Phil Hobden: So, Billie welcome to the podcast! It’s great to have you here and great to talk to you about the always exciting and interesting topic of artificial intelligence. So, let’s jump straight in.

I guess the first question I want to ask is, and I guess probably the question that’s on everyone’s mind: should accountants be scared of AI? Is it coming for them and their jobs?

Billie Mcloughlin: No, it really upsets me to read this in the media all the time because what it’s coming for is the future of the profession. The fact is, if I was 17 again, deciding what I wanted to do as a career, and I was reading that by the time I qualified, there would be no job for me, I’d probably decide to go into a different industry.

So, the most concerning part is actually the future of the profession. We want to try and encourage new people to still want to be accountants. It’s still a great industry to be in, and if anything, artificial intelligence makes it more exciting because it means we can diversify what we do for our clients.

So, it’s coming for the future in a sense, but not in the way that we think it is. There’s a huge amount of opportunity. Everybody’s heard the saying now that it’s not coming for your job—it’s the people who don’t use artificial intelligence that will get replaced by it. And I do think that’s true to a certain extent.

I think we need to make sure that we’re aware of AI, understand the opportunities, and grasp them where they’re readily available—especially if our clients are using it. We’re seeing this day-to-day with it being in mobile phones now. Our clients will be using it in their businesses, but I think people need to not underestimate how much it will be expected from their point of view that we’re using it as well, as an industry.

Phil Hobden: Yeah, I think that’s a really good point. Do you think there’s a risk that there will be two tiers of accountancy firms—or two tiers of people in accountancy firms—those that embrace AI and those that don’t? Kind of like the start of the cloud journey, right?

Billie Mcloughlin: Yeah. And to be honest, I think the cloud journey, even now, isn’t necessarily two tiers, but definitely staggered. Because even now, I speak to practitioners who may be acquiring a firm, and they’re still adding up on spreadsheets.

I had one experience a few years back, not too long ago, maybe about 5 years ago. They were acquiring a practice where the team was still putting information on a spreadsheet and then adding it up with a calculator. It blows your mind that you wouldn’t even think that exists anymore.

So, it’s amazing to think that actually, it’s not just two tiers. There’s a huge amount of difference between the most technologically advanced firms and the least technologically advanced. It’s definitely a huge, huge gap. And what it will do is just drive a bigger wedge between those firms.

Something I say a lot to people is to get your ducks in a row—not just assuming AI can revolutionise everything. You have to get the basics in order. If you’re not already using things like cloud software, you’re not going to get the benefit of using artificial intelligence.

Phil Hobden: I actually just… I wept a little bit when you said about someone using Excel and then using a calculator. It’s like, you can… A, you can do that on Excel. But B, the fact they’re still using Excel? A little tear came to the corner of my eye there.

So, I guess if it’s not going to replace their jobs, it’s not going to replace them. It is still quite overwhelming, right? AI has gone from being nowhere to everywhere, and I guess that it is quite overwhelming. So, practically, with the firms that you talk to on a daily basis, and from your experience, where’s a good place for an accountant to get started with AI? What’s the entry point they can kind of grow from?

Billie Mcloughlin: Yeah, I just want to go back to what you said. 100%—I agree about the overwhelm. It’s kind of just been launched into our laps of, “Here’s AI, and you must use it.” In any other walk of life, whether it’s personally or professionally, if someone just threw something at you and said, “You need to use this, or you’re going to get left behind,” and you didn’t even know what it was, you’d be more reluctant to use it.

AI is becoming a bit of a subjective term now. You could get a room of 10 accountants and ask them what AI is, and every single one would come up with a different answer. It’s a bit like asking AI itself—you’d probably get 10 different answers.

So, there’s this aspect of understanding what it is first. But that’s not the most important thing. There’s also the element of understanding how AI can help you. For example, when I speak to practitioners, they say, “Okay, I want to start using AI in my audit department.” And I’ll ask, “Okay, but what for?” It’s not just about using AI; you need to know what benefit you want to achieve from it.

If you haven’t started with AI yet, and you don’t fully understand it, the best place to start would be using AI note-takers. We’re seeing a lot of people use this now. I join calls quite often, and we’re joined by an AI note-taker. For those who aren’t aware, this is essentially a robot that joins the meeting alongside you and acts as a virtual assistant, taking notes for you.

