Skip to content

Discogs - Batch Upload Feedback - 2025-08-27

Metadata

  • Date: 2025-08-27
  • Company: Discogs
  • External Participants: Jamie McKinstry (Controller), Stephanie Burns (Senior Accounting)
  • Palm Participants: Emma Sjöström, Jennifer Pearson
  • Type: Customer Call
  • Domain Areas: Cash Forecasting
  • Recording: https://tldv.io/app/meetings/68af24991db842001467cc6d/

Summary

Context

Feedback session with Discogs team on the new batch upload feature for forecasts. Stephanie walked through exporting open bills from Ramp and attempting to upload them to Palm. Session focused on understanding friction points and what would make the feature usable for their workflow.

Key Discussion Points

  • Exported open/scheduled bills from Ramp (filtered by status: not paid)
  • Current limitation: all uploaded rows go to same category
  • No bank account identifier in Ramp export - mapping fails
  • GL account and Department columns exist but don't map cleanly to Palm categories
  • Discogs pays most bills from same bank account - simplifies their use case
  • Feature positioned as "fine-tuning" forecasts, not primary data source

Pain Points

  • No automatic categorization - Must upload one category at a time, manual mapping required
  • Bank account mapping fails - Ramp export has no bank account identifier
  • Duplicate risk - Uploading same invoice twice, or uploading something ML already predicted
  • Manual effort vs value - "I don't know that it's worth Stephanie's time to put together the spreadsheet"
  • ML overlap concern - Recurring payments should be predicted by ML, not manually uploaded

Feature Requests & Needs

  • Automatic categorization - Read GL account/department and auto-assign Palm category
  • Duplicate detection - Flag if vendor name + amount already exists (manual or ML-predicted)
  • Bank account dropdown - Select target account in UI rather than requiring it in file
  • Vendor name field - Add proper vendor field to forecasts (not just free text note)
  • Upload from account page - Entry point that implicitly knows the bank account

Jobs & Desired Outcomes

Job: Fine-tune ML forecasts with known upcoming payments

Desired Outcomes: - Minimize the risk of duplicate forecast entries (manual + ML overlap) - Reduce the manual effort to format and upload AP data - Increase forecast accuracy by confirming payment timing

Job: Adjust payment timing to manage cash flow

Desired Outcomes: - Minimize uncertainty about when large payments will actually go out - Increase ability to see impact of moving payment dates on forecast

Domain Insights

  • Batch upload as fine-tuning: "I would view this upload more of like a fine-tuning the forecast, not like... the machine should be doing it right. But this is like, yes this payment is coming, but we're planning on paying it this week versus what you predicted."
  • Recurring vs one-offs: ML should handle recurring; uploads are for confirming timing and one-offs
  • Duplicate detection key: Vendor name is the critical identifier - "How do I know what you're adding is real if I can't tie it to... a description of what it is"
  • Cash-comfortable vs cash-constrained: Discogs doesn't need to watch cash closely; high-growth companies watching for payroll would use this more
  • Payment runs: They pay bills out of one main bank account, simplifying their setup

Action Items

  • [ ] Add bank account dropdown to batch upload UI
  • [ ] Implement automatic categorization based on GL account/department
  • [ ] Add duplicate detection (vendor name + amount matching)
  • [ ] Consider vendor name as proper field (not just free text)

Notable Quotes

"I would view this upload more of like a fine-tuning the forecast... this is like yes this payment is coming, but we're planning on paying it this week versus what you probably predicted." - Jamie McKinstry

"It definitely feels like if you have accurately got all of the big payments you're expecting and then you upload this file, it should be like 'Oh, there's that big payment' and match it and maybe update the amount to the actual amount." - Stephanie Burns

"I don't know that it's worth Stephanie's time to put together the spreadsheet in the right format if we're just fine tuning like a week here, a week there" - Jamie McKinstry

"My key takeaways will be like the automatic categorization and duplicates. That would make it feel usable for you guys." - Emma Sjöström

Full Transcript

Date: 27/08/2025, 17:30

00:00 Emma Sjöström: Oh lovely. Hi. Sorry. I just I'm trying out like so many different AI now, taking assistance and I was just like, nope, not you. Not? You tldvq. How are you?

