Randy and Brian continue their CES 2026 coverage by focusing on artificial intelligence (AI) trends and new products that can help accounting firms, finance teams, and their clients. The Accounting Tech Lab is an ongoing series that explores the intersection of public accounting and technology.
View the video below:
Or listen to the podcast using the below player:
Transcript (Note: There may be typos due to automated transcription errors.)
SPEAKERS:
Randy Johnston, Brian F. Tankersley, CPA.CITP, CGMA
Brian F. Tankersley, CPA.CITP, CGMA 00:00
Welcome to the accounting Technology Lab. Brought to you by CPA practice advisor, with your host, Randy Johnston and Brian Tankersley.
Speaker 1 00:09
Welcome to accounting Technology Lab. Brian and I are coming to you live from CES 2026 and this particular Yeah, you do. We want to use the badges. Here we go, badges. And it turns out, we’ve enjoyed our time here at the show. The main things that were being shown were AI powered things and robotic things. And this session, we want to talk about the AI items, The Good, the Bad, the Ugly that are out there. Now we saw some really ugly AI things, but I want to finish up with the really good things. So we’re going to just introduce a little bit of the ugly things. For example, cooking devices, AI microwaves and AI washers and AI dryers and AI lawn mowers and AI pool cleaners. Yeah.
Brian F. Tankersley, CPA.CITP, CGMA 00:55
I mean, can we agree that if you spill ink on your shirt or that or that if you put a stake in the microwave, AI is not going to save you. You know, it’s, it’s, it’s almost like some of those folks are just checking boxes to do things. But it was a it was nonetheless, there, there. There is more stuff here. And we did see a pretty significant shift in how business how they think businesses are going to deploy AI, in particular, small
Speaker 1 01:25
businesses, yeah, now the AI robotic shifts, you know, again, kind of desktop units, maybe. But we did see and stopped and talked with many, many different vendors that were building software products that didn’t seem all that mature. For example, I’m thinking of the AI medical record note taker as an example. And you know, again, I watched and listened to these various pitches where they were trying to use AI to accomplish various goals. But almost all the software products we saw on the floor were only partially baked.
Brian F. Tankersley, CPA.CITP, CGMA 02:03
And I will say this that, you know, we know in startups, you can’t, you can’t solve everything at once. We’ve worked with, you know, the two of us have worked with dozens of startups together. And I will just say that one of the things that you do in a startup is you set up with what’s called a minimum viable product. That is the that’s the smallest amount of functionality that you can get away with charging people money for. And the idea is that you start out with it, and then you iterate, and you get better and better and better. Well, the problem with with a lot of these tools is they were minimal and they were products, but they were not viable, and they just didn’t generate reliable results or even usable results.
Speaker 1 02:45
Yeah, I understood some of their visions, but they needed to spend more work on
Brian F. Tankersley, CPA.CITP, CGMA 02:53
their market. Their idea of what was minimal and the consumer’s idea of what was minimal were quite different,
Speaker 1 02:59
yeah, but there were a couple of items that were clearly beyond minimal viable. And one of those was what we saw at the Qualcomm booth with the dragon wing AI on premises appliance. And they built a device that could actually run large language models on that device, another vendor, competitive vendor, u green, also had their network attached storage. Now we’ve talked to you about that in the infrastructure ces session, but the NAS devices that they have the network attached storage actually have local AI powered data organization, and they’ve got different names for that, but they are doing the universal storage and filing of documents and retrieval of documents down at the content level. And they’ve got an AI chat and voice memos and so forth, all pretty impressive stuff.
Brian F. Tankersley, CPA.CITP, CGMA 03:55
So one way, I mean, if you think about it, if you’re going to put your files on your NAS, isn’t it amazing that you can use AI against it, and then that AI never goes anywhere? And that’s the thing that we’re we’re talking about here, is because everybody’s worried about how they’re going to keep their data private, you know, because you want to show your data to your AI models, but you don’t want somebody else to be using those same AI models and be able to extract your data. And so it’s it, you know this, the fact that this Nas has built into it, the all of these AI capabilities to index documents to and to recognize pictures, to to do all kinds of to reorganize documents. I mean, it’s, it’s amazing the things that this can do with an internal AI model,
Speaker 1 04:43
yeah, so maybe some of you have used the photo options from Amazon or from Google or from Apple. All three are similar cloud based type capabilities, but this universal search with U Green was amazingly. Good, because it would recognize in pictures various topics they demonstrate.
