WEBVTT

00:00:00.017 --> 00:00:03.174
Hey, everybody, and welcome back to the Operational ITAM podcast.

00:00:03.488 --> 00:00:06.704
I'm Bill Van Nort, and I hope you brought the flashlight I told you to bring,

00:00:06.884 --> 00:00:09.897
because today we're going hunting in the dark. Shadow IT.

00:00:10.605 --> 00:00:14.488
The tools your business bought without telling you. The servers nobody owns.

00:00:14.755 --> 00:00:16.874
The subscriptions hiding in expense reports.

00:00:17.333 --> 00:00:21.628
The episode I warned you way back in episode one might make your security team sweat.

00:00:22.366 --> 00:00:26.557
If you're new to the show, welcome. Glad to have you. This one stands alone

00:00:26.557 --> 00:00:28.670
just fine, but it connects to everything.

00:00:29.024 --> 00:00:33.487
The gates from episode two, the ledger from episode three, and the audit findings

00:00:33.487 --> 00:00:34.974
we spent two episodes fighting.

00:00:35.624 --> 00:00:37.507
Because here's the through line. An

00:00:37.507 --> 00:00:41.406
awful lot of what goes wrong in asset management was born in the shadows.

00:00:42.085 --> 00:00:45.626
Today we find out where the shadows come from. And I'll warn you now,

00:00:45.760 --> 00:00:49.581
the answer is going to be uncomfortable. Because some of it comes from us.

00:00:50.330 --> 00:00:54.388
Good morning, good afternoon, or good evening, wherever you happen to be listening from.

00:00:54.783 --> 00:00:58.927
This is the Operational ITAM Podcast, the show where we take the unglamorous

00:00:58.927 --> 00:01:02.068
machinery of enterprise technology and make it make sense.

00:01:02.735 --> 00:01:08.018
I'm your host, Bill, and today's episode requires a little humility, mine included.

00:01:08.535 --> 00:01:12.285
Welcome in, grab your coffee, grab your headphones, and let's get into it.

00:01:12.912 --> 00:01:16.679
First, the definition, because I promised you I'd always define my terms.

00:01:17.521 --> 00:01:22.188
Shadow IT is any technology—hardware, software, cloud service,

00:01:22.333 --> 00:01:27.097
and lately, artificial intelligence—acquired and used inside your organization

00:01:27.097 --> 00:01:30.965
without the knowledge, approval, or management of your IT function.

00:01:31.876 --> 00:01:34.721
The marketing team's design subscription on a corporate card,

00:01:35.058 --> 00:01:39.197
the developer's personal cloud account running a workload that quietly became production,

00:01:39.731 --> 00:01:43.010
the department that bought its own laptops because yours took too long,

00:01:43.515 --> 00:01:47.727
And the newest resident of the shadows, the AI chatbot half your workforce is

00:01:47.727 --> 00:01:49.910
using and most of them aren't mentioning.

00:01:50.497 --> 00:01:53.869
We'll get to that one after the break, and it deserves its own segment, believe me.

00:01:54.432 --> 00:01:57.827
Now before we talk about scale, I want to do something a little unusual for

00:01:57.827 --> 00:02:01.573
this show. I want to defend the people in the shadows.

00:02:02.530 --> 00:02:05.859
Because the standard industry framing, shadow IT as villainy,

00:02:06.199 --> 00:02:11.308
employees as rule breakers, IT as the wronged party, is not just self-serving,

00:02:11.691 --> 00:02:13.020
it's analytically wrong.

00:02:13.566 --> 00:02:17.043
And if you start from the wrong diagnosis, you'll prescribe the wrong cure.

00:02:17.815 --> 00:02:21.217
Here's the honest truth from somebody who has sat in the CIO chair.

00:02:21.664 --> 00:02:23.515
Shadow IT is a demand signal.

00:02:24.299 --> 00:02:28.020
Every unsanctioned tool is a business need that showed up at your front door,

00:02:28.130 --> 00:02:30.261
found it locked, and went around back.

00:02:31.114 --> 00:02:35.809
One industry survey found that only about 1 in 8 IT departments can actually

00:02:35.809 --> 00:02:40.123
keep pace with the demand for new technology requests. 1 in 8.