00:07 Jamie McKinstry: Great. Have you been?

00:09 Emma Sjöström: A good I was just telling Stephanie, I just came back from vacation this week.

00:14 Jamie McKinstry: Alright.

00:14 Emma Sjöström: so, It's been a nice, taking some time and yeah.

00:20 Jamie McKinstry: Yeah, it's always hard to get back into the swinging. Things after being gone, like the work,

00:25 Emma Sjöström: Exactly first week it's just like Oh my God so much context so many things but it's good. It's cool too.

00:33 Jamie McKinstry: Good grateful. I'm excited for today. I think Stephanie, are you gonna be okay, driving? I think a good example, perfect.

00:43 Emma Sjöström: Amazing. Did you get my note in Slap?

00:47 Stephanie Burns: Yeah, I saw that.

00:50 Emma Sjöström: Really cool. So I'll be helping you out and telling you what you need to do in the Again. We just

01:00 Jennifer Pearson: Hell.

01:02 Emma Sjöström: Sorry, we just got started.

01:04 Jennifer Pearson: oh,

01:05 Emma Sjöström: But yeah, I would love it. If you would, would you mind screen sharing when you export from ramp as well? Just so I can really See, and what you're doing and how it works for you guys.

01:17 Stephanie Burns: Here. Yeah. Should I start or do you want to wait for Christian?

01:22 Emma Sjöström: No, Christian census regards but he's not able to make it today. He's traveling and doing conferences and it was just a bit of a hectic schedule.

01:33 Stephanie Burns: No worries. Okay, let me share my screen just Happened to be in Bill Pay.

01:43 Emma Sjöström: All right.

01:44 Stephanie Burns: Okay. So what kind of time period? Should I be exporting?

01:53 Emma Sjöström: I think, if we think about what we, what is currently viewable in the app, I guess it would be a bit mid-november-ish.

02:03 Jamie McKinstry: Yeah, I

02:04 Emma Sjöström: but, But it's fine to import. However, for a head you got. I think for today we won't be actually importing it so it would be nice to see. How much data were also talking about here?

02:18 Jamie McKinstry: Yeah, so I think like, instead of thinking it from a date perspective, we should think of it from like what's not sketch or like, what's not paid. So, when I was thinking about it before this call, I think we should just filter. If you click on the searcher filter,

02:33 Stephanie Burns: Okay.

02:33 Jamie McKinstry: Or.

02:36 Stephanie Burns: So, you just

02:39 Jamie McKinstry: Over. To the left. Right there. If you click there and then go to status, I click the arrow for status I would select everything besides paid because that would capture any bills that are open that are waiting to be paid or scheduled to get paid. Waiting for a match like an approvals basically.

00:00 :

03:07 Stephanie Burns: Well, I don't know, probably rejected but maybe not archive. Well I guess they don't have archived. Yes you can't take your arcade with you.

03:14 Jamie McKinstry: Yeah. Yeah.

03:17 Stephanie Burns: Okay.

03:20 Emma Sjöström: Nine.

03:22 Stephanie Burns: and then, We can customize the columns a bit and In the export, I don't know if that, maybe we'll just try with all first.

03:35 Emma Sjöström: Yeah.

03:37 Stephanie Burns: And then we just hit export.

03:39 Emma Sjöström: What I would suggest before, that is just this, all the same type of Payables, or other different categories.

03:51 Stephanie Burns: Do you mean like, are they all going in the same, like Gl account? Or do you mean? Like, I mean, they're

03:59 Emma Sjöström: so,

04:00 Stephanie Burns: Bills. All of our open bills.

04:04 Emma Sjöström: Yeah, so if in palm, if you were to map them to a corresponding category that we've set up for you guys in palm with, with these all be in the same category.

