Main content

2023 - the year in AI Use Cases

Jon Reed Profile picture for user jreed December 22, 2023
Generative AI forces an AI readiness conversation; AI is a gut check on automation and analytics maturity. Get your data platform right, and getting AI right gets a lot easier. With that in mind, here's the high points from ten AI and automation projects we documented this year.


For better and sometimes for worse, generative AI-dominated enterprise conversations in 2023. AI fatigue became a real thing, especially from the keynote stage - but the impact of AI is undeniably real also. The urgency to make sense of AI took the diginomica team across ethics, data governance, and into the perils of disinformation on democracy itself. Even the OpenAI drama required dissection.

But in an enterprise context, this ultimately comes down to results. As such, generative AI is only one form of AI - and AI, no matter how much sex appeal marketers might infuse, should be subject to the same project discipline - and KPIs - as any other technology.

I've been critical of the flaws in generative AI - Large Language Models in particular. I've also blown a gasket or two on why "responsible AI" posturing by OpenAI and other "big AI" vendors is farcical. What I'm really after is precision: enterprise customers should know what AI's limits and potentials are, without the braggadocio and exaggerations. These tools are powerful enough.

With live enterprise projects as the criteria, generative AI isn't there yet - we'll have to wait till 2024 for that. But we did review several gen AI projects in the planning stages, and I recently documented one live gen AI use case. AI projects provoke an essential data conversation, whether it's robotics on the shop floor, or customer-facing service bots. Here's some of the best conversations we had this year.

IFS Connect 2023 - why talk about generative AI when robotics and connected manufacturing are ready for prime time?

You can see the heavy payloads this robot can handle. As Ivkovic says in an IFS video on the project: 'We expect to save over $1.5 million per year. In addition, every employee affected will be retasked into higher skilled positions, greatly helping us with our labor shortage.'

Why? I lost count of the number of vendors that overlooked the chance to showcase mature AI projects in 2023 - opting for generative AI pronouncements instead. But at IFS Connect, I spoke with a couple of well-established robotics projects that are delivering eye-opening results, while easing labor shortfalls.

CCE 2023 - the great AI debate shifts from content creation governance?

On the enterprise side, audience questions about data governance were a surprisingly common thread. Given generative AI's current sex appeal, I was struck by the contrast with the practical-but-essential data questions these CXOs emphasized. Example: "Who owns data governance?" How's that for a concise can of worms? A detailed guide to "best practices" in AI data governance - if such a thing even exists - was beyond the scope of this event. But field lessons were shared: bear down on the current state of your data governance, prior to ambitious AI projects.

Why? I expected the CXOs at Constellation Connected Enterprise to jump right into a generative AI conversation. They did - but like many events I attended this year, the conversation soon shifted to the nitty gritty pain points of data governance, and readiness for "responsible" AI.

Can AI solve the messy data predicament?

The solution was to build a Knowledge Graph that can identify 60 different data formats, including JSON, AVRO, Parquet, XML, and RDP, including semi-structured data via its Data Parsing of WebLogs/Log File. – crawling through the universe of data and using AI techniques to create an Enterprise Knowledge Graph that encodes relationships between files and depicts all of the relationships in the entire corpus of data.

Why? In one of his best posts from a substantial year, Neil tackles the problem of messy data from a practitioner's view - including lessons from an Enterprise Knowledge Graph project. The extent to which AI can make data preparation and governance easier is one of the questions to track in 2024/2025.

Deutsche Bank doubles down on generative AI after laying foundations with Google Cloud

2022 ended with Deutsche Bank deciding to double down on generative AI. However, whilst the management board was excited about its possibilities, it still viewed it as a ‘technology thing’, according to Perez. However, this all changed with the public release of ChatGPT. He added:

'We went to the board on the 7th December…on the 30th of November ChatGPT happens. Ever since then, the big thing that has happened is that for the first time business leaders want to transform their business. We were fortunate enough to do a lot of the groundwork, prep and priming of the pump to seize the moment. But the biggest thing that has happened, it's not the technology, it's that it has transcended the technology community.'

Why? I included Derek's quote from Deutsche Bank in detail, because it illustrates the potent effect consumer adoption of ChatGPT has had on the enterprise. Again we see that emphasis on AI readiness. But data preparation is of limited use until you get business buy-in. With about 25 generative AI use cases on the fast track at Deutsche Bank, this is another story to watch.

Workday Rising EMEA - Mondelez International tests out generative AI for employee self-service

90% of employees now self-serve for tasks where it's offered, but can still face issues when attempting something unfamiliar. So the company has been testing whether generative AI can power a more conversational interface for these less frequent interactions.

