Advanced AI in Excel & Databases

1 x 2.5 Hour Bespoke Session (online, with one 15 minute break)

For teams whose work lives in spreadsheets: messy documents turned into structured data, then into analysis, dashboards, and a database an AI agent can actually use.

Summary:

A bespoke session for finance and data teams, run on the organisation’s own kind of material and on freely available tools, so nothing depends on a paid rollout. The first half follows one Excel pipeline end to end: an unstructured document becomes a clean table, the table becomes an analysis sheet, then a formatted report, then charts, then pivots, and finally a KPI dashboard, with the prompt for every step written out. The second half moves from spreadsheets to databases, showing why an agent with a live database it can read and write has memory where an agent with a fixed knowledge base does not, and building a working library agent live. It closes on talking to your own data and on the wider tool picture: which chatbot for which job, what is genuinely worth paying for, and no-code workflows.

Course Breakdown:

Excel and Generative AI

• What is realistic with AI in Excel
• Which tools actually do it
• Use case: data creation from sources, an unstructured document turned into a structured table with a verification column and a rule against inventing values
• Use case: analysis built from the extracted data, on formulas rather than hard-coded numbers
• Use case: Excel’s designing abilities, reformatting a sheet without touching the data underneath
• Use case: visual analysis, charts and conditional formatting with a written insight under each one
• Use case: pivot-based analysis, with the findings stated in plain language
• Use case: automating text extraction with logic, one prompt per new column
• Use case: KPI dashboard builder, a flat monthly sheet turned into a dashboard with trends and risk colouring

Database management using AI agents

• Online database options
• What a database really is
• Building the database first, before you build the agent
• How agents work with databases, and where they struggle
• Use case: a project and prompt library agent, built live, with version history so the library compounds

Talk to your data

• The problem: an export that looks like a calendar but is really a picture of a table
• The options weighed, and why connectors decide which tool wins
• Building a team calendar agent, and testing it with plain language questions

Further in

• Free versus paid chatbot capabilities, feature by feature and price by price
• What a workflow is
• The no-code tools for building one
• Workflows for translation
• Summarisation
• Text and blog generation

Outcome:

Participants leave having run the whole pipeline themselves, from a messy document to a working KPI dashboard, with reusable prompts for every step. They will know what a database is, how to build one before pointing an agent at it, and where an agent stops being reliable. They also leave with a clear view of which tool suits which job and whether they need to pay for any of it.