Your Boss Asks About AI Strategy—Here’s What to Say
4 min read
Your boss walks into the meeting and says: “We need an AI strategy. Get on it.” Your only thought: Where do I even start? According to Bitkom, 41% of German companies already use AI-but most without a real strategy. The good news: An AI strategy doesn’t have to be a 50-page document. For starters, three use cases, a 90-day plan, and an honest assessment of your current state are enough.
Key Takeaways
- 41% of German companies are already using AI-but most without a documented strategy (Bitkom, 2025).
- A 90-day plan is enough to get started: identify use cases, launch a pilot, and measure results.
- Budget for Phase 1: €20,000 to €50,000 and four to eight hours per week are sufficient for an initial pilot project.
- Focus on quick wins, not moonshots: invoice verification, proposal drafting, support ticket responses-the best entry points are unglamorous but effective.
- Your boss wants numbers, not PowerPoint slides. Track metrics from day one: time saved, error rates, cycle times.
The Situation: Your Boss Expects a Plan
You’re a team lead, project manager, or have just stepped into your first leadership role. At a recent conference, your boss heard that competitors are “all-in on AI.” Now they expect you to deliver a plan. The problem? You know AI is more than just ChatGPT prompts-but you’re not sure where to start.
Good news: you don’t need to be an AI expert to craft an AI strategy. What you do need is a deep understanding of your team and its pain points. Because the best AI strategies don’t begin with technology-they start with the processes that eat up the most time today.
The 90-Day Plan: 3 Steps Instead of 50 Pages
Month 1: Find where it hurts. Talk to five to seven people on your team. One question is enough: “Where do you lose the most time on routine tasks?” Typical answers include reviewing invoices, preparing quotes, sifting through error logs, or responding to support emails. Write everything down. Sort by: How often does it happen? How much time does it take? How prone to errors is it? Tasks that are frequent, time-consuming, and rule-based are your top AI candidates.
Month 2: Launch a pilot. Take your best candidate and implement it-not perfectly, not at scale, just make it work. A quote-generation assistant that auto-fills 60% of standard proposals is worth more than a white paper on “holistic AI transformation.” Budget: €20,000 to €50,000 for an initial pilot project, depending on whether you work with an external service provider or use a platform solution. Expect to invest four to eight hours of your time per week on oversight and feedback.
Month 3: Measure and report. This is where you’ll determine whether your boss gives the green light to scale. Compare results against your baseline metrics: How much time does the team save each week? How has the error rate changed? What’s the team’s adoption level? Turn this into a one-pager: starting point, action taken, results achieved, and next steps. Your boss doesn’t want technical details-they want to know if it’s worth it. Give them exactly that.
Quick wins that almost always work
The best AI entry points for SMEs are unglamorous-and that’s not a bug, it’s a feature. Low-key use cases offer high success rates and deliver measurable results fast. The most common quick wins in 2026:
Proposal drafting: AI pre-generates standard proposals; you review and customize them. Time saved: 40 to 60 percent per proposal. Invoice verification: AI automatically matches invoices against purchase orders and flags discrepancies. Support response drafts: AI drafts replies to routine customer inquiries; your team reviews and sends them. Knowledge assistant: An internal chat tool trained on manuals, error databases, and process documentation-eliminating time wasted searching for the right information.
What to avoid: Starting with your most complex process right away. Choosing predictive maintenance as your first use case requires clean machine data, sensor infrastructure, and a data pipeline strategy. That’s not a quick win-it’s a six-month project. For those laying the groundwork, our article on data quality in SMEs offers a solid starting point.
What You Can Tell Your Boss
If he asks tomorrow morning what your AI strategy looks like, you now have an answer: “I’ve spoken with the team and identified three processes where AI can immediately save measurable time. I propose we launch a 90-day pilot project with a budget of approximately 30,000 Euro. After three months, we’ll know whether it’s scalable.” This isn’t buzzword bingo-it’s a plan any CEO can understand. And that’s exactly what sets a pragmatic AI strategy apart from a theoretical one.
Frequently Asked Questions
Do I need technical expertise to develop an AI strategy?
No. You need process knowledge, not deep technical know-how. The key skill is identifying tasks that are repetitive, rule-based, and time-consuming. Implementation can be handled by an external service provider or an off-the-shelf platform solution.
What does an AI pilot project typically cost for SMEs?
Typically €20,000 to €50,000 for a well-defined pilot with a concrete use case. Add four to eight hours of your weekly time for oversight and feedback. Platform solutions like Microsoft Copilot or specialized AI tools further reduce entry costs.
Which AI use case is best for getting started?
Pick the process that occurs most frequently, consumes the most time, and is most rule-based. In many companies, that’s quote generation, invoice reconciliation, or internal knowledge search. Avoid complex processes like predictive maintenance for your first project.
How do I convince my boss to approve an AI budget?
Frame it in euros, not features. Example: “We create 200 quotes per month, each taking 45 minutes. With AI support, that drops to 20 minutes-saving 83 hours monthly. At an hourly rate of €55, that’s €4,600 in monthly savings.” A €30,000 pilot pays for itself in under seven months.
Can you really build an AI strategy in just 90 days?
Ninety days is enough to run a first pilot with measurable results. This isn’t a full enterprise-wide strategy-it’s a proof of concept. If the pilot succeeds, you’ll have the data needed to make informed decisions about scaling and additional use cases.
Editor’s Reading Recommendations
Source, header image: Pexels / Thirdman (px:7652126)