Why? If we want to see successful generative AI projects in 2024, we'll need smart design that accounts for gen AI's current limitations, with humans in the loop for adult supervision where necessary. But one clear strength of generative AI is the impressive scope of its potential use cases. And, unlike the clunky bots of the past, the prospect of gen AI bots is appealing to users - a topic Phil takes up in this Workday Rising EMEA case study. Also see my Workday Rising story, Want to be ready for AI and automation? Get your culture right. Workday's financial customers weigh in.

Can enterprise LLMs achieve results without hallucinating? How LOOP Insurance is changing customer service with a gen AI bot

Imagine my surprise when yet another PR campaign about 'our gen AI doesn't hallucinate' turned into something different: my first published use case on a live gen AI project. Another curve ball: this isn't an internally-facing bot or digital assistant. This is a customer-facing chat bot, operating in a regulated industry (insurance).

Why? In my first live use gen AI use case, I was able to press LOOP Insurance: "Does Quiq's AI bot hallucinate"? The answer was no - though I did find some limitations in the bot, which I explain. But the key points are this: LOOP's live bot is getting results, including: customer self-service rate increased by three times, to more than 50% automated resolution, and a 55% decrease in email tickets. This type of project demonstrates you can achieve a gen AI success in a focused area, without having to overhaul your entire data infrastructure. Quiq's CEO also shared details on how their architecture avoids the level of hallucinations we see in consumer tech bots like ChatGPT. A top priority in 2023 is clearly to have frank, "lessons learned" discussions with other live gen AI project leads.

AI is ‘letting lawyers lawyer’ at Clyde & Co

Something that would take a person about half an hour to do, assuming a medium complexity personal injury claim, is now down to five minutes at Clyde & Co. The company stresses that while it’s a machine pulling out the data and making the initial valuation, a human specialist still checks its findings. But, says Rourke, his colleagues really appreciate the time saved by the system pulling out data from what can be thousands of pages of notes. 'This enables us to gather data at a much more granular level - to capture it and then analyze it on a much more macro level.'

Why? Another live LLM use case on diginomica - this one from Gary. This one is indicative of two aspects of gen AI for the enterprise: 1. human in the loop where needed, and 2. industry-specific LLMs, one way that we may see more accurate/useful results (Clyde & Co is using a legal-specific LLM from Luminance, which was trained on 150 million+ "verified legal documents."

No more frantic timestamp hunting! How Zappos overhauled planning with Planful

Mann stressed at this point that cost savings or reducing headcount was not the goal. Rather, the time saved was able to be repurposed into “more value add” so that analysts could spend more time with partners, understand their business better, and drive those partnerships forward. Zappos has gone from needing a full business day every quarter for top-down planning, with manual inputs and unreliable data, to an hour for a consolidated quick view.

Why? Alex documented a planning project that illustrates how companies will be able to shift from better use of data to AI. One crucial issue: are you trying to empower/enable your people "add more business value," or are look going solely for headcount reductions? If the answer is only - or primarily - the latter, I  don't like your long term business prospects. For more on the predictive side of AI, check out my 2022 Planful use case, How ProMach powers its acquisition strategy with cloud finance.

UiPath Forward - how Kelly Services is preparing for the future of work by taking the pain out of recruitment

This is about more than reducing paperwork. Calling back to the demonstrations during UiPath Forward's keynote, the opportunity to provide recommendations for cost comparisons based on different contracts, insights on applicant availability and training could take recruitment to a new level. The combination of automation and AI in this context is a strong example of not taking away jobs from anyone - but getting the right people into the right roles, faster.

Why? Another example of how automation lays the groundwork for AI. And: mature automation is more than just eliminating paper processes. As Alex rightly notes, the principles of automation matter. If you want to automate in a way that puts your talent in higher impact roles, you must design for that outcome.

How AI is helping disadvantaged UK young people get into top universities

The Dataiku team then ran its predictive modelling software over this data to understand how to better support strivers applying for Brilliant Club support. It also carried out a technology transfer where it helped the Club extend its in-house data analytical capabilities... This is delivered as a visual data tool that can easily be used by non-data scientist managers at the Club, Ballaera says. They run the tool regularly to see what pupils may be struggling, and immediately suggest aid and interventions for those who are red flagged, she adds.

Why? If AI is out of reach of non-profits, then collectively our AI pursuits have failed. Fortunately, a series of non-profit use cases, such as this one by Gary, imply that there are ways to make AI feasible, even with non-profit budget/data limitations. However, great care will be needed to make sure AI algorithms don't misrepresent risk factors or improperly screen out those in need. That said, any AI project brings that type of risk management.

A grey colored placeholder image