AI-first company

Strategy · Adopt

Budget for people and process, not just tools

Licences are the small part of the cost. Plan most of the time and money for training, redesigning work, and helping people through the change.

By
Last reviewed
Corrections
Report an error
When Do first. Before you start with AI, or as you start.
Effort Low (days, one person)
Evidence Emerging

Three surveys agree that people and organization, more than technology, hold AI back. All rely on what respondents say. The 10-20-70 split is a consultancy’s rule of thumb, not a tested result.

Fits
Industries All industries
In the guide Mid-market companies

What it is

When you plan an AI project, cost the whole change: training, time for staff to redesign their work, communication, and the slower weeks while people learn. Do not stop at the price of the software.

A well-known rule of thumb comes from the consultancy BCG. It says 10% of the effort in an AI programme goes into algorithms, 20% into technology and data, and 70% into people and process change. Treat the exact numbers as an opinion. The direction has support.

What the evidence shows

  • Skills are the strongest barrier. In a survey of 108 specialists in the Netherlands, a shortage of skilled staff held AI progress back more than anything else. IT systems and data problems had no direct effect beyond chance.
  • Training and consultation go with better outcomes. OECD surveys in seven countries found that workers did better where employers trained them and consulted them about AI.
  • Barriers are organizational. In a survey of 179 leaders in Poland, barriers were mostly organizational, not technical. Managing the change was a problem for trained and untrained leaders alike.

How to do it

  1. Before approving a tool, list the non-tool costs: training hours, redesign workshops, time for checking output, support.
  2. Put a figure on each, in hours and money.
  3. Plan for a dip in output while people learn.
  4. Fund training for the people who use the tool every day, not only for specialists.
  5. Give managers time to redesign their team’s work.
  6. Review the split after the first project, and adjust the next budget.

What to measure

  • Share of the AI budget spent on tools, and on everything else.
  • Training hours per user.
  • Take-up: the share of intended users who use the tool weekly after three months.

Where it fails

  • Using 70% as a target. The right split depends on the project. A simple built-in feature needs little change work. A redesigned process needs a lot.
  • Training with no use for it. Training months before people get the tool is wasted.
  • Ignoring the technology. Data quality and integration with existing systems are real barriers too, most of all once you roll out beyond a pilot.

By company size

  • Medium: you buy and set up tools; you do not build them. So nearly all your own effort is people and process.
  • Mid-market: expect data quality and integration to take a larger share once you roll pilots out more widely.
  • Enterprise: change work spans many units. Budget for it centrally, or each unit will underfund it.

Company cases

Companies that did this

Sources

  1. van Veen, Zhu and Ravesteyn (2026). How AI Implementation Barriers Shape AI Maturity? The Interacting Roles of Skilled Employees, Technological Infrastructure and Data Quality. CONF-IRM 2026 Proceedings, 17. What this study did and found
  2. Lane, Williams and Broecke (2023). The impact of AI on the workplace: Main findings from the OECD AI surveys of employers and workers. OECD Social, Employment and Migration Working Papers, No. 288. What this study did and found
  3. Wojewodzic (2026). AI Transformation in Polish Organisations: The Gap Between Individual Adoption and Organisational Maturity. e-mentor, 2(114), 75–83. What this study did and found
  4. BCG (2022). 5 Rules for Fixing AI and Machine Learning for Your Business (the 10-20-70 rule). bcg.com

Claims on this topic, checked against these studies: