Skip to main content

Accelerating the Monthly Financial Closing With Genie

Summary

  • This video walks through how finance teams can replace manual VLOOKUP-based workflows with Genie-powered reconciliation and variance analysis, demonstrated through a mid-size retailer scenario on the Databricks Data and AI platform.
  • A finance controller uses an AI/BI dashboard to spot reconciliation gaps between CRM invoices and accounting numbers, then queries Genie in natural language to investigate root causes and draft variance commentary against budget.
  • Databricks Apps serve as the governed finance team interface for applying corrections, updating numbers, and generating the final reporting deck within a single unified workspace without requiring data exports or spreadsheet workarounds.

Accelerating the Monthly Financial Closing With Genie

Watch: Accelerating the Monthly Financial Closing With Genie
Closing the books every month is still slow and manual for many finance teams. In this video, we will walk through a monthly financial closing for a mid‑size retailer and show how we tackle two of the most painful steps with Databricks One: reconciling gaps between CRM invoices and accounting numbers and running variance analysis to explain differences between actual results and budget. We will follow a finance controller using an AI/BI dashboard to spot issues, Genie in natural language to investigate root causes and draft variance commentary and a Databricks app to apply the corrections safely, update the numbers and generate the final reporting deck. Attendees will see how business users can make financial closing faster, simplify investigations and produce more reliable reports with less manual effort — all from a single, unified workspace.
Talk By: Christophe Chieu, Solutions Architect, Databricks ; Francis Laurens, Solutions Architect, Databricks ;

Chapters

FAQs

How can Genie accelerate the monthly financial closing process?

Genie allows finance controllers to query financial data in natural language, identifying reconciliation mismatches between CRM invoices and accounting numbers and explaining variance drivers against budget. This replaces the manual process of exporting to Excel and running VLOOKUPs across multiple disconnected systems.

What steps of the monthly financial closing does this video address?

The video focuses on two of the most painful closing steps: reconciliation, which involves matching invoices across CRM and accounting systems, and variance analysis, which explains differences between actual results and budget. It also touches on adjustments, accruals, and producing the final CFO reporting deck.

Is Databricks a replacement for accounting software?

Databricks is not an accounting software and should not replace a consolidation tool or accounting system, as the presenters explicitly note in this video. Financial and accounting logic can be replicated on Databricks to accelerate investigation, reconciliation, and reporting workflows alongside existing systems.

What role do Databricks Apps play in the financial closing workflow shown in this video?

Databricks Apps serve as the governed interface through which finance users apply corrections to the data, update numbers, and generate the final reporting deck. This keeps all actions auditable and contained within the unified Databricks workspace rather than scattered across email threads and spreadsheets.

Full transcript

[00:11] Good. Hi everyone. Thank you for being here today. Uh I hope you're having so far a very good time at our data summit and we're very happy to have you here today and uh I'm Christoff Chu. I'm a social
[00:23] architect at data bricks. Uh my background is like I come from data science and I worked for like several years in insurance and financial services and I had the the the pleasure to work on the monthly closing and for a
[00:36] fun fact or so when I left my former company for data bricks I thought I would be done with the monthly closing and I turn out my wife is a controller so actually I still hear monthly closing at home and that's how I came up with
[00:48] this topic today. actually even leaving the industry I still in touch with monthly closing >> and hi nice to meet you all so my name is Francis Lawrence actually I'm a
[01:00] solution architect working in Paris France as you can guess it with my accent actually and ask so I was working in some reconciliation product in finance prior uh in my prior experience
[01:12] and also architect in many startups and French unicorns And yeah and today we will talk how you your company your finance department can accelerate the monthly causing with
[01:25] genie and just before we continue just among the audience can you raise your hand if you are directly or have been directly involved in the monthly closing. Yes. Oh
[01:37] wow. Okay.
[01:51] still I'm just going to set say a few words about what the monthly closing is what is the scope what are the changes. Then uh Francis here will show you a cool demo uh in which we will see how
[02:04] you can access monthly closing with Jenny. After that we will do uh a review of what we have seen during the demo and we will provide further details about the components that we used and finally
[02:16] we will wrap up the session and have some time for Q&A before we continue. So I just would like to provide some clarity about some key details uh because the session is only 40 minutes. So we will skip some topic
[02:29] such as we're not going to talk about how you ingest data from source systems uh uh and or accounting software and data bricks. We're not going also to talk about how you create data model
[02:42] behind um the financial statements as income statement income statement cash flow statement balance sheet. Yeah. And finally, datab bricks is not an
[02:54] accounting software. Just to make clear, you can create and replicate some financial and accounting logic and data bricks. But I would recommend to be very careful to build your entire
[03:08] consolidation tool or accounting software and data bricks. Not that you can do it or it's not recommended. It's just like it takes lots of effort and time. So you may want to weight the the
[03:21] the reward versus the cost of effort and time. Now on that let's start. So what is the monthly closing? and very few words. Uh
[03:33] at the beginning of the next month, uh your financial team, your accounting, they must make sure that all transactions during the months were recorded, that all balances made sense and that the financial
[03:46] statements with the appropriate documents were generated for the your CFO and uh the the leadership. And the task leading to the financial statement can be broken into four bucket of works.
[04:00] We have the reconciliation. Uh we're checking that sales, revenues, uh subleddgers, um payroll systems match with what we have uh posted in our
[04:12] general ledger. Checking so that the bank statement match the cash. Checking also that if uh inter company balances cancel out. We have also acru and
[04:24] adjustments. So booking expenses that have happened but are not invoiced yet, calculating amortization. Um updating your reserve. We also have valuation uh
[04:37] in compliance. So valu valuating uh your assets and inventory and u handling handling the the leases and tax. And finally the last mile um
[04:49] variance analysis and reporting where we dig into the difference between the numbers and the budget and we consolidate the numbers into the finance
[05:01] statements with the appropriate comments for your CFO and the leadership to review. Now all of this is very essential for the financial relies. They are the foundation of the financial
[05:15] health of all companies including yours. But uh the process might have have some problem. It is very long. It drags on several days. It is very intense. It
[05:28] puts lots lots of um pressure on the teams and people involved and also many tasks are very manual and this is what we're going to address today.
[05:41] Now why does the monthly coding take so long? Well, you have many reasons uh but two all of them come come up systematically reconciliation and balance analysis.
[05:55] When it come to reconciliation, we're talking about matching thousands of invoices, records across different channels, different uh systems to make sure that the numbers align.
[06:09] And when you come to varance analysis, we're talking about um explaining the difference between the numbers that you calculated and the budget. And for that you need to dig at a very fine granular
[06:22] level uh to find the the the drivers behind the numbers. And you have for for those two many tools. You have your Excel tool, you have BI tool, but you also different communication channels.
[06:34] Uh we have teams, Slack, sometimes also emails where you get information that contains information that you must take into consideration. um if your company have multiple entities, some of them
[06:46] also might run their own system and those information must also be taken into consideration. So in the end it's it's it's a bundle of of many tools that that can that that that create uh
[06:59] inconsistent logic and that can cause um missing information and drag the process and make it very heavy uh very painful and very stressful. And so now let's
[07:12] just take a quick look uh at the the workflow when it comes to reconciliation and variance analysis. So here is our traditional workflow. So that's somehow what most of the finance teams live