00:02:40.832 --> 00:02:43.989
Which means, for the overwhelming majority of organizations,

00:02:43.989 --> 00:02:49.224
the official answer to, I need a tool to do my job, is, functionally, wait.

00:02:49.991 --> 00:02:56.190
And people whose bonuses, deadlines, and careers are on the line do not wait. They swipe the card.

00:02:57.003 --> 00:03:01.969
Would you wait? Be honest. I've been the guy whose process was being routed

00:03:01.969 --> 00:03:07.237
around, and when I finally stopped being offended long enough to ask why, the answer stung.

00:03:07.795 --> 00:03:10.331
Because my process was slower than their problem.

00:03:11.202 --> 00:03:14.517
Write that down, because it's the thesis of this whole episode.

00:03:15.086 --> 00:03:19.399
People don't route around IT because they're villains. They route around IT

00:03:19.719 --> 00:03:22.586
because the sanctioned path is slower than their problem.

00:03:23.218 --> 00:03:27.699
That's the empathy. Now here's the accounting, because sympathy for the cause

00:03:27.699 --> 00:03:30.782
does not erase the cost. Scale first.

00:03:31.915 --> 00:03:37.163
Gartner has estimated that 30-40% of IT spending in large organizations happens

00:03:37.163 --> 00:03:40.989
as shadow IT, outside IT's visibility and control.

00:03:41.871 --> 00:03:45.643
Sit with that. If your official technology budget is $50 million,

00:03:46.665 --> 00:03:50.753
there may be another $15 or $20 million in technology spend scattered across

00:03:50.753 --> 00:03:55.871
expense reports, departmental budgets, and purchase cards that your ledger has never heard of.

00:03:57.009 --> 00:04:01.313
And the visibility gap is worse than the spend gap. Research has repeatedly

00:04:01.313 --> 00:04:06.163
found organizations formally tracking on the order of 100 cloud services while

00:04:06.163 --> 00:04:08.544
actually using close to 10 times that number.

00:04:09.490 --> 00:04:12.758
10 times. Your ledger from episode 3?

00:04:13.489 --> 00:04:17.286
If you haven't hunted the shadows, the right hand column, the deployment side,

00:04:17.588 --> 00:04:20.711
isn't just incomplete. It's a rounding error of the truth.

00:04:21.338 --> 00:04:25.082
Now, the risks, and there are three, in escalatory order.

00:04:25.813 --> 00:04:30.092
Risk 1. Waste. Shadow purchases mean duplicate tools.

00:04:30.370 --> 00:04:33.627
Three departments buying three different project management platforms,

00:04:33.760 --> 00:04:39.420
none at volume pricing, all renewing automatically, none appearing at any renewal negotiation.

00:04:39.914 --> 00:04:45.085
Remember the SaaS numbers from episode 3? Roughly half of purchased licenses going unused?

00:04:45.567 --> 00:04:49.243
Shadow spend is where that problem goes to hide, because nobody can right-size

00:04:49.243 --> 00:04:51.244
a subscription nobody knows exists.

00:04:52.004 --> 00:04:56.103
Risk 2. Audit exposure. Think back to the two-parter.

00:04:56.724 --> 00:05:00.636
Where do you suppose those findings come from? They come from right here.

00:05:01.414 --> 00:05:05.222
A department stands up software on a server nobody told licensing about.

00:05:05.971 --> 00:05:12.026
A team downloads a free tool, and we learned in Episode 3 exactly what free can turn into.

00:05:12.589 --> 00:05:16.972
One licensing change, and that friendly little runtime is billing you per employee.

00:05:17.825 --> 00:05:21.999
When the audit letter lands, the shadows get counted. At list price.

00:05:22.498 --> 00:05:24.147
Whether you knew about them or not.

00:05:24.814 --> 00:05:29.197
Ignorance is not a defense recognized in any license agreement I have ever read.

00:05:29.789 --> 00:05:32.105
And I have read more of them than any human should.

00:05:32.819 --> 00:05:38.160
Risk three, and it's the big one, security. Unsanctioned tools mean unmanaged tools.

00:05:38.282 --> 00:05:42.675
No patching, no multi-factor enforcement, no offboarding when the employee leaves,

00:05:42.964 --> 00:05:44.945
corporate data sitting in personal accounts.