04:17 Jamie McKinstry: They wouldn't not and we don't use that category in ramps so that might be. Something, but we do like and when Stephanie exports it you'll be able to see that there is a column for GL account. So they're in some descriptor as to what

04:31 Emma Sjöström: Huh.

04:34 Jamie McKinstry: the expense is. I don't know if it Palms able to like, read that and take a guess or What?

04:42 Emma Sjöström: So how it currently works. Is a we would expect the file to contain only transactions or like future events for a specific category. so if it's at all possible to apply a filter for the GL account or anything else, that would proxy that for you guys already at the export level, I think that would be a nice way to do it.

00:00 :

05:10 Jamie McKinstry: Yeah. Instead of many exports Stephanie because I think it'll come and Csv, maybe we just take the all of it and then manipulate it.

05:17 Emma Sjöström: Okay.

05:21 Jennifer Pearson: Especially because sometimes the columns move around and stuff. Don't they like when we've done that this set? Things before sometimes they're not consistent.

05:29 Stephanie Burns: Yeah, me. And change, what I'm sharing. So you can see the file.

05:37 Emma Sjöström: All right.

05:54 Stephanie Burns: Okay. All right. It was giving me like don't have an infinity mirror. So okay. So, see if I can. it's like, We have.

06:30 Emma Sjöström: You have some form of categorization going on there, right? Column fold. No ad.

06:38 Stephanie Burns: 80.

06:39 Emma Sjöström: Or am I?

06:40 Stephanie Burns: Yes. Um, I'm not loving that you have. I guess it does have the GL account but I'm not living these multiples. Like that's not very useful.

07:03 Emma Sjöström: So, would it be tricky for you at this point? To kind of just like, say Which corresponding Palm category. Things were going to.

07:14 Jamie McKinstry: We'd have to create a key, I think.

07:17 Emma Sjöström: Yeah.

07:17 Stephanie Burns: well, I think like,

07:18 Jamie McKinstry: Based on GL account. And even then, I don't even know if DL account would be the best one, because sometimes like, Departments. Like Might be a better one. To pivot after because, you know, like we'll post something to prepaids. Versus an expense account.

07:45 Stephanie Burns: Most of what this is is. Either web hosting or operational payments, right? Like we're not gonna have Mmm, employee expenses in here.

07:57 Jamie McKinstry: And payroll related ones.

08:00 Stephanie Burns: There. Yeah. Good point.

08:02 Jamie McKinstry: But yeah, otherwise there's no taxes. Postage is so minimal these days. So it's really in like three buckets, I would say.

08:14 Jennifer Pearson: Because there is this date that This data is in Palm already.

08:19 Stephanie Burns: No, because we're only giving poem the credit card transactions. If it's been paid, it would be in poem.

08:28 Jennifer Pearson: Yeah.

08:28 Stephanie Burns: like, if it's gone to the pink account,

08:31 Jennifer Pearson: With these and this the same information in the files.

08:38 Stephanie Burns: I mean, I think it looks a little different in the credit card one.

08:42 Jennifer Pearson: Right.

08:43 Stephanie Burns: But yeah, like, I mean the mapping that exists for our credit card transactions. And our bank transactions.

08:52 Jennifer Pearson: I guess Emma is that when you, when we do these type of uploads is there. No

08:53 Stephanie Burns: Should apply to this.

09:02 Jennifer Pearson: categorization feature moment not

09:03 Emma Sjöström: No. For now, it's as since a lot of our users are able to export kind of like, Oh, these are all from Category A or these are category B, so they can put filter in their system. And for example, payroll or they could get like an export. That's already sort of categorized.

09:23 Emma Sjöström: I'll be not the same name as the poem category. We just it's a V star versions. So we just said, Okay, we'll solved the categorization later. But we can still for the purposes of this. Exercise, We can still use this file. But could I ask you to just remove a few lines because everything that we upload gets saved in the database, although you don't see it.