Brian F. Tankersley, CPA.CITP, CGMA 05:04
So you could, you know, just like you do the CAPTCHAs that will say, show me all of the, all of the blocks in this grid that contain a stop sign or something like that, this had, this did similar things. So, you know, show me. Show me pictures of dogs out of my bed of my photo collection, yeah.
Speaker 1 05:21
So they had created a thing that they called the Olya, U, L, i, y, a, a, I chat, which was a large language model that you could make queries against. So what I’d like for you to do is picture how you might have used any of the large language models, Claude, chat, G, P, T, copilot, whatever, and you’ve searched for things. Well now search for things on your NAS by name. Now you and I aren’t really pack rats, but we’ve been
Brian F. Tankersley, CPA.CITP, CGMA 05:49
doing things. I’m a digital hoarder.
Speaker 1 05:52
Okay, well, maybe you are, maybe I am. But you know, the problem for both of us is we’ve been doing things a long time electronically, and we have kept things that we’ve done in the past for a long time, but we have so much content that we’ve created through our own labor, not through AI labor, but through our own labor, that it’s actually hard to file everything, and these tools are automatically organizing these files and allowing you to search.
Brian F. Tankersley, CPA.CITP, CGMA 06:22
So think about it as a document management system where you don’t have to go in and enter the keywords associated with that document, so you can find it later. It will this tool generates the keywords, and to some extent, will even let you create fields on the fly, and then it will go find those. And the brilliance of doing this in a NAS is that most of your server type computers run it two, 5% 10% when they’re really cranking, okay. The beautiful part about this is you have processors that you’re already paying the electricity on, and they can sit 24 hours a day and index, you know, in my case, my 10, my 12 terabytes of data and and then, you know, in a few days, I have a really strong index, and it doesn’t marginally cost me anything.
Speaker 1 07:08
And so friends on this one, their feature, they call this was aI album. You can do a semantic search, like we’re talking about, or you can do a similar search. Here’s something.
Brian F. Tankersley, CPA.CITP, CGMA 07:19
Find other things. Yeah, here’s the child, find me. Find me. Other pictures of this child,
Speaker 1 07:24
or you can just do flat image recognition, where it will classify faces and animals and so forth. So we were fortunate enough to have, you know, breakfast with one of your relatives yesterday, and you were looking for a photo, and I couldn’t find it. I could tell you were frustrated because you were using the Apple searching mechanism, and it just wouldn’t come up. Well, I bet the photos on
Brian F. Tankersley, CPA.CITP, CGMA 07:45
there, by the way, the photo, the photo I found out, was not on there, only because this person died in 1999 before I got a cell phone, before cell phones had cameras widely and so that’s the only reason that that I didn’t have that one. But I will say that it is very exciting because, you know, with with very few of us load our old photos into Apple or Google or other places. But you know, many of us used to have accounts at Flickr and other places like that. And the beautiful part about this is that is that you can, you can run it against everything you have.
Speaker 1 08:22
And my theory, Brian, is that here in 2026 AI is going to be dominantly powered by voice, yes. And if I’m right on that call, and I think I mentioned that we learned that at the CCH user conference, and was like, Oh, that does make great sense. And we’re seeing a lot of the AI engines now being driven with voice well, this you green Nas also can do voice memos, so you can translate audio, translate a transcript, and so forth in the privacy of your NAS.
Brian F. Tankersley, CPA.CITP, CGMA 08:55
And you can record meetings and feed them in, and then the tool will summarize them as well.
Speaker 1 09:01
Yeah. And during the last two weeks of December of 25 and the early part of 26 I think I’ve answered this transcription question for CPA firms a dozen times or more. And of course, you have used otter for a long time, and we’ve used teams and zoom. And you know, there’s actually fireflies. There’s a 10 plus of these tools,
Brian F. Tankersley, CPA.CITP, CGMA 09:22
yeah, and they’re transformational, okay? And you should be looking at them, but you also need to look at the privacy
Speaker 1 09:29
of them, yeah, and that is the big issue. But this privacy problem is really not a big deal if you’re doing it this way. And then the major breakthrough we expect in 2026 on document management is the content of documents will be extracted, a naming convention can be applied, and all of a sudden you’ve got automatic organization of your files. Yeah.
Brian F. Tankersley, CPA.CITP, CGMA 09:54
So, so it’s like your your your personal assistant that used to file things in the paper do. A’s, you probably had a personal assistant. I was never senior enough when, when we still had secretaries, but, but I will say that your personal assistant that filed things away from you for you. You now have AI that will do the same kinds of things. And so you can describe through chat what you want, or it can create the tags, and you can click on links, and it’s very much like, again, a document management system, where where the tags and the labeling and the flagging of what attributes this document has and what it contains can be generated by AI.
Speaker 1 10:36
So the documents, photos, downloaded files, it’s all organized.
Brian F. Tankersley, CPA.CITP, CGMA 10:42
If you want to see an example of what this looks like, I’m not ready to roll this out publicly, but you can go see it at wiki, dot, CPA, te.ch, and they’re actually summaries of the last 100 or so episodes of our podcast. And so I actually feed the otter AI transcript of it, the audio file, and that into it, and it generates a 200 to 200 word summary. It generates a list of things mentioned. It creates five interesting, five interesting quotes out of it. And it also creates 20 different social media posts, of which I pick two or three and send them out. So so this summarization of data, you know, but again, I’m using chat GPT for that, and I have to pay a lot to get the chat GPT teams version that that has the right privacy. The beautiful part about this is your NAS is doing this, and there’s no monthly
Speaker 1 11:37
on it. No monthly now, this company also had a camera system. Now they call it the sin gear syn, C, A, R, E series, and they had indoor and outdoor cameras that were also AI powered so well.
Brian F. Tankersley, CPA.CITP, CGMA 11:56
Now, now that’s not out yet. It’s going to be out second half of the year. But the thing about it is that it will recognize people and animals, and it can say, this is one of my people, and this is not a people that lives here. And so that’s a particularly interesting thing. You know, ubiquity has cameras, and of course, we’ve got ring and blink and all kinds of other things like this. But on this one again, the content and the AI is stored on this NAS, yeah.
Speaker 1 12:24
And you know one of the and just call me old school on this one. Part of the reason I haven’t put in a ring doorbell, for example, is I know the security is not tight enough for my tastes, and they’ve also been hackable and so on. And the bottom line here is, wait a minute. Can I get a doorbell that’s now on my secure system?
Brian F. Tankersley, CPA.CITP, CGMA 12:45
Well, and I have cameras in my garage and all around my home, but sometimes, when I come in from from doing a particularly nasty, smelly task, I get undressed right at the at the washing machine, I load the washing machine, and then I walk upstairs and jump straight in the shower. Okay? Well, I don’t want some AI engine seeing things that they can’t unsee, you know. And, and, you know. So, so the beautiful part about this is this never goes anywhere. So you would have, you would be the only person you and the people that could see this system were the only people that could potentially see things that couldn’t be unseen.
Speaker 1 13:25
And our Ketu associate, Steve Yas, has done a lot of this same type of programming in his own home, keeping everything in NASA’s and secure, but he built it all with open source content, and it will do things like recognize that the mail is being delivered, and when there’s a person that’s at the door, and so he’s got all
Brian F. Tankersley, CPA.CITP, CGMA 13:45
this bill, but he spent hundreds of 1000s of programming time if you had to pay for
Speaker 1 13:51
it, if you had to pay for it, and the amount of hours. And what I’m thinking about here is this is all off the shelf and private,
Brian F. Tankersley, CPA.CITP, CGMA 13:59
and it is, and they’re there. The tool set is so much more robust than what Steve had to create for himself.
Speaker 1 14:06
Exactly So, indoor cameras, outdoor cameras, doorbells and an overall panel. Now, one kind of creepy thing that they demonstrated, I thought, was baby crying in a room and in the other room, it tells you baby’s crying, you know. So it had some interior and I guess that’s okay.