[07:25] today. So you when it comes to reconciliation usually you you take your your data out of the the source system. It can be sales, it can be payroll, it can be warehouse and you take also your data out of your ERP or your accounting
[07:39] software where your general ledger lives and then you compare then you put everything in usually an Excel spreadsheet uh yours and also one that
[07:51] you show your colleagues and then you you will take another number let's say uh the the total revenue that you posted in your general ledger and the total sales revenue coming from your sales system for example and you will compare
[08:04] and usually there's a difference and that's where that's that's when the the fun starts. So then you have to cross reference thousands of thousands of invoices and records to see where it doesn't match. And then you do that
[08:18] usually with a VLOOKUP. And then after that you have to work out when you found those you have to work out the root uh cause behind the difference which can be sometimes an invoice a missing invoice.
[08:31] It can be also a credit note that that was not uh taken into consideration. It can be an exchange rate trip. You have many many reasons and you have to repeat repeat repeat until uh the difference
[08:45] between the two systems uh equals zero and you deal with all the the subsystems. Then so it's already quite heavy. Then you do then once you did everything you're good to go the reconession you can start to calculate uh the the KPIs
[09:00] such as operating profit gross margin working capital and so on and so on and you compare that with your budget and you if you see some difference then you need to dig into the difference and see
[09:13] why you have a difference and then those difference you will add them as you add the the reason as a comment and to the into the final statement that will be generated and that you will share with
[09:26] your CFO and your leadership. So as you can read here it's already some it's already very heavy and yeah we spend a lot of time uh doing
[09:38] manual research, manual investigation, manual corrections and with manual uh task co comes the risk. So by doing things more manually you increase the
[09:50] chance of making more risk uh which can affect the financial statement and because it's very time consuming also the finance team usually have then less time to spend on analysis. So I think
[10:03] this is why we're here today. What if there's is there a way to to make it faster, simpler and more reliable and I think that's what my colleague Francis
[10:15] can answer. Thank you Kristoff. Um so um actually just prior that we start I have just one question who is actually doing his reconciliation through just
[10:28] Excel export files. No shame actually I was doing the same in in the past but and I just wanted also I think it's really important to to stress on that. I mean the pain that
[10:41] Kristoff mentioned are real. I mean you need to export in a timely manner. Sometime you also forget to have the sources and even for vookup sometime it can be something that is not really easy to do. So definitely this makes the
[10:57] this process very less stressful. So the idea is that we're going to see how we can make it way more smooth and easy. Um so for this scenario um I'm going to
[11:10] be u the c um a person working directly in a midsize company was like around 8 million monthly revenue and we are on day three of the January close. So there
[11:24] is two day left and let's see how things goes. the next slide.
[11:38] So I think that you now know this interface very well. It was presented a lot of during the keynote. But this is actually genuine. So genuine is um pro in which you can have many kind of assets like
[11:50] potentially dashboards, you can also have genaces or you can also have app Spoiler alert, we're going to go through these three different items through this demo. The
[12:03] idea is at first compared to Excel files or other things. You have everything within one single platform and ch the cake all your data is protected and
[12:15] governed in one place. So you no longer have to deal with many emails or SFTP to exchange this data. So everything is located into one place. And we're going to check and go to our first dashboard.
[12:29] So this dashboard contain different tabs. So there is the first one with reconciliation cost process, another one regarding P&L revenue KPIs, a third one with balance
[12:42] sheet and working capital and a third one a fourth one regarding budget. So what we can see sadly unfortunately there is 90 conc and we have two days left.
[12:55] Secondly so we can still have also because all the information gathered into one place we can have some views about the P&L. So as you can see the revenue is pretty is pretty good margin
[13:07] is a bit under and we have for example the balance sheet as now so the R is pretty high but fortunately the cash position is good. So I think we can say that we are fine
[13:19] and regarding budget forecast I think that the forecast revenue and business revenue as we can see is are pretty aligned so nothing to declare. So we can now focus on these 93
[13:32] unreconciled items and see what we can do and this is where we can ask Jenny to help us. So the idea is that we can direct directly go to a gen space and we can
[13:48] directly ask a question to I have this amount of it times how can you help me solve it. So no no longer requirement about V lookups no longer
[14:00] requirement about export we directly ask the question and we have directly the results. So we can see that we do have uh two kind of root causes
[14:14] for missing references. First one is variance issues with uh 75 transaction which are not matching references. It's not matching but with different account amounts and a second
[14:27] one which is a little bit uh worse which means that they are not recording the system and the idea is through that gen space with the genie agent it automatically give us
[14:41] immediate action that we can do. So it gave me a plan to figure out exactly what I should do and as it was also we do had for um in the systems it