09:49 Emma Sjöström: So maybe we can upload like 10 lines or something. Just to try it out.

09:56 Stephanie Burns: Okay. Yeah. Well, definitely take out the multiples.

10:02 Emma Sjöström: What?

10:02 Stephanie Burns: And but yeah, I can.

10:08 Emma Sjöström: And it doesn't matter which ones you upload, you can just keep the top 10 ones as long and then. Delete the rest.

10:15 Jamie McKinstry: Right. Yeah, the category really doesn't matter since we're gonna test it to the

10:16 Stephanie Burns: Okay.

10:19 Jamie McKinstry: same one. So anything below, Ryan Line, 11.

10:23 Stephanie Burns: Open palm first.

10:24 Jamie McKinstry: Right. Yeah.

10:27 Emma Sjöström: Exactly. Nice. All right, let's let's go on to the Paul map.

10:50 Jamie McKinstry: I did notice that it made me log in with my Google. That was awesome.

10:58 Stephanie Burns: Okay.

11:00 Emma Sjöström: Nice. So there's currently one place in the app where we've just added this batch upload feature. It's a It's on the Forecast page exactly in the top right corner, you have an ad forecast and if you click that, you'll get two options. And we're going to try this one.

11:21 Emma Sjöström: Exactly. And then we'll just upload the file you created. Nice. And here this is the caveat right now, right? The restriction that we have in a product that all the kind of Rose will go into the same category.

11:39 Stephanie Burns: Okay.

11:42 Emma Sjöström: And now we used an element automatically map. Or suggest a mapping for you guys. Between. Account, which is the palm bank account, and the amount that is due. The value date. and also we have it's it's going to be optional This note concept, which you guys have a description column so that might be nice extra context for you guys to have on every added forecast item, The you can change these If you look at the account Palm field for example, you'll notice that the Banking Partner ID is like a dropdown so you can click it.

00:00 :

12:30 Stephanie Burns: Yeah, I was just wondering so banking partner idea. Do they got that from this

12:31 Emma Sjöström: So you can change like all the the AI, got it wrong. It's actually some other field and this is this is where we haven't set you guys up because we we store some sort of nothing typically between the some, some feel and and the bank corresponding bank account in palm, But that we can make this work for you guys, if you find value of it out of it.

00:00 :

13:05 Stephanie Burns: spreadsheet.

13:08 Emma Sjöström: I sure hope so.

13:11 Stephanie Burns: I'm just trying to figure out what that is.

13:13 Emma Sjöström: Yeah.

13:15 Stephanie Burns: Oh yes, but it's all blank.

13:18 Emma Sjöström: Is it in the spreadsheet at all as a column?

13:25 Stephanie Burns: Um, okay.

13:26 Emma Sjöström: So it's safe to say it gets wrong here.

13:30 Stephanie Burns: Yeah, I guess. I mean if we had anything there, it might figure eight but we don't, I'm not sure we have anything in this file that indicates what account it's gonna actually be paid out of Yeah.

13:47 Emma Sjöström: That's okay, let's move ahead with the, if you put, let's pretend we have some way of identifying the account.

13:54 Stephanie Burns: Okay.

13:55 Emma Sjöström: And if you hit confirm, You will see that some Processing is taking place. We are using a mix of AI and non ai take in the background to go through all your line items. And then present the industrial ones back to you. And we call it invalid for now just why it's not possible to ingest this data right now and as the it's because we're not really Able to asset now.

14:36 Emma Sjöström: Mark this to your corresponding bank account I would love to just take a break now and just get your initial like, thoughts and feedback. Would this be possible to you? If it would, you know, work.

14:59 Stephanie Burns: What I mean? Definitely the I mean the this is the setup is not that hard obviously. You know we have some particular issues to work through. What are your thoughts Jamie?

15:17 Jamie McKinstry: Yeah. I mean, I'd be curious like how everyone else is already mapped. Like did we not set up our categories correctly, or

15:27 Emma Sjöström: oh,

15:27 Jamie McKinstry: Did.

15:28 Emma Sjöström: It's a super early feature, so everyone is not map. It's that some some systems provide maybe their own representation of a bank account, then it's really easy right to kind of just do that direct mapping.

15:39 Jamie McKinstry: Okay.

15:44 Emma Sjöström: but what we're also, if you find this valuable honestly, one thing that we could do that is not complex and wouldn't require a lot of mapping is to provide an entry point to this, but from the page of the account that you care about

16:04 Jamie McKinstry: Gotcha. Like a drop-down menu of like manually selecting which bank account, we want everything to apply to.

16:09 Emma Sjöström: That's also a great idea. And or we could just like at the account page in the app you could just upload from there and then it would implicitly know like you know. But that's it that the drop down is also a great idea.

16:24 Jamie McKinstry: Yeah, I think the dropdown would probably help us because our data isn't in a Format yet that like helps, you know, so it would be like us having this Excel talk and then we'd have to fill out the columns to be what we want, but they'd all be pretty similar.

16:42 Jamie McKinstry: Whereas, if I could just like, Collect in there. Like upload the data that I do have and then dropped on that'd

16:48 Emma Sjöström: Yeah.

16:50 Jamie McKinstry: be quicker, I think.

16:51 Emma Sjöström: would you prefer like One file per account, still? Or would you like a drop down like on each line item in the UI or boat or

17:03 Jamie McKinstry: um, I mean in my perfect world. I feel like palm would do, it's like thing and reading my document and automatically categorizing it like it, you know, does with our other being transactions. But Essentially because I mean I'm not a coder, I'm not techie at all but you know it's able to do that with our other being transactions and it does it pretty accurately.

17:29 Jamie McKinstry: So it feels like if I'm doing it feels more prone to error I guess.

17:34 Emma Sjöström: Fair enough. So if in a perfect world, let's say we've solved the automatic categorization of the file. But there are still no, let's say bank account, mappings. Would you? how would it still feel usable to you then even if you'd need to either on a file by file basis, select, which bank account it applies to or In line in the UI like row by row.

00:00 :

18:00 Jamie McKinstry: So, for us, we usually pay our bills out of the same bank account. So, for us, it's like not really the same because, you know, we have like a bank account for cash comes out and that's usually the one we use. So, For us specifically, yes, that would be totally fine.

18:21 Jamie McKinstry: A concern that I do have with it. Getting uploaded is potential duplication and

18:25 Emma Sjöström: Yeah.

18:26 Jamie McKinstry: I just don't understand. Well I just don't know enough about the back end to know if it can like flag those or if it's something that I would have to be able to identify when I'm looking at like the forecast.

18:44 Emma Sjöström: Gotcha.

18:44 Jamie McKinstry: And then the drill into it, which I know that's still getting worked on too. So then I'd like really feel unsure, whether or not the forecast is accurate until I saw the results and they're, you know, both like our, you know, first uploading this file weekly, there could still be invoices that we've already uploaded before so the system would need to be able to fly bed and be like, Oh, this shouldn't get uploaded because This is the exact same as last week.

19:26 Jamie McKinstry: And that would also take time on our end to make sure that it's not, you know, it's more prone to human error. Again, if I'm having to make sure that it's not

19:43 Emma Sjöström: That makes total sense. Would you say that the concern is both around like maybe

19:44 Jamie McKinstry: has not been uploaded. um,

19:46 Emma Sjöström: importing the same?

19:47 Jamie McKinstry: And then, yeah.

19:47 Emma Sjöström: Item. Hitwise And/or, is it about like having the machine learning together with this, or like

19:48 Jamie McKinstry: I mean it's because I know Palm can predict, you know, recurring payments. So if

19:50 Emma Sjöström: how it would play together? Yeah.