Brian F. Tankersley, CPA.CITP, CGMA 14:26
I’d rather hear the baby crying and then have it say, babies cry, yeah, you know. But maybe that’s just me.
Speaker 1 14:32
Maybe it is just me. But these cameras had fairly sophisticated night vision in them, as did the Amazon ones that we saw, and so forth. So this cross camera awareness was like, oh, okay, maybe.
Brian F. Tankersley, CPA.CITP, CGMA 14:47
But bringing the AI to figure out, you know, because one of the problems I have with my current blink cameras, which are some first some second generation, is that my dogs will go up in my bee yard, where I have a bunch of these cameras. Dollars, and it will say that there’s a person in my bee yard or a vehicle. Well, no, it’s my 90 pound Goldendoodle or my 15 pound Bichon Bresee, you know. And so it’s, you know, it’s, we’re detecting my dogs as vehicles, and doing the wrong thing with the AI. It’ll do quite a bit better, yeah.
Speaker 1 15:17
And one of the things that they say is they will do multimodal AI identity recognition. So again, picture in your office, picture in your home, that it begins recognizing all your family members or all your employees, and knows those are legit. But when somebody non legit shows up, it identifies an alert.
Brian F. Tankersley, CPA.CITP, CGMA 15:36
You danger Will Robinson? Danger. Danger Will Robinson?
Speaker 1 15:40
In fact, they actually talked about their smart alerts, the monitoring zone, the warning zone, and then the actual alert zone. So again, I don’t want to overstate what this vendor’s done, but I like the privacy profile.
Brian F. Tankersley, CPA.CITP, CGMA 15:54
But remember, the cameras are still six months out. So you know our rule with ces things, if somebody says something’s coming out in the fourth quarter, it’s 25% chance it’ll ever make it to market. Okay? If it’s coming out in the second half of the year, it’s usually 6040, 7030, given the amount of time that they’ve spent on this, it’s clear that they’re already testing these devices. So we know they’re going to come to market. We’re just not sure exactly when, but it looks like a very interesting thing they’re doing.
Speaker 1 16:25
And you know, this is the vendor again, we found a few years back and said they’re doing the right things, and they convinced me all the more this year. Now, by the way, the vendor does have other things, for example, mice and cell phone and lots of other stuff, but we just wanted to talk about the AI piece here with you, because we saw a lot of fake AI here, or useless AI This seemed useful.
Brian F. Tankersley, CPA.CITP, CGMA 16:51
And I will say that that I’ve taken over doing a short term managed service project, managed service provider project for somebody for the net for the last few months. And one of the things I had to do was my company had to buy out the existing equipment that was being rented, and it included a Synology NAS. And having looked at the Synology NAS compared to this, it’s a whole it’s a different thing. You know, the Synology NAS is like a vacuum cleaner, and this is like a Roomba, you know, it’s a this thing solves, it’s this thing create, identifies problems, solves the problems. And the Synology just sucks so,
17:33
and the Synology is a good NAS,
Brian F. Tankersley, CPA.CITP, CGMA 17:35
it’s a good NAS, but, but in the, in the vacuum cleaner analogy, it just, it just picks up what you pointed at, as opposed to getting around the furniture and all that.
Speaker 1 17:45
Or, you know, using an old vacuum cleaner, UK name, Hoover’s up the data. Yes, servers up the data. All right, friends, we wanted to give you insights of what was happening here at CES that we thought applied to the practice of accounting, and we’re very cognizant about things that we think are a business value and a personal value to you, and we know that you count on us to be authentic and accurate reporters on this stuff,
Brian F. Tankersley, CPA.CITP, CGMA 18:14
almost authentic to a fault. Sometimes, my apologies if I offended any of your sensibilities.
Speaker 1 18:20
It’s all good. We’ll catch you again in another accounting Technology Lab podcast. All the best.
Brian F. Tankersley, CPA.CITP, CGMA 18:27
Thank you for sharing your time with us. We’ll be back next Saturday with a new episode of the technology lab from CPA practice advisor. Have a great week.
= END =
Thanks for reading CPA Practice Advisor!
Subscribe Already registered? Log In
Need more information? Read the FAQs