A case study in why organizations should both incentivized their employees to explore AI uses that help them & have a Lab of dedicated AI builders
Here, Cornell's finance & AI teams created a /treasury Claude skill that recovered $100k in back payments. https://innovationhub.ai.cornell.edu/articles/how-cornell-recovered-100000-in-unidentified-payments-with-ai/
Positive users praise Cornell's custom Claude pipeline for recovering $100K via time savings and in-house data tools, while negative users argue it reflects prior finance oversights rather than genuine AI gains.
No Digg Deeper questions have been answered for this story yet.
#184low hanging fruit everywhere with those with the arms to pick them
A case study in why organizations should both incentivized their employees to explore AI uses that help them & have a Lab of dedicated AI builders
Here, Cornell's finance & AI teams created a /treasury Claude skill that recovered $100k in back payments. https://innovationhub.ai.cornell.edu/articles/how-cornell-recovered-100000-in-unidentified-payments-with-ai/
More on Leadership, Lab, and Crowd: https://www.oneusefulthing.org/p/making-ai-work-leadership-lab-and
low hanging fruit everywhere with those with the arms to pick them
@emollick there’s like a dozen startups trying to solve this exact problem right now
there’s also something to say that they were able to do this all on their own, quite easily it seems, with just a simple skill add on for claude code
@emollick might be wrong but the $100k recovery sounds brittle until the treasury team can update the skill themselves did Cornell address what happens when the payment data structure changes and the lab isn't around?
@emollick the AI didn't find new revenue. it found $100k the finance team had already lost track of. that's not an AI story. that's a performance review with extra steps.
@emollick 🔥🔥🔥🔥
@emollick Combining domain expertise from finance teams with specialized AI builders creates practical tools that address real operational challenges. This collaborative model often leads to measurable improvements in efficiency and financial outcomes.
@emollick @alexolegimas I like the part where they say this takes a long time because the bottleneck is now the staff response time
@emollick This is a classic @emollick paper... it has changed a lot of how I work with businesses and business leaders when considering AI within their businesses.
Thank you, sir. It's changed my persepective and helped thousands of others, I'm sure!
@emollick $100k recovered from a custom skill is the kind of ROI that should be in every CFO's deck this quarter. The dual approach is the real lesson though, bottoms-up exploration finds the use cases nobody would've thought to build top-down
@emollick The quiet lesson here: the finance team always knew the $100k was lost somewhere. What they lacked was the ability to build the tool without waiting six months for IT. AI didn't find the money, it collapsed the distance between the person with the problem and the fix.
@emollick Frameworks are just org chart cosplay. Leadership budgets it, labs demo it, crews ignore it until it overwrites a Friday spreadsheet. The bottleneck is whoever deletes the legacy workflow instead of defending it.
@emollick Inhouse teams building on company data consistently beat one size fits all software tools.
@emollick @emollick The dual-track matters: employee exploration finds the real use cases, but the Lab structure is what gives those wins a path to scale. The $100k wasn't just found—it became a replicable workflow. Wondering if Cornell plans to open-source the skill architecture.
@emollick The $100k headline undersells it. The real unlock is that the finance team built the skill themselves instead of waiting on a central AI group to hand one down. Domain experts shipping their own tools beats a top-down rollout every time.
@emollick Hours of investigation compressed to minutes. Humans stay in the verification loop. $100K from one batch. That's the deployment pattern that works.
The $100k is the headline. The real unlock is the structure underneath.
People who know where the pain is (finance team) + people who know how to build (AI lab) in the same room.
Most orgs have one or the other. That combination is what turns a one-off win into something that actually scales.