19:54 Jamie McKinstry: it's all like our credit card bill was, or we paid this big renewal last year at this time. I would think that Palm would learn or know that that's coming. But if I uploaded and ramp as well, And the amount might not be that or upload it from the this ram detail as well.

20:13 Jamie McKinstry: The amounts might not be the same because, you know, of upticks every year or uplifts, because the price increases so, I don't know if it know that or if like I don't know how many how big teams are for each company that uses Palm, but There have to be a lot of communication to unlike who's doing.

20:33 Jamie McKinstry: What because I know you can separately add somewhere. Like we have distributions like you can manually add in a distribution if you know that it's going to happen. But if someone does that and someone else doesn't know and then they upload it in this bill or you know, Fisheries in that example, but it just feels like there's a risk there as well.

00:00 :

20:56 Emma Sjöström: Gotcha, gotcha. I'm really interested in the duplicate thingy. Would you? If we focus on first of all that, the case of duplicate entries, let's say If I understood you correctly, like someone manually adds a forecast or like one big payment coming out and then somebody else perhaps using this feature uploads, the same one again.

21:20 Emma Sjöström: Or even manually, just like a one-off.

21:24 Jamie McKinstry: Yeah.

21:25 Emma Sjöström: What would you? What would you consider a duplicate? Like If we look at just information or data like what would be considered a duplicate for you? Would it be like the exact amount and they they're anything more?

21:39 Jamie McKinstry: The vendor name. because that's going to be in our data that we upload for ramp and then I would hope that when someone goes into ADD in the you know the known upcoming payment, they would type in a vendor name because otherwise like that's like the most important I think detail to me is who's How do I know what you're adding is real? If I can't tie it to like, Or don't have a description of what it is.

00:00 :

22:12 Emma Sjöström: Gotcha. so in this case, now if we just look back and pull up the flow or scroll up a little bit on the model and look at these mappings here that we created, Would you just so I understand? Like would you say that the Description field here should be mapped to Vendor name? For example, if you used this feature,

22:36 Jamie McKinstry: um, Yes, if we tag vendors, which I should notice, but I don't. I don't think we have a vendor field in the forecast module, right? Or do we

22:50 Emma Sjöström: Not not yet. But we do have this kind of free text field for manual ad, that you can just kind of type in whatever you need.

23:00 Jamie McKinstry: Yeah. And that would be very useful, right? Because like in our accounting system, the vendor name is a vendor name. Like I use that name to pivot and like on all our spreadsheets like it has to match the exact Field. Similarly, if it's not a name, it'll be like, you know, bigger companies

23:13 Emma Sjöström: Yeah, fair enough. That's right. It's a really good point.

23:17 Jamie McKinstry: have numbers for their vendors, that they used to like, Pivot and Use. So there's like a vendor field that we can. Populate, I think that would be, that would solve the potential duplicate issue from an upload. I'd hope it would because it'd be like, Well, you're saying you're gonna pay and double the amount are you sure?

23:47 Emma Sjöström: So if I'm just asking you, then let's say, Okay, if this magically works the way you like it to, it would be able to ideally I find the correct category. And it would be able to identify duplicates. Would it feel usable to you then? Would it be something that you?

24:11 Jamie McKinstry: Yes. Yeah.

24:12 Emma Sjöström: Yeah.

24:13 Jamie McKinstry: Yeah, I'm trying to think. Well now my brain just going really well per hour like what's possible here, but yes, it would be usable or like very useful because we then be able I'd feel like our forecasts is more accurate but again, I think that kind of goes back to, like, I I don't know too much that much about the machine learning.

24:36 Jamie McKinstry: I feel like it's probably accurate. if it's a recurring payment though it's like I guess my question is like do we even need to be uploading those like, if we pay Jen $5,000 per month, like does that

24:51 Emma Sjöström: Yeah.

