[Chris Argent]

And is GENai ever going to compete? @adambr ? https://openai.com/index/improvements-to-data-analysis-in-chatgpt/

[Chris Argent]

Ps – great chat about the above here – https://gencfo.codeomega.co.uk/themes/copilot-on-excel-magic-or-make-believe/

[Tanbir Jasimuddin]

Co-Pilot – so far, summarising and transcribing Teams calls, summarising emails

Power Query – connecting to a data source, selecting only the columns I want, renaming, reordering. Some cleaning tasks like splitting, or removing characters. Reformatting columns (e.g. decimals, dates). Adding calculated columns, then pushing either into Power BI or Excel. I then schedule a refresh.

[Chris Argent]

@thefpandaguy @momentman @lynnet tagging for support.

I used Co-pilot to create a formula today, I didn’t know it existed before I asked. I followed my own advice (see below) and learnt the formula, validated it, and now use it freely.

Note, I didn’t tag it “AI generated” as the human aka me, took over the responsibility, once I had validated it.

Fun fun!

[Lynne Titley]

Hey all ????

I’m routinely using Copilot to draft meeting notes/actions from call transcripts, & draft some of the content of papers on specific topics. This week I’ll be using it to come up with some training ideas, inc. programming/scheduling, & materials for a specific series of training I’m intending to deliver.

For Power Query, I’m going to give a handful, off the top of my head, of more finance-specific potential use cases, for those who haven’t dabbled yet and want to better understand the possibilities. Which of these might be relevant for you will depend on what other data warehousing/reporting solutions you have in place:

1) ‘Transform’ data to make it ‘usable’ – e.g. you have a formatted report that comes out of an ERP system with a bespoke layout – maybe a debtors/creditors report, and you want to strip out all the ‘noise’ (titles/subtotals/blank rows/irrelevant bits) and shape that data into a ‘structured’ data set (table) that you can use for analysis, e.g. pivot. PQ’s key benefit is that once you’ve written the query, you’ve taught PQ how to transform that data set, and when you rerun the source report later (or it runs automatically), a quick refresh of the query will repeat those steps to transform the new version without further work. Bit like a macro;

2) Consolidate multiple monthly versions of a report into a single file (i.e. database effectively) for analysis – e.g. many monthly TBs, many monthly payroll files (careful with your PII security obvs), many monthly project cost reports. PQ can consume source data from a folder with multiple files, not just one;

3) Disaggregate a single report/data source into separate data sets for different people to see – e.g. budget reporting – you could set up separate query files that pull from the same centralised budget and actuals source data, but filter down on a criteria, like a name or cost centre, to show only the relevant records;

4) Mash different data sources – PQ can ‘merge’ tables on defined criteria. An equally if not more powerful option here is to clean your data using PQ and then create a data model using Power Pivot. Example: budget vs actual analysis – your budget data might come from a different source to your actuals. Merge them into a single data set for variance analysis;

5) Collate data – similar to 2) but instead of thinking about getting recurring versions of the same report from a system, maybe you’re issuing a template out to 10, 20 or 50 colleagues that you need them to complete & return. Use PQ to consolidate the returns into a single data file – lose all that copy & paste!

These are all things I’ve done quite a bit of with PQ in different businesses & circumstances.