There's a lot of money lying on the table such that processes and systems are already going to be running on the border of human comprehension... can I see the signal in this noise?
Meanwhile projects that do something new tend to be successful and profitable, but mere reimplementation digitally tends to result in BS being added to the process as noise until the signal to noise ratio has dropped once again to what was, after all, acceptable before automation. "Well now you got all the time in the world to really stick it to the XYZ department so here's the new requirements" then they fire back causing lower productivity that if nothing were tried to begin with.
You have to realize automation is old, and our ancestors were ignorant not stupid. If you could increase the odds of winning a game by 5% using a million bucks of IBM 7094 mainframes and a million LoC in COBOL then gramps would have done it and made a profit to boot and maybe you can make a couple dimes by modernizing the hardware, but I assure you the algorithms and constraints are baked into the business logic cake and the win/loss percentage won't change.
One way to automate that does succeed is fixing something mgmt didn't realize is broken. I never knew there was a way to do that. I never knew we as a company blew X thousand labor-hours per week on hidden task Y.
IF football never tried statistics and record keeping, then a new system probably using contemporary tablets could really rock. But this isn't the case.
A good, although semi-controversial example, is moving from verbally asking for records over the radio to mobile data terminals to whatever cops use now (probably "cop space app" on their phones or something) has many effects, some even good, but catching criminals isn't one. Fundamentally you've changed who spends five minutes dorking around looking for warrants and dramatically changed how they spend those five minutes, but you haven't really changed anything in terms of end results, although intermediate metrics can be gamed (like dispatcher labor hour minimization).
Tablets make very poor band-aids.