24:51 Jamie McKinstry: Do really need upload it or should this should machine just know it? Which then it's like will then should I just be entering in the one-offs? that way during that process, so,

25:05 Stephanie Burns: You definitely feels like if you're if you have accurately got all of the big payments, you're expecting and then you upload this this file that it should be like, Oh, there's that big payment and match it and and maybe change update the amount to the actual amount, you know?

25:26 Jamie McKinstry: Yeah, I guess I would view this upload more of like a fine-tuning. The forecast not like Because the machine, it should I would imagine is doing it right. But this is like yes this payment is coming, but we're planning on paying it this week versus what you probably predicted or what you might have predicted.

00:00 :

25:45 Emma Sjöström: Gotcha gotcha. Yeah, that's also very valuable feedback. It makes sense.

25:54 Jamie McKinstry: Yeah, so I guess I would say it's Oh it's a nice to have to find tune and make it more accurate.

26:01 Emma Sjöström: 100%. So how often, how much, like how often do you think you would do something like this? If it worked? Perfectly for your. For you and your setup.

26:18 Jamie McKinstry: Asking for more like discogs or like as a company like my general thoughts of like how a company would use it really good.

26:24 Emma Sjöström: Oh, both would be super interesting.

26:27 Jamie McKinstry: um, Discogs, I'm not sure how. How much I would use it unless it was like super easy and automated I guess um because I don't know that it's worth Stephanie's time to put together. You know, those spreadsheet in the right format if we're just fine tuning like a week here, a week there of the expenses because we have like, cash isn't too big of a concern port? Well, it sounds terrible.

26:53 Jamie McKinstry: We don't need to be like watching our cash of that closely like a week there a week. Here isn't gonna make that big of a difference. On like an overdraft. but if I were, if we were like a bigger business, Um where there's just a lot more change like a high growth size.

27:13 Jamie McKinstry: Then I think this would be more important because we're Like the company I used to work out. We were watching our cash closely, like it was whether Are we gonna make? We need to make sure we have funds for payroll, you know, constantly because we're not trying to float cash and we're playing with the when we're actually going to pay a vendor and when we get the invoice and figuring out when we're gonna like how we make it happen.

27:42 Jamie McKinstry: So I could see this being very helpful because I'm like, you know, if I upload that data then I could play around with the date that I'm gonna pay it like, you know, AWS was

27:48 Emma Sjöström: Yeah.

27:49 Jamie McKinstry: our biggest vendor. And that's the one that we try to push as much as possible. And so if I could change the date and see how that impacts the forecast, If you don't.

28:01 Emma Sjöström: Makes a lot of sense. Thank you. I know we're running out of time, but do you have any additional thoughts or

28:10 Jamie McKinstry: No, I love it. I hope I'm not coming off that this isn't it? Like this is very great.

28:14 Emma Sjöström: No, no. It's I'm aware. It's early version of this feature and we try to do it deliberately because we want to make sure we kind of iterate in the right direction. So I love like old types of feedback because it helps me in my job a lot to make sure we're focusing on the right things, right? So, it's great.

00:00 :

28:35 Jamie McKinstry: Okay.

28:36 Emma Sjöström: Yes.

28:38 Jamie McKinstry: But yeah, no other questions on on my end.

28:40 Emma Sjöström: Yeah. Stephanie, any other thought?

28:45 Stephanie Burns: And not the moment. I mean I think it's definitely something that could have a lot of value once it's a little more fine-tuned

28:56 Emma Sjöström: Yes. Amazing. My my key takeaways still will be like the automatic categorization and duplicates. That would make it feel usable for you guys. Super, thank you so much for your time. This was really, really helpful.

29:16 Jamie McKinstry: Of course, thanks for walking us through it and showing us what's to come. It's

29:20 Emma Sjöström: Of course, I hope to see you soon again and show you something else.

29:21 Jamie McKinstry: exciting. Yeah, sounds great.

29:28 Stephanie Burns: Okay, thanks.

29:29 Jennifer Pearson: Thanks.

29:29 Jamie McKinstry: Thanks. Bye.

29:31 Emma Sjöström: All right.