I use one called Fireflies. The reason I use this one is because not only does it do a transcript, but I can take sound bites from it, and I can search in the transcript. So, when I’m back in practice and have a meeting with a client—say, a farmer who tells me they’ve bought a new tractor—I can easily search for the word “tractor” in the transcript and find the relevant part of the conversation. It makes it easy to reflect.

There are other great use cases for these note-takers. They can send the transcript to everyone in the meeting, give you an idea of whether the connotation was positive or negative, and summarise action points and questions.

The reason I recommend starting here is because these benefits are quickly achieved, and from there, your team can get excited about using AI. They’ll start thinking, “Oh, that’s really helped me. How else can I use this?” So, it’s also great for team engagement.

Phil Hobden: Yeah, I love that. And, I mean, we joke—I always joke that sometimes more AI note-takers join the call in Silverfin than actual people. It’ll still join if you don’t, so sometimes you have like four or five note-takers and only three people. It’s like, “Oh, they’re really good.” I agree—they are really good! A great introduction tool. And like, we use a specific piece of AI here called Gong. Gong will even craft follow-up emails based on the action points and notes discussed in the call.

I’m not overly a fan of AI-crafted communication because I think it takes away the personality, but it’s a great tool if you’re not comfortable crafting that email yourself, or if you’re a junior in the office and not quite as confident. It’s a great jumping-off point to start with. It helps you get comfortable and get started.

Billie Mcloughlin: Yeah. And I quite often say, when using tools like these as a starting point, it doesn’t have to be your end point. So, if Gong crafts that email for you, that’s fantastic—use it as a starting point, then add the personality and the human element yourself.

Phil Hobden: I’ll tell you another great use I found recently. So, I’ve started some volunteering projects outside of Silverfin, which is super exciting, but it means I’m taking a lot of notes. And I mean, I’d love to put mine against your partner’s and see which ones are worse.

What I decided to do was scan my notes into ChatGPT and get ChatGPT to transcribe them for me, so I could then put them into a Google Doc and have a much clearer revision tool. And it works brilliantly. It’s so good at transcribing even the worst handwriting. I’d boldly say 98% of the time it’s accurate—and I’m definitely not accurate 98% of the time!

Billie Mcloughlin: Yeah, I mean, I use a Remarkable tablet to take my notes. It’s great—it syncs with Word and other apps. But as you say, it almost eradicates the use case for that, because now you can upload a photo of your paper notes and use something like ChatGPT to organise them. If you have a tab for specific topics, clients, or industries, you can always have that information on hand. You can even ask questions about the notes rather than sifting through them all.

Phil Hobden: Yeah, I think my ultimate plan is to put all of my notes into its own GPT and then use it as a search tool. I think that’s what I’m going to try and do. But that might be a little step too far for where we are on the podcast today!

So, look, let’s take note-takers as a good starting point, right? Great tool, could have a real impact within the practice. But I use Gong, I think you use Fireflies… and then there are a few others, or many others as well. So, in terms of going through that process, what hints and tips would you give to an accountant looking to find the right tool? Should they try everything, or should they be laser-focused and go on recommendations?

Billie Mcloughlin: Yes, don’t try every tool. There are millions of them out there. And also, this might be a little bit unfair today, and maybe a little bit controversial, but there are a lot of apps and software claiming to have AI because they know it’s a buzzword. If you go onto the website of many apps and software vendors now, the first page will say “AI” because it catches the eye.

hil Hobden: It needs to replace cloud, right?

Billie McLoughlin: Yeah, it has. Yeah. And even if people don’t fully understand what it means or why AI is being mentioned, it’s like, well, it has to have AI. And again, I still have these conversations with people now where they say, “I want a certain tool. It’s got to be on the cloud.” And I’m like, “But why?” If you don’t know why you want something to do it, then you’re just following the trend for oblivious reasons. But yes, people aren’t trying everything.

Crowdsourcing information is by far the best way to do it. So staying informed but not consumed is such a big tip in this world because there’s so much out there. If you can tune into a few trusted sources, it’s helpful. For example, the Digital Podcast is a great one. I watch Jason Staats quite a lot on YouTube. I personally have a membership called Tech Talk for members so they can stay up to date with those changes once a quarter rather than having to do all the research themselves.

As a practitioner, you don’t have time to research every product. So instead, tune into the right sources, listen, and if something sparks interest, then go ahead and research it. Another thing that works really well is a member of ours who runs a monthly “lunch and learn.” Within their team, they say, “Okay, the tax department wants you to look at this,” or “the audit department needs to look at that.” Then once a month, they all get together to discuss either new software or new features in existing software. They’ll ask, “How have you used that this month? What did you find beneficial?” It’s a great way to share use cases and crowdsourcing information works really well.