[14:54] definitely take some ownership for some people to be able to fix it. So no longer vookups, no longer export. We directly have uh uh actions that can be
[15:06] uh taken uh and do yeah to make this to make to fix this. The idea is that now it's time to fix it because okay we know what is missing but
[15:19] there is still things to do and this is why we have created a databicks apps to make it. So instead of saying send me an email then I'm going to run a process for you. Oh no it is not the right
[15:31] format because you send me an Excel. I was expecting to have a Google spreadsheet or I don't know why. We created a datab bricks app for this. And the idea is that um we can ask directly some people that
[15:44] have overhip to directly send potentially corrections about the issues that was puted previously. And so the idea is that automatically we can check the format. So we no longer have this
[15:57] kind of exchanges like it is not the right format. It is it is okay. And we can also directly um um allow people to trigger the reconciliation job. So okay
[16:10] what's the benefit? You are no longer waiting for so some people that can have a lunch break during this two t days to wait. So we can directly launch a reconciliation job and The good thing is
[16:23] that for also it is error prone in that case of you prevent user from doing actions that could be too wide and we can see the result afterwards.
[16:36] It's working. So we are happy due to the fact that we have created this project that makes really easier to spot what is missing and how we can fix it. Then we can we can see we can spend our
[16:50] time in analyzing the P&L and and also the revenue and we can also for for this use Jenny. So the idea is that I would like to
[17:04] write an analysis. I don't want to just send a screenshot or to spend time to run an email and to make it clean. The idea is that I can just ask Jen to analyze what is happening and to
[17:16] generate me a report and this is what we can see directly within this chart and it it tells you for uh so basically what what Jenny says is that okay we
[17:28] have uh better revenue we are uh we don't have a good KPI with gross margin but actually we have a positive operating outcome so this is compensated
[17:40] and this is something uh that is favorable for revenue variants. So that's fine and ch the cake we can do directly a PDF export and share and share them to to the people.
[17:55] So yeah, so this is what we have seen during this demo and I see that could be really uh impactful and the good thing also with what we've seen recently potentially with other tool like Omni
[18:08] that you could also share with uh other folks in the company to have some comments and to share all this information together. So what did we see?
[18:20] I just go to the next slide. I think it is broken. Okay. So what was the solution that we saw today? We saw first AIBI dashboard
[18:34] in order to detect. So I think for this kind of project it is good just not to do fully everything in natural language but just to have like a dashboard just not to spend much
[18:46] time to ask for the question but just to have like a first overview about what happened and we we we were able to highlight efficiently and quickly gaps. Then we used DNA for in for
[18:58] investigations and we were able to to make some root cause analysis and to draft a variance commentary later. And then we were a
[19:10] able to resolve and report things thanks to a datab bricks apps which really uh help uh user to interact with the platform and to be uh not to be also
[19:22] potentially lost because there there would be many KNS or things to to uh potentially could be possible. So all the step can be very well documented and we can really uh um fix directly things
[19:37] in production and something that is really important all those things were possible. Why? Because we had all the information in one place in the lakehouse and this
[19:49] is all the benefits that because all those services were powered by unity catalog. So your data is governed. You can control access and you are able with natural languages to interact with this.
[20:02] So this is really uh what we've seen with this solution.
[20:14] And so Kyaways I think it's time to uh potentially recap what was saying. So we went from days of Excel spreadsheet analysis to a streamlined workflow in one platform in one place for all the users.
[20:28] It is faster but also more efficient. So the idea is that there is like no more exporting, privoting, copy pasting and also by the way exporting data and send
[20:40] them through email about finance data is not really secured.
[20:53] not mentioned in the slide but definitely for security that is better that it is stayed within the platform. And lastly, it is more accurate because you um you were able to uh have those
[21:07] KPI that were created with metrics that you can trust and you can also trace also with Unity catalog from the sources to the final report. So this is all the benefit that you have
[21:21] and yeah so we saw there for the reconciliation but definitely this is a replicable pattern that you can take home for the same kind of workflows. So
[21:36] first you would like to detect things, investigate and then to be able to reconcile in potentially uh or also the topic like interact
[21:48] company AP uh and do variance analysis. It could be also outside of finance scope potentially. But definitely this is a good way to be able to work uh
[22:02] efficiently and not being like in a pure data chaos with many file or sources that are spread across your drive and so on.
[22:19] So uh just to add because I think as a joke but actually quite true we are I think the 18th of June. So it means like probably in two we have two three solid weeks before your next end of the month closing. So if it's something that resonates for you at the end of session
[22:31] feel free to share with your CFO to get it approved and then have your data team work on implementation so that you know you too can accelerate the monthly closing with Jenny. Thank you.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.