Martin KogutBy Martin KogutNovember 20, 20268 min read

AI adoption in mid-sized businesses

What makes an AI pilot worth scaling? Survey respondents report revenue and time savings. Getting those results in your own business starts with choosing the right workflow and testing it with the people who do the work.

What businesses report about AI

In September 2026, Bitkom reported that 57% of the 603 German companies it surveyed used AI. The companies had 20 or more employees. That puts AI on the agenda for many teams, but leaves each one with the same decision about where to start.

Two vendor surveys point to possible benefits.

  • 91% of respondents using AI said it helped increase revenue, according to Salesforce’s release on its 2024 global survey of 3,350 business leaders at companies with up to 200 employees. The percentage applies to the AI-using subset, and reflects their own assessments.

  • In Thryv’s May 2025 US survey of 540 decision-makers at businesses with 1 to 100 employees, 58% of respondents using AI reported saving more than 20 hours a month. This was self-reported.

A practical place to start is helping people do a specific task better. For example, a tool can help a sales team prioritize leads, while account owners decide whom to pursue.

What holds businesses back

In a 2025 preprint, Serena Proietti and Roberto Magnani used a four-level maturity model to assess 36 Italian SMEs across several sectors. None of the companies in that sample reached the top level.

Their discussion points to three barriers that deserve attention.

  • Limited knowledge. Start by identifying one work task where AI could help. Ask the people who do it what a useful result would look like.

  • Cost and resources. Budget for integration, training and support as well as tools and model usage.

  • Digital readiness. Spreadsheet-based work may be a reasonable starting point if the input is consistent and the team can test the output. A fragmented process may need work first.

Before you pick a tool, ask who owns the process, what inputs you can use and how the team will judge the result.

How to help teams adopt AI

In McKinsey’s 2021 global survey of 1,034 people who had been involved in organizational transformations, fewer than one in three reported that their company’s transformation improved performance and kept those improvements over time. In July 2024, Gartner forecast that at least 30% of generative AI projects would be abandoned after proof of concept by the end of 2025. That was a forecast. It cited data quality, risk controls, costs and unclear value as reasons.

Give the people who use AI a role in the rollout.

The team needs time to try the new workflow, report errors and help define what human review looks like. Set aside time for that work alongside technical implementation.

Four ways to support the change

  • Visible leadership support. Ask a senior sponsor to use the workflow, review the early results and help resolve problems.

  • A point person in each team. Give someone time to collect feedback, answer questions and help colleagues try the new workflow.

  • Clear communication. Explain why you’re making the change, what it means for the team’s work and when a person needs to step in.

  • An early result. Pick a small case and aim to show what improved in the first month.

A five-step plan for your first AI project

Use the timing as an example. Adjust it to your data, team and integrations.

1. Review your workflows (weeks 1–2)

Look for repetitive, data-heavy or error-prone work. Record how long it takes today and what errors cost.

2. Choose one use case (week 3)

Compare potential use cases by impact (time savings or revenue potential) and feasibility (available data and implementation effort). Start where an improvement matters and you can test it.

3. Run a pilot (weeks 4–8)

Test one workflow with real users, a time limit, a person responsible and clear measures of success.

4. Measure and share the results (weeks 9–10)

Compare time, quality and rework against your baseline. Share what improved and what still needs attention.

5. Decide whether to expand (week 11 onward)

Apply what you learned to the next workflow. Build internal skills and bring in outside support when the team needs it.

Funding for eligible advisory work in Germany

Germany’s Federal Office for Economic Affairs and Export Control (BAFA) supports eligible individual business advisory work for qualifying SMEs. The grant can be up to €2,800 per engagement in eligible 80% regions, with eligible consulting costs capped at €3,500. In the 50% regions, the maximum is €1,750. AI strategy work must meet the program’s advisory criteria.

BAFA applies the EU SME definition, which covers staff, financial and ownership criteria. The business must meet the German location requirements, and the consultant must meet BAFA’s requirements.

Check the official criteria and application sequence before commissioning work. Our guide to AI funding for German SMEs covers other options for implementation and research.

Start with one decision you can test

The surveys show what some participants reported. Use a pilot to check what AI changes in your own work. Set a baseline, make time for people to learn and measure the result.

Start with one workflow and agree on the evidence you need to decide whether to expand.

If you’d like to identify a promising workflow, talk to hypescale. We help mid-sized teams choose a use case, test it and plan what comes next.


Sources

  • Bitkom (September 2026). Survey on AI use in 603 German companies with at least 20 employees. Survey release. Bitkom (German)

  • Serena Proietti and Roberto Magnani (2025). Assessing AI Adoption and Digitalization in SMEs: A Framework for Implementation. Preprint. arXiv

  • McKinsey & Company (2021). Losing from Day One: Why Even Successful Transformations Fall Short. McKinsey survey

  • Salesforce (2024 fieldwork). Small & Medium Business Trends, sixth edition. Global vendor survey of 3,350 leaders. Salesforce survey release

  • Gartner (July 2024). Forecast of generative AI projects abandoned after proof of concept by the end of 2025. Gartner forecast

  • Thryv (May 2025 fieldwork). US survey of 540 small-business decision-makers. Thryv survey release

  • BAFA. Official overview of German business advisory funding for SMEs. BAFA guide (German)