Phil Hobden: Yeah, I think you make a really good point about the podcasts and articles that, like Accounting Web, for example, are really good at breaking down some of those tools and things that people can use.

And absolutely, listen to other accountants, right? Like us. We all know there are hundreds of Facebook, LinkedIn, and other groups, even WhatsApp groups, with colleagues. I’m in a couple myself. I’ve infiltrated a few accountant groups, and every so often, I see someone ask, “Oh, we want to solve this problem, what tools are good?” And I’ll just pop in and go, “Hey, look, we use this because we find it’s X, Y, and Z.”

We’ll talk a little bit about ethics and security later because I actually think that’s probably one of the bigger concerns people have.

You mentioned that every bit of software is putting AI on their website, and I think that’s a fair point. I think you’re absolutely right. What question should an accountant be asking a vendor around AI when looking at their products? So they don’t just get sucked in by the hype of AI, which turns out to be yet another ChatGPT, right?

Billie Mcloughlin: Yeah, so with AI, I think what people do is they kind of lead with the wrong part of it—they lead with the AI and how that can help. Whereas actually, what they need to start with is: “Okay, as a practice, what is our problem, and how can this product or software solve that problem?” It’s kind of irrelevant whether it’s artificial intelligence, general intelligence, or just RPA (robotic process automation). It doesn’t matter as long as the solution solves the problem.

So the questions should really be about what problems you have and how that software can solve them. And if they’re using artificial intelligence, then I guess the questions would be:

• “How is that different to just using software?”
• “How can we get the most out of it?”

The other issue we have with AI, whether it’s built into software or using a standalone product, is: Do we know how to use it properly? So whether you’re using ChatGPT or a co-pilot built into a bookkeeping tool, the problem a lot of people are going to have is that this thing’s been placed in our laps, but where’s the education around it? Who knows how to ask the right questions?

Prompt engineering, for example, just came out of nowhere, and nobody actually knows what prompt engineering means. It’s essentially just a way of asking a tool a question. What we need to do is make sure our team is trained in asking the right questions, making sure they can use AI properly, and also recognise if AI is producing something incorrect.

One of the questions I always ask a tool when they say they’ve got AI is: “What if AI reaches its limit? If I ask your tool a question and it doesn’t know the answer, what would happen?” Quite often, in standalone products like ChatGPT, it might “hallucinate,” which essentially means it makes up an answer. We don’t want that in a tool where we rely on the answer to give to a client. We want the product to say, “My capabilities end here. I don’t have an answer.”

So I think that’s probably one of the biggest questions. But try not to lead with the AI. Instead, lead with the solution to the problem you’ve got.

Phil Hobden: And actually, I think that’s great advice across the board for any form of technology, right? Like, I always say that you shouldn’t get sucked in by the great coffee cup or notebook at the event. Just make sure that when you’re looking at software, it’s solving a specific problem you have.

And I guess that leads me to another question. You work with a lot of firms and you’ve been in practice yourself. What do you think are some of the key areas where firms could save time or maybe drive efficiencies with AI? What kind of impact could it have in an ideal situation?

Billie Mcloughlin: Yeah. So this is why I set up Friday. I feel like everybody talks about AI and how it’s going to change your life and revolutionise everything, but actually, nobody talks about how it does it. So this development has come out by Microsoft—here is how an accountant can actually use this development in your practice.

People need to not get wrapped up in the idea of how AI can change everything and just take it a step at a time. Some of the best use cases I’ve done demos for in past presentations—let me take you through a bit of the client journey:

