes and pain points, not on generic examples.
• Agents built in the session. Hands-on demonstration, with solutions put in place on the spot.
• Advanced module: server-side automation. For processes that have to run unattended: integration and deployment pipelines, model access, audio generation.

What the research measures, function by function

The figures below come from public studies, run in different settings and with methodologies that do not always line up. They show what is possible; they are not a promise of results.

Customer service

AI-assisted agents resolve, on average, 14 to 15 per cent more cases per hour; for the least experienced agents the gain reaches 34 per cent. Brynjolfsson, Li and Raymond, study of nearly 5,200 agents, NBER 2023, Quarterly Journal of Economics 2025.

Sales and marketing

At function level, McKinsey credits AI with a revenue uplift above 10 per cent, driven by content drafting, idea generation and campaign personalisation. McKinsey, The State of AI, November 2025.

Software development and IT

A coding task gets done roughly half again as fast; at function level, costs fall by 10 to 20 per cent. Peng et al., GitHub and MIT, 2023; McKinsey, November 2025.

Legal and compliance

In a randomised controlled trial, lawyers completed complex tasks 50 to 130 per cent faster, with clearer and deeper analysis. Schwarcz et al., 2025.

Consulting, analysis and research

Consultants equipped with AI handled 12 per cent more tasks, around 25 per cent faster, with quality higher by about 40 per cent within the scope AI covers. In document synthesis, working time falls by around 30 per cent, and by up to 61 per cent for free-form summaries. Dell'Acqua et al., Harvard and BCG, 2023; Microsoft Security Copilot controlled trial, 2024.

Writing and communication

Writing time falls by about 40 per cent, while the quality of the output rises by about 18 per cent. Noy and Zhang, Science, 2023.

Human resources and recruitment

AI speeds up CV screening and large-scale shortlisting, though the measured gains are less firmly established than in the functions above. McKinsey; sector surveys, 2025.

Finance, accounting and operations

In some administrative activities, 60 to 70 per cent of working time is technically automatable. McKinsey, 2023.

At organisation level, 88 per cent of companies now use AI in at least one function, yet 94 per cent admit they are not getting significant value from it. The gain comes from redesigned processes and governance, not from the tool on its own. McKinsey, The State of AI, November 2025. At employee level, daily users of generative AI report productivity benefits far more often than occasional users: 92 per cent against 58 per cent. PwC, global survey 2025, close to 50,000 workers across 48 economies.

A candid caveat. Outside the territory AI has mastered, quality can slip: roughly 19 per cent fewer correct answers (Dell'Acqua et al., 2023). And when seasoned developers used AI on complex codebases they knew inside out, their tasks took 19 per cent longer, even though they felt they had worked faster (METR, 2025). The gain is not automatic; it has to be governed. That is exactly what the consulting and the training are for.

Trade finance, working capital & fintechs

Where it all began, and a full advisory practice in its own right. Two decades in the field at major international banks, across three kinds of engagement.

• Structuring trade finance transactions. Supply chain finance, securitisation, discounting, export finance, receivables finance, letters of credit and guarantees.
• Working capital optimisation. A diagnosis of the operating cycle, and the financing arrangements to match.
• Advisory for foreign fintechs. Market-entry strategy, Europe in particular, and go-to-market business plans.

It is this grounding that makes the AI offering uniquely relevant to finance and fintech.