00:05:45.461 --> 00:05:49.666
The breach research year after year links a meaningful share of security incidents

00:05:49.666 --> 00:05:54.384
to unmanaged, unsanctioned technology, and breach costs run to the millions.

00:05:55.219 --> 00:06:00.113
Every zombie asset, remember the zombies? A device on your network your record's never heard of?

00:06:00.665 --> 00:06:04.426
Every shadow app is the software version of a zombie, drawing data,

00:06:04.726 --> 00:06:07.671
drawing risk, invisible to every control you have.

00:06:08.187 --> 00:06:11.693
Your security team can't protect what your asset program can't see.

00:06:12.437 --> 00:06:16.136
That sentence, by the way, is how you get your security team to fund your asset

00:06:16.136 --> 00:06:17.691
program. You're welcome.

00:06:18.352 --> 00:06:21.644
If you're enjoying the show, do me a favor, like and subscribe,

00:06:21.876 --> 00:06:25.556
post your comments, and share this episode with a department head who's quietly

00:06:25.556 --> 00:06:27.147
built their own technology stack.

00:06:27.698 --> 00:06:31.790
And when you send it, tell them it's not an ambush. This one's on their side.

00:06:32.319 --> 00:06:36.422
All right, and now the segment I promised. Let's talk about the newest,

00:06:36.579 --> 00:06:39.863
fastest growing shadow in the building, shadow AI.

00:06:40.438 --> 00:06:44.542
Everything we just said about shadow IT applies, but compressed and amplified.

00:06:45.250 --> 00:06:49.761
Employees adopted generative AI tools faster than any technology in my 30 years,

00:06:50.278 --> 00:06:53.267
faster than any governance process on earth could keep up with.

00:06:54.010 --> 00:06:58.370
And surveys suggest a majority of the people using AI at work are doing it quietly.

00:06:59.113 --> 00:07:03.076
One widely cited finding put the share of workers hiding their chatbot use from

00:07:03.076 --> 00:07:05.121
their employer at around 70%.

00:07:06.015 --> 00:07:10.024
Now here's why this shadow is darker than the old ones. When somebody used an

00:07:10.024 --> 00:07:15.367
unsanctioned file sharing app in 2015, the data sat somewhere it shouldn't. Bad.

00:07:16.168 --> 00:07:19.146
But when somebody pastes your source code, your customer list,

00:07:19.477 --> 00:07:24.161
or your draft contract into an unmanaged AI tool, that data doesn't just sit.

00:07:24.643 --> 00:07:28.712
It leaves. Into a third-party system entirely outside your control.

00:07:29.560 --> 00:07:32.439
The payload changed, and the breach data has caught up.

00:07:33.124 --> 00:07:37.594
IBM's most recent cost of a data breach research found roughly one in five breached

00:07:37.594 --> 00:07:42.194
organizations was compromised through Shadow AI, and those incidents ran hundreds

00:07:42.194 --> 00:07:45.186
of thousands of dollars more expensive than the average breach.

00:07:46.114 --> 00:07:49.226
This is not a future problem. This is a current line item.

00:07:49.847 --> 00:07:55.849
But, and here is where 30 years of pattern recognition earns its keep, we have seen this movie.

00:07:56.523 --> 00:08:01.666
Shadow AI in 2026 is Shadow Cloud in 2012 wearing a smarter costume.

00:08:02.194 --> 00:08:06.298
And we know how that movie ends, because the industry already ran the experiment.

00:08:06.931 --> 00:08:12.454
Bans alone failed. Blocking the popular tool just pushed usage to personal devices

00:08:12.454 --> 00:08:15.842
and lesser-known tools that were harder to see and worse on security.

00:08:16.411 --> 00:08:20.714
What worked was visibility plus a sanctioned alternative. Give people an approved

00:08:20.714 --> 00:08:24.451
tool that's actually good, make it easier to use than the shadow version,

00:08:24.700 --> 00:08:26.709
and watch the shadows thin out on their own.

00:08:27.312 --> 00:08:30.459
Keep that in your pocket. It's about to become the whole strategy.

00:08:30.912 --> 00:08:35.514
Which brings us to a listener question right on cue. Dave from Milwaukee writes,

00:08:36.025 --> 00:08:40.872
Bill, my CISO wants to block every unsanctioned app at the firewall and be done with it.