1. Marketing and Awareness: To attract new clients, you could take frequently asked questions from a note-taker or ones within your practice and throw them into something like ChatGPT to make them nice and neat. Then ask, “These are the questions I get asked. Could you help me attract more clients or clients to use this specific software?” As an accountant, we’re not marketers, but AI can help with marketing tasks like blogs for websites or social media posts. Again, just make sure you add that human element.
2. Onboarding Clients: I’ve created an AI avatar of myself that I can feed a script to. Every time I onboard a new client, I just change the client name and get a personalised video. I don’t have to record the video each time. It takes about 10 seconds. There are also great statistics around video content, such as how many hours of YouTube are watched daily and how much content is uploaded to TikTok. Clients are getting used to consuming information through video.
3. Bookkeeping: There are some fantastic use cases in software. For example, asking software questions. Major bookkeeping tools now introduce chatbots for this. For trainees, ChatGPT is great for spreadsheets. You can ask it to create a formula like, “I want a formula that will make a pie chart from this data and highlight anything over £1,000.” It will create it instantly. You can even ask it to create a double-entry journal for a specific person.
4. Compliance: Use AI to explain tax queries. For example, feed a link to HMRC legislation and ask it to reinterpret it into layman’s terms, either for a junior team member or a client. You could even ask it to explain it differently depending on the industry. So a plumber may want it explained differently from a manufacturer.
5. Advisory: For advisory, you could use it for “what if” scenarios. For example, “Here’s an anonymised P&L from January to March. What would happen if revenue went up by 10% and theme costs went up by 20%?” There are lots of use cases for this, especially with ChatGPT or some of the tools that integrate AI.

Phil Hobden: I love that. There are so many great tips.

My wife uses it for website HTML coding. So she’ll get it to write the code for her. I just sit there thinking, “If I had this 20 years ago when I was building websites, it would have been a total game-changer.” Sadly, I didn’t!

Let’s talk about security and ethics now, right? It’s a big concern for people. Are people too concerned about the security and ethics around AI? Are they right to be concerned, and what can they do to make sure they’re being as safe as they can be?

Billie Mcloughlin: Yeah, I mean, they’ve got a right to be concerned, as accountants. Ethics and security are built into our studies, so I understand that. When I first started talking about AI a couple of years ago, I used to kind of brush over this because nobody was interested—it’s boring. But every time I’d finish a session, the questions would always be about ethical implications.

It’s interesting to look at how much we use Google now and how much data it has on us. People have kind of just accepted it. I think AI will eventually follow that route.

But there is concern, and people like Elon Musk, for example, don’t have the greatest reputation for trustworthiness, so there are concerns. I wouldn’t be putting my bank details into ChatGPT, even if I felt it was safe. And when you’re using something like that, it claims to redact and anonymise information when training the model again. But I still wouldn’t trust it.

A few tips I would say for using products with AI built in:

• Make sure it meets the same security requirements as the rest of the product. If you’re already happy with Microsoft or a bookkeeping product, ensure the AI element is just as secure.
• If you’re using a paid AI product, it’s essential because if you’re not paying for it, you’re the product. If you use a paid-for version, the security is much better, and using something like Microsoft Teams or ChatGPT within Microsoft is a good idea because it has enterprise-level security. Microsoft has a significant shareholding with OpenAI, so they use similar security.
• There’s also the option, when using these versions, to go into the settings and toggle the option to stop the AI from learning from your data.

Finally, I recommend having an AI policy. As accountants, we didn’t get into this industry expecting to deal with these issues. But it’s the reality now. Having an AI policy ensures that your team understands how to use AI safely and sensibly. You can guide them on what data is safe to use and how to anonymise sensitive data.

Phil Hobden: I love that. I think there are some really good practical tips there. So my last question, very quickly, before I ask the final question: What’s your favourite AI tool, Billie? What’s the one you couldn’t do without right now?

Billie Mcloughlin: Well, I think it would have to be ChatGPT. It’s definitely up there. But I think, if Microsoft keeps developing the way they are, I think Copilot is going to be the one. All of my information is essentially already in Microsoft. What I always say is it’s crazy that we can go online and find more about a rare frog in the Amazon rainforest than we can about our own business. The thought that we can go into something like Microsoft or Dropbox and ask questions based on our own data and get that information back instantly just feels like where we should be.

So, right now, it’s ChatGPT, but there are some other great tools emerging too.

Phil Hobden: Amazing. Last question then. What’s the one piece of advice you’d give accountants to cut through the overwhelming noise that is AI?

Billie Mcloughlin: I think I’ve already said this, but stay informed, but don’t consume everything. Make sure you’re not getting overwhelmed or spending too much time on AI unnecessarily. Stay informed but don’t get consumed by it, and make sure it’s always relevant. When you’re looking at something, keep in mind the problem you’re trying to solve.

Don’t try to be the hammer finding a nail. Make sure you have a use case for it before you run with it. Don’t think you need it to solve everything because, at the moment, we’re not at that stage.

But please don’t be frightened. Your clients are expecting this. Your team is expecting this, and it’s really important we embrace it. Otherwise, there might be a time when we get left behind. I don’t want to say something too cliché, but it’s probably going to have to be my last point.

Phil Hobden: Thanks, Billie. Great to chat.