AI for Small Business

Build a clear mental model of what AI assistants can and cannot do reliably in a business setting.

• How language models produce answers and why they sometimes sound certain when wrong
• Prompting basics: context, task, format, and examples
• Tool landscape and free-tier baselines
• Hallucination examples and simple verification habits
• Choosing tasks that suit AI: drafting, summarising, organising, brainstorming

Hands-on lab: Complete three everyday business tasks in an assistant and log the time taken, errors found, and edits required.

Map your workflows and rank opportunities so effort goes where value and safety are greatest.

• Process inventory: where time goes in a typical week
• Effort, value, and risk scoring matrix
• Quick wins versus larger projects
• Data sensitivity: customer, employee, and financial information
• Writing a use-case card with owner, inputs, output, and reviewer

Hands-on lab: Build a scored list of ten use cases in Google Sheets and select the top three with written reasons.

Use AI to respond faster and more consistently while keeping people in charge of sensitive conversations.

• Building a reply library for frequently asked questions
• Tone and brand voice guidelines for customer messages
• Summarising reviews, surveys, and support emails for themes
• Escalation rules: complaints, refunds, health, legal, and safety topics
• Accessibility and multilingual communication checks
• Disclosure when customers interact with automated responses

Hands-on lab: Create a reply library of ten templates with escalation notes, then test them on five realistic customer messages.

Turn scattered documents and repeated tasks into checklists, standard procedures, and searchable knowledge.

• Converting notes into standard operating procedures
• Meeting summaries with action owners and deadlines
• Using NotebookLM to question your own documents with source citations
• Drafting quotes, proposals, and supplier messages
• Version control and approval steps for important documents
• Confidentiality rules when sharing files with AI tools

Hands-on lab: Upload non-sensitive business documents to NotebookLM, produce a quick-reference guide, and check each answer against its cited source.

Use AI to explore data and plans while checking numbers independently and recognising its limits.

• Explaining spreadsheet trends in plain language
• Scenario planning and SWOT drafts as thinking aids
• Why AI arithmetic needs independent checking
• Competitor and market research with source verification
• Budget and cash-flow questions to ask, and when to involve an accountant
• Bias and gaps in AI-generated market assumptions

Hands-on lab: Analyse a sample sales sheet with an assistant, recompute three of its figures by hand or formula, and report any differences.

Plan, run, and evaluate a small experiment, and write the rules your team will follow.

• Choosing a baseline and success measure before starting
• Pilot scope, duration, owner, and stop rule
• Collecting feedback from staff and customers
• Staff guideline: approved tools, prohibited data, review requirements
• Privacy, intellectual property, and vendor terms to read before adopting a tool
• Deciding whether to scale, adjust, or stop

Hands-on lab: Write your pilot plan and staff guideline, then present both for live peer critique and revise them.