Griffin Chase

If your introduction includes the school you went to, I’m not interested.

I’ve been critical of OpenAI lately, but for the past three weeks my family has been dealing with a health issue with my dad, and a ChatGPT shared project with live document syncing has been essential to organizing and understanding everything happening. Me, my four siblings, my mom, and my dad have faced an onslaught of information from various doctors and nurses, which we’ve captured in hundreds of text messages and documents and scans and you name it. ChatGPT has helped us collect this information in a single place, make sense of it, and interrogate it to make the most informed decisions possible. Also, credit where due: Claude played an important role as well, by ingesting iMessages and synthesizing summarizes from them to upload to ChatGPT, as well as by extracting text from a bunch of HEIC document scans. I think those of us, like me, excited at AI’s potential get frustrated when we can see issues so clearly, like ChatGPT’s bad design skills, and Claude’s increasing instability and confusing usage consumption. But at times like this I’m reminded of how incredible this technology already is, letting me and my family make sense and act on hundreds of pieces of information, empowering us in the face of a disjointed and fragmented healthcare system.

I’ve seen this floating around, most people ick-ing at this as Claude derangement syndrome. But guess what! This CEO probably just wants to get great outcomes fast, and I’d bet he’s thrilled to hear something other than “it’s not a priority” or “let me revise this and we’ll come back next week” or “thanks for the input but we have experts on the team.” Maybe your team isn’t producing work up to their standards. Maybe your company is slumping along at single digit or negative growth and the board has said it’s up or out. Maybe he’s trying to drive up AI adoption. Maybe the CEO job sucks and AI makes it more fun. Maybe AI raises the floor and the ceiling on everything. I will say I’d rather a CEO who cares about driving up quality of marketing copy than one that’s checked out. I’d rather be lead by a Claude Boi than an exec team that forms a “AI innovation committee” and is just starting their limited 2 quarter trial of copilot. I’d rather just have the answer handy “yep, we did deep research on this via ChatGPT and figured out how to automate brand assets consistently thought nano banana, do you have any feedback?” SLOP ENGINES BEWARE for sure, but I generally think folks don’t have enough insight into how corporate anxieties and power dynamics make the exec job a lonely one, and I’m not surprised CEOs are reaching for tools that shorten feedback cycles, openly brainstorm, and produces work (not more meetings.) Look. Do I think there are better styles of leadership? For sure. Do I think that means some teams don’t need a major wake-up call? No. Ok have fun. Tell Claude hi.

There is so much low hanging fruit for AI in accounting and finance functions. Here are a few ideas for anyone with accounting domain expertise and the engineering skills to build: 1. Master Data Bleach Master data is a train wreck in most businesses; product data, customer data, inventory data, etc. The basics (i.e. prices, product codes, etc) are normally correct (because they have to be to get cash in and out), but the attributes that drive quality reporting are often inconsistent. This is often the limiting constraint on reporting (rather than reporting capability itself). And can be tens of thousands of records across hundreds of fields. So not easy to keep live. A bot that crawls all company master data looking for inconsistencies in naming, taxonomy, coding etc. And then produces a list of proposed changes, which are then reviewed by the human to accept or reject those changes. The bot then goes and makes the changes. It does this daily / weekly learning each time and improving it's accept %. Probably doesn't save any heads in master data management in the short term, as there will be an increased review workload. But this would unlock huge reporting capability and time savings downstream where the data is being worked many times at different touch points in the organization. Solve the problem at source. 2. Accounts Payable Support AP teams have to maintain contact centers to answer - what is normally - some variation of the question 'when will I get paid?' Quite a bit of the AP function has been automated through RPA bots over the last 15 years or so. (I.e. supplier statement reconciliations). But handling supplier queries and inbound has been mostly untouched. And hearing how fast voice powered AI is coming for customer call centers (first and second level support) - feels like a small leap to stretch this tech across to AP functions. I expect it will be slower on the accounts receivable team (B2B at least), because businesses will be more reluctant to hit their customers with AI generated outbound. (And the cost of a mistake is higher - i.e. don't get paid on time and it blows the cashflow up) 3. Automated Preparation of Annual Reports Preparing 10-K/Qs is a pain in the ass. There are a number of inputs. There are the numbers (which are jsut mapped from the GL). Then there is coordinating a good quality commentary (MD&A), making sure your risk factors are up to date. Control disclosure, regulatory changes, Etc. Different jurisdictions for multi-nationals. And the formatting is a pain too. There is a surprisingly large number of people involved. It's inefficient, and hard to bring together. So this would be cool ... A dedicated tool that takes previous annual reports, control reports, audit report, board packs, local regulation etc and prepares a draft of the annual report. It would asking any questions of management it needs to (as if it were the controller preparing it) and helps manage version control, and workflow to completion. It will still need verifying, but getting to a good quality first draft is most of the work - and crucially is the bit that takes time. 4. Contract Review & Payment Terms Optimization It's hard to track inconsistencies in terms and conditions across thousands of commercial contracts. So many times I've found $$$ left on the table thorugh inconsistencies. It's left me wondering "how many other contracts do we have where we've lost 15 days like this". I presume it would be easy to create an application to: a) review new contracts before approval to find inconsistencies with previous contracts / house terms b) review all contracts to synthesize where your procurement / sales teams are leaving money on the table I know what you are thinking... Chat GPT / Claude, etc can do this already. You just have to know how to prompt it, etc. Or there is already some software tool that does this. That misses the point. Outside of tech circles, no-one in accounting and finance teams are really thinking about AI. They are too busy doing what they've always done. They need the solution put into their hands that works perfectly alongside what they already have. It just needs to take some existing human work, and do it faster and better. Low friction roll out. These are also (by definition) the least efficient teams where all the people / cost savings are - i.e. it's where the money is. Plus there is the issue of needing a safe and compliant private environment on which to manage the data. This will be crucial for getting CFOs on board (who need to get theit baord and auditors comfortable with it.) So a static application / environment that is fool proof for the user (and in time can become trusted) is the key. Any other ideas? Forget the long term ... I know there are way more transformational ideas than the above. I'm interested here specifically on what could be at us VERY quickly... anything else?