Jenny Blewett FCA, Head of Finance at Cambridge-based Costello Medical, champions the fusion of technology and finance.

Incorporating AI into her daily work routine has been transformational, saving vast amounts of time, improving efficiency and allowing her to do things she didn’t believe were possible. Here she shares the practical ways that AI is transforming her job every day.

08:39 – Ask Copilot to go through my emails

I’m back from a lovely few days in Cornwall. The first thing I do is turn on Microsoft Copilot and ask it to summarise my emails.

I’m specific in my prompt as I require a list of actions triaged urgent, non-urgent and nice-to-have so it gives me a really good idea of what's in my emails.

Copilot took a while to get to know me, it now knows my writing style, my role in the business, what I'm interested in and who I interact with.

08:57 – Automatically respond to emails

The version of Copilot I use has a “Draft with Copilot” button, I use it frequently. However, I always proofread the emails before I hit send. They often contain AI giveaways like exclamation marks and double dashes. For me, AI is a starting point and not the end game.

09:45 – Explain a new piece of legislation to my team

The EU VAT rules are shifting towards mandatory structured electronic invoicing, I need to email my team about what’s changing. I expect I’ll just add a short introduction and draft most of the email myself.

Instead, AI surprises me by pulling together the relevant guidance and clearly summarising the impact by region, which means my team can validate the information quickly.

10:15 – Record, summarise, action meetings and more.

I set AI to record our bi-weekly finance team catch up. We check in with it mid meeting to see if there is anything we might have overlooked. It occasionally highlights points we have not considered.

We ask AI to summarise the action points while everyone is still in the room, so we’re all clear on what needs to happen next, and we have space to refine the actions if required.

12:20 – Do some learning

I need to learn about SaaS (Software as a Service) accounting. I have some very specific accountancy related questions about capitalisation of development costs.

Copilot helps me write board-level responses. I then send the paper to my team to validate it, to make sure there aren’t any hallucinations in it.

I use the updated Copilot with links to source documents, so we can scan the sources and sense-check the information.

13:06 – Grab some food

Somewhere between back‑to‑back meetings and a calendar that suggests lunchtime is optional, I set myself a simple goal; find lunch within five minutes of the office that is low in calories, high in protein, and not a biscuit at my desk.

Within seconds, AI produced a shortlist of realistic options within walking distance, factoring in protein content, calorie density, and the critical constraint of not exceeding a five‑minute radius.

14:36 – Use specific AI functionality for Excel

I need to cleanse some data and get rid of the dashes from bank sort codes, AI immediately suggests “=SUBSTITUTE(A1,”-“,””).

I love AI when working with data, particularly in Excel. Newer functions like GROUPBY and PIVOTBY have fundamentally changed how I work, and AI has been invaluable for helping me get the most out of them.

14:45 – Run automated reports

I need to update the monthly revenue reports within Power Query. Power Query is the data shaping engine in Excel and Power BI which allows me to clean, and transform data automatically, so instead of manually fixing spreadsheets each month, I press refresh and let the steps run themselves.

When I first started using Power Query, AI helped me translate models built in Excel into Power Query in baby steps.

Likewise, I showed it a single-row Excel calculation and asked Copilot to generate the equivalent DAX (Data Analysis Expressions, the language used in Power BI) for an entire dataset, to enable me to create a profit and loss statement within Power BI.

15:47 – Using AI for coding

I want to build a new cash flow model.

I use AI extensively for coding tasks. I regularly pull finance data directly into Excel using APIs (application programming interfaces), which are essentially secure connections that allow different systems to talk to each other and automatically exchange data and then manipulate the data using VBA, Excel’s built in scripting language.

Now, instead of jumping straight into code, I can talk the problem through with Copilot and ask a more fundamental question first: what is the best tool for this job? Should it be Power Query, Power BI, VBA (visual basic for applications) code, or even a Power Automate workflow? It gives me the ability to access new tools.

17:05 – Use AI to streamline a finance process

We are reviewing our accounts payable process. I like to keep processes under regular review and find ways to improve them. We use AI to evaluate the current process and suggest improvements. The trick is to keep feeding it back to refine the process.

AI then assists me in writing a VBA script that automates the cross-check between the bank file and accounts payable file, to ensure everything matches or to highlight anything that doesn’t.

19:02 – Convert a friend to the benefits of AI

In much need of some exercise, I head to Zumba. After class, my friend pushes back on the idea of using AI.

I convince her to try it and ask my phone to: “Plan an itinerary for a family of four going to Italy that would like history, pizza and pasta.” Another friend joins and we try: “Malta culture vulture one week.” They are converted.

20:35 – Homework with my son

He asks me to explain longitudinal drift. After answering “longitudinal what?” We ask AI to actively teach a 12-year-old what longitudinal drift is. It asks questions to assess knowledge level and breaks it down for us.