In almost every workshop I teach, someone tells me the same thing: “AI is for big companies.” And it is true that big companies got there first, with engineering teams and budgets that neither you nor I work with. But when you look at what they actually did, almost all of it comes down to tasks your team already does every day: answering customers, finding information, reviewing documents, making content, keeping track of inventory.
Here are 20 AI use cases companies already implement, each with a company that did it and talked about it publicly (with a link to the source). Under each one I show you the version sized for a small business or a team. I can implement any of these 20 with you: we set it up with tools that already exist, and I teach your team to use it.
Customers and sales
1. Customer-service virtual agents
Klarna put an AI assistant in its app to handle refunds, returns, payments and invoice questions. In its first month it had 2.3 million conversations, two-thirds of its service chats (source).
On your team: an assistant that answers frequently asked questions on your website, email or WhatsApp with your real information, and passes to a person whatever it cannot answer.
2. Agent assist in the contact center
Telstra gives its agents, before and during each call, the customer’s history, a summary of why they are calling, a transcript and AI suggestions (source).
On your team: automatic summaries of calls and meetings, and a draft reply ready for a person to review and send.
3. Searching the company’s knowledge
Morgan Stanley built an internal assistant that finds and explains the firm’s own content for its financial advisors. More than 98% of advisor teams use it (source).
On your team: an assistant that answers from your manuals, policies and procedures (for example, a Claude project with your documents), so nobody has to ask the same question three times.
4. Lead scoring and next best action
HubSpot scores contacts and deals inside its CRM, ranks opportunities and explains which factors move each score (source).
On your team: your CRM’s AI set up properly, or a spreadsheet with AI help, that tells you who to call first today and why.
5. Marketing content and creative
Coca-Cola launched “Create Real Magic,” where creators made artwork with the brand’s assets using GPT-4 and DALL-E (source).
On your team: a brand kit with instructions (prompts) so your team makes posts, flyers and emails in your style, with a person who approves before anything goes out.
Operations
6. Personalized recommendations
Netflix turns what each person watches, browses and clicks into title suggestions made for them (source).
On your team: product recommendations in your online store, or follow-up emails based on what each customer bought.
7. Demand forecasting and inventory
Walmart uses models that forecast demand and decide how much inventory goes to each store and fulfillment center, down to differences by ZIP code (source).
On your team: AI reads your sales spreadsheet and tells you what to reorder and when, before you run out.
8. Delivery route optimization
UPS uses AI and machine learning to map the delivery route for each driver (source).
On your team: an app that optimizes the routes for your deliveries or service visits, and AI to organize the day’s schedule.
9. Predictive maintenance
Siemens installed sensors and AI at the Sachsenmilch dairy plant to detect equipment problems before they happen. Catching one failing pump in time saved the plant a six-figure sum (source).
On your team: a maintenance log for your equipment or vehicles, and AI that reviews it and reminds you what to check before something breaks.
10. Visual quality inspection
BMW Group uses image recognition on the production line to catch parts that are missing, out of place or defective, in real time (source).
On your team: your people take a photo of the finished job or the packed order, and AI compares it with the checklist before it goes out.
Finance, insurance and documents
11. Payment fraud detection
Visa gives every card payment a fraud risk score in real time. It estimated this helped prevent about $25 billion in fraud in one year (source).
On your team: the fraud protections your payment processor already includes, set up properly, plus alerts when an unusual payment comes in.
12. Financial crime detection
HSBC uses AI to monitor transactions and find new patterns of financial crime. It reported finding 2 to 4 times more cases, with 60% fewer false alarms (source).
On your team: a monthly review of your transactions with AI help, flagging what does not add up so your accountant can check it. It is support, not financial advice.
13. Insurance underwriting and pricing
Lemonade uses AI and data from its app and platform to underwrite and price some of its policies (source).
On your team (if you run an agency): AI organizes the client’s information and prepares the quote file, and the agent decides.
14. Insurance claims
AXA uses AI decision systems in its claims work across 15 countries, to speed up legitimate claims and detect fraud (source).
On your team: AI reads the forms, photos and emails of incoming requests, sorts them and prepares each case for the person who decides.
15. Document review and data extraction
JPMorgan Chase built COIN, which reads credit agreements and pulls out the key terms in seconds, work that used to take more than 360,000 lawyer-hours (source).
On your team: invoices, contracts, receipts and forms: AI moves the data into a spreadsheet and you verify it.
Technology, talent and science
16. Coding copilots
GitHub engineers use Copilot on GitHub’s own code to fix bugs, restructure code, prepare changes and review them (source).
On your team: if you have a website or internal tools, Claude Code or Copilot to make changes faster. That is how I build this site.
17. Cybersecurity
In an internal Microsoft study, security analysts who used an AI agent to review suspicious emails found the malicious ones up to 550% faster (source).
On your team: teaching your people to recognize phishing (fake emails that try to steal information), with AI as a second opinion, and turning on the protections your email already has.
18. Recruiting
Unilever used AI-assessed video interviews to hire recent graduates. The company said it saved 100,000 hours of interviews (source).
On your team: AI helps you write the job post, organize resumes and prepare interview questions. Who gets hired is decided by a person, with care not to carry over bias.
19. Medical imaging support
At Mayo Clinic, AI does the first pass on imaging studies, such as tracing and measuring structures on CT scans, to support radiologists (source).
On your team (if you run a medical office): diagnosis stays with professionals and certified systems; AI helps the office with appointment reminders, forms and work summaries. It is not medical advice.
20. Drug discovery
Amgen uses generative models that design protein molecules and simulate which ones could be safer and more effective, millions at a time (source).
On your team: faster research: AI summarizes studies, compares sources and prepares the first draft, and a person verifies every citation.
What they have in common
Look closely at the list: in almost every case, AI takes a first step (a draft, a priority order, an alert) and a person decides. That is what I always tell teams: AI makes you more efficient, not smarter. It is a strategic tool, not magic.
An honest note: each company measured its results in a different way, so they cannot be compared with each other, and they are not a guarantee of what will happen in your business. What they do prove is that these tasks are already done with AI, every day, at real companies. To see what that means in practice, I recommend reading what it means for a company to use AI.
Where to start
You do not have to do all 20. Start with one: the task that takes the most time from your team every week. Find the case above that looks most like your work, and try that version first.
Want these tools compared in depth? Check the unbiased reviews.