00:08:41.359 --> 00:08:44.773
Gut feeling says that's wrong, but I can't articulate why.

00:08:45.887 --> 00:08:49.945
Dave, your gut is smarter than the firewall. And here's the articulation.

00:08:50.473 --> 00:08:55.105
Blocking treats the symptom and feeds the disease. The need that drove someone

00:08:55.105 --> 00:08:57.828
to that tool doesn't vanish when the tool gets blocked.

00:08:58.426 --> 00:09:02.112
It goes to their phone, their home network, their personal account,

00:09:02.536 --> 00:09:06.663
where you have zero visibility instead of partial visibility.

00:09:07.226 --> 00:09:09.966
You haven't eliminated the shadow, you've made it darker.

00:09:10.715 --> 00:09:15.375
Block the genuinely dangerous stuff, absolutely, but pair every block with a

00:09:15.375 --> 00:09:20.624
sanctioned path, or you're just running an expensive program to make your blind spots blinder.

00:09:21.263 --> 00:09:23.312
So how do you actually find the shadows?

00:09:24.409 --> 00:09:28.089
You promised your flashlight. Here it is. Five beams.

00:09:28.589 --> 00:09:32.015
None of them require new budget, and every one of them uses data you already

00:09:32.015 --> 00:09:35.147
have, which by now you know is my favorite kind of project.

00:09:35.716 --> 00:09:37.934
Beam 1. Follow the money.

00:09:39.076 --> 00:09:43.190
Pull your purchase card and expense report data and filter for software and

00:09:43.190 --> 00:09:44.347
subscription merchants.

00:09:45.183 --> 00:09:49.130
Then pull your accounts payable vendor master and look for technology vendors

00:09:49.130 --> 00:09:50.802
that procurement never onboarded.

00:09:51.452 --> 00:09:54.471
The shadows are self-documenting. People expense them.

00:09:55.179 --> 00:09:58.842
It is the single richest shadow IT data source in your company,

00:09:59.063 --> 00:10:02.232
and it's sitting in finance, one polite email away.

00:10:03.016 --> 00:10:08.210
Beam 2. Follow the identity. Your single sign-on and identity platform logs

00:10:08.210 --> 00:10:12.150
show what applications people authenticate to, including the ones that were

00:10:12.150 --> 00:10:13.349
never formally sanctioned.

00:10:14.034 --> 00:10:15.642
And look hard at OAuth grants.

00:10:16.338 --> 00:10:20.350
Every time an employee clicks Sign In With Your Work Account on some third-party

00:10:20.350 --> 00:10:22.875
app, they've connected that app to your environment.

00:10:23.455 --> 00:10:27.176
That consent log is a shadow inventory writing itself in real time.

00:10:27.838 --> 00:10:30.044
Beam 3. Follow the network.

00:10:31.029 --> 00:10:34.407
DNS queries, web gateway logs, egress traffic.

00:10:35.069 --> 00:10:38.720
Your network team already sees every cloud service your organization talks to.

00:10:39.277 --> 00:10:43.544
They've just never had a reason to hand asset management the list. Give them the reason.

00:10:44.397 --> 00:10:48.687
Beam 4. Follow the endpoint. Your endpoint management platform,

00:10:48.914 --> 00:10:52.750
the one I've been telling you to lean on since episode 1, knows every installed

00:10:52.750 --> 00:10:55.764
application, including the ones that never pass through a gate.

00:10:56.408 --> 00:11:00.698
That's the same normalization work from episode 3, pointed at a darker corner.

00:11:01.470 --> 00:11:06.805
And lastly, beam 5, the one everyone forgets. Just ask.

00:11:07.606 --> 00:11:13.538
Run an amnesty survey. Tell people, in plain language, no punishment, no gotcha.

00:11:14.177 --> 00:11:18.142
Tell us what you're using and why, and we'll try to make the good stuff official.

00:11:18.943 --> 00:11:23.291
You will be astonished what people volunteer when the question isn't an accusation.

00:11:24.225 --> 00:11:29.629
And the why answers? That's free consulting on exactly where your service catalog is failing.

00:11:30.528 --> 00:11:34.789
Which brings us to the response. You've found the shadows. Now what?

00:11:35.863 --> 00:11:39.619
Three buckets. And notice there's no bucket labeled punish.

