Cash Flow Q&A: AI, Machine Learning and Forecasting

Posted 25 August by Barb Easter in Accounting, Bookkeeping, Cash Flow, Small Business

Barb and Blaine discuss the impact that machine learning and artificial intelligence will likely have on bookkeeping and accounting practices, and what clients want when we ask them about high ROI services.


Hi Everyone,

Barb Easter, Director of Partnerships here with Blaine Bertsch CEO of dryrun.

Today we are going to talk about robots. More specifically we are going to talk about the impact of artificial intelligence and the advances that it has made into the accounting and bookkeeping industry and in fact in small business and Blaine is going to tell us all why we don’t have to worry.


I’m going to talk a little bit about it in the sense of forecasting and where AI and machine learning falls short.

Now, I’m no AI, machine learning expert but I do my little bits of research, conversations, everything to kind of know what’s going on and in fact I’m sure that a lot of that technology will start working it’s way more and more into dryrun in the sense that it is going to help some of the automated concrete tasks.

The area of forecasting that I think can be really misleading is when businesses or bookkeepers, accountants, think that an automated system can do an accurate forecast.

Before I had this business, the type of business I had had big bills, big invoices and a lot of our customers had the same sort of things, big jobs, paid out in fazes, everything else, and I know how complex business is…

When you try and rely on a machine to say well, you know the last three invoices from this customer were paid on average 20 days late, so you are getting paid this 20 days late can actually be very misleading.

If you are not having those conversations because when you need that money, when that money has to come in, and it’s just an algorithm telling you when it may come in there is a little bit of usefulness there but the most important thing is quite frankly, there has to be some sort of conversation, some getting on the phone, getting on email and finding out what is going on in your clients business.

Because I can’t tell you how many times I’ve had that phone call and said, how come this invoice is running late…the reason isn’t because we always pay 20 days late, the reason is yah well, so and so is not in town and can’t sign off on the cheque or our supplier hasn’t delivered x or we haven’t got paid here and you start to find out, man that thing isn’t coming in for 30 days, 60 days, sometimes you get on the phone and the worst thing is I don’t think I’m getting that money…

And then you have to change that forecast real quick because now, things have changed and so what happens is invariably people dive back into a spreadsheet because reality doesn’t match what the robot told you. And reality is more important, reality is where you are going to get into some serious trouble.

So that is why in dryrun, even though we automate all kinds of stuff and we will be looking at all these different technologies, in the end, there is always the ability for that human intervention.

To say, I talked with this person, I know we are not getting this money in or no, I’m not paying that bill that they sent because they haven’t supplied what they promised and it’s not getting paid tomorrow and it’s not getting paid until they deliver what they promised or whatever the circumstances.


I sometimes have a sneaking suspicion that when I am talking to our users that when they are testing different softwares that are in the market that the idea of an automatic forecast is touted as being the latest and greatest and purely just the best.

And you know, every time and over time I more and more strongly disagree for exactly that reason.

So, it’s all well and good to say yes you forecasted and checked the box because it automatically told you what last years figures imposed onto this years figures but what I’m hearing from you, long story short, is that that’s not useful.


No. Like I know this so well because and we’ve talked about this a million times, dryrun is really a big part of what happened in 2009 with my last business.

The economy through 2007 and early 2008 it was flying high, we had work coming out of every orifice we had, like we did not have…


That’s a lot of work Blaine.


Yeah we had a lot…we didn’t even think about it. Cashflow, oh it’s in my head and I hear that from a lot of businesses it’s in my head, I know those cheque’s are coming in…


You could afford for it it to be in your head…


Yeah and you could kind of…we could kind of half stay ahead of it but it was still a stupid way to do it, I understand that now.

2009 is when we really got hit and suddenly it didn’t matter what happened last year it was the polar opposite, not only were things not getting paid, things were delayed, there were contracts that were getting eliminated partway through money flushed down the toilet.

Everything changed. It was such a dramatic switch and you know what, if I was using something automated it would look just Rosie. Yeah, we’ve got all this work…


Yeah, it would have led you right off the edge of a cliff…


Yeah exactly and what happened was I knew all this stuff in my head but it was so complex, you have so many moving parts, so many invoices that are supposed to come in but now all the terms are…when it’s going to come in is completely different than we thought, all the bills we’ve got to pay out, well now suddenly everything is in flux, and it got way too complicated and you actually have to plot it out on a timeline, which for me at that time was a spread sheet.

It’s the difference between surviving and not surviving. And fortunately we survived and we ended up thriving through an awful lot of hard work and a lot of very careful operational forecasting let’s say to make sure we knew where we were going.


And a robot couldn’t have guided you through that either.




That’s the ultimate wildcard factor. Right?




So it’s just wildcard compounded with wildcards. As you go through a deal, or a business cycle or any series of circumstances.

So what I’m hearing from you is that automation has it’s place in terms of manual data replacement…


Yes, absolutely. We do as much as we can and we are increasingly doing more and more of that and getting feedback.

In the end, every deal, every dollar and especially the bigger the invoice the bigger the bill, virtually every one has a story behind it.

If you sit down and talk to a business owner, you’ll find out very quick, Everyone has a story.

When you start talking about this client and then that client and that client they are all different. Each job had it’s own story and unfortunately what comes along with those stories is usually a bunch of heartaches.

It’s important to have the most realistic, risk adverse forecast that you can create which means err on the side of caution, move things around. Get a real look at what is going on, what is most likely to happen.

Don’t trust those due dates today, it isn’t going to cut it.


I can hear a certain kind of user say well, having those trends available, understanding those trends in business is valuable.

Yes, it’s valuable, but that’s valuable for economics it’s not valuable for an individual business cashflow forecast. And that’s the difference. We’re not economists.

It’s not economic software, it’s cashflow forecasting software. And I think that’s a really important distinction.

We’re not saying that trends should be disregarded all together, trends have their place it’s not what is most useful to the small or medium business.

So I wanted to make that distinction, I wanted to drill that home again.


Yeah, for small and medium business, well trends, I actually I’ll be honest with you, I don’t think they even help that much.


No, but that’s an economic thing.


Yeah, but if you want to look at other data’s for a small or medium sized business that’s going to make a huge difference in your business that’s where we are talking profit margin for different services or different products and your sales and revenue forecast, how much it costs to acquire a customer, things like that…that concrete operational information is going to make a big difference for your business.

Don’t get fooled into thinking that some generic trend or what’s happening with other businesses is going to apply to you, because until we are all Microsoft sized…it’s not going to be a big help.


Yeah, you can’t, this is not the one, this is not the box that you want to check and pretend like you’ve gone through the motions and haven’t really walked through your most likely scenarios for cashflow.

But Blaine, thank you so much, we are almost out of time again.

This is a really chewy topic so I appreciate you taking the time to walk us through this little overview and for anybody that has more questions or feedback on this video or any of our other materials, we’d love to hear from you.

You can chat us up at

Like the article? Please share...Share on LinkedIn
Share on Facebook
Tweet about this on Twitter
Email this to someone