00:11:40.751 --> 00:11:46.573
Bucket 1. Adopt. A lot of shadow tools are genuinely good. That's why people chose them.

00:11:47.084 --> 00:11:52.239
Sanction them. Negotiate real terms. Bring them into the ledger and let the team keep their tool.

00:11:52.912 --> 00:11:55.234
You just converted an adversary into an ally.

00:11:56.564 --> 00:12:02.615
Bucket 2. Migrate. Where 5 shadow tools do one job, consolidate to one sanctioned

00:12:02.615 --> 00:12:06.316
choice, with a transition period and actual help, not a memo.

00:12:07.773 --> 00:12:12.568
Bucket 3. Retire. The genuinely dangerous ones, the data leaking,

00:12:12.719 --> 00:12:17.032
the unlicensed, the abandoned, get shut down, with an explanation of why.

00:12:17.322 --> 00:12:22.669
Because when IT says because I said so, that is how the shadows formed in the first place.

00:12:23.889 --> 00:12:27.585
And then the permanent fix, and you knew this was coming. Gate 1,

00:12:27.760 --> 00:12:29.902
Episode 2, the request gate.

00:12:30.749 --> 00:12:35.684
A fast, clear, well-stocked service catalog is the single best shadow IT prevention

00:12:35.684 --> 00:12:39.874
technology ever invented, because the shadows are exactly the shape of whatever

00:12:39.874 --> 00:12:41.721
your catalog can't deliver quickly.

00:12:42.435 --> 00:12:45.958
Speed up the front door, and the back door traffic drops on its own.

00:12:46.829 --> 00:12:51.364
Let's take it to the library one more time. The head librarian is doing her

00:12:51.364 --> 00:12:54.439
rounds one evening, and behind the boiler room she finds.

00:12:55.275 --> 00:12:58.886
A second library. Shelves the patrons built themselves.

00:12:59.356 --> 00:13:03.797
Books nobody cataloged. Some wonderful. Some borrowed from lenders with,

00:13:03.983 --> 00:13:06.409
let's say, alarming terms and conditions.

00:13:07.123 --> 00:13:11.744
Her first instinct is fury. But then she reads the note tacked to the doorframe.

00:13:12.487 --> 00:13:15.598
Acquisitions take six weeks. We needed the books now.

00:13:16.353 --> 00:13:21.032
And a wise librarian, the kind who keeps her job, doesn't burn down the reading room.

00:13:21.566 --> 00:13:25.804
She catalogs the good books, replaces the dangerous ones, thanks the patrons

00:13:25.804 --> 00:13:29.193
for showing her exactly which collections the library was failing to stock,

00:13:29.866 --> 00:13:31.886
and then she fixes the acquisitions desk.

00:13:32.641 --> 00:13:37.761
Because the secret library was never a rebellion, it was a suggestion box she'd been ignoring.

00:13:38.556 --> 00:13:43.926
One closing principle, and the arc continues. Hardware is a custody discipline.

00:13:44.349 --> 00:13:46.137
Software is an evidence discipline.

00:13:46.904 --> 00:13:51.101
Audit defense is a process discipline. Settlement, a commercial discipline.

00:13:51.954 --> 00:13:55.518
Shadow IT? Shadow IT is a service discipline.

00:13:56.244 --> 00:14:00.164
The shadows are not a security problem you punish or a compliance problem you

00:14:00.164 --> 00:14:03.128
audit. They are a service problem you out-compete.

00:14:04.063 --> 00:14:07.807
Make the sanction path faster than the shadow path and the shadows shrink.

00:14:08.603 --> 00:14:12.393
Let the sanction path stay slow and no policy on earth will save you.

00:14:13.455 --> 00:14:17.025
You don't fight shadows with rules. You fight shadows with light.

00:14:17.628 --> 00:14:20.513
And with a front door that opens faster than the back door.

00:14:21.152 --> 00:14:25.376
Class dismissed. Homework. And this one might be the most eye opening yet.

00:14:25.899 --> 00:14:31.959
Get 90 days of purchase card and expense data from finance. Software and subscription merchants only.

00:14:32.371 --> 00:14:36.475
Count the distinct products. Then count how many appear in your asset records.

00:14:36.836 --> 00:14:41.021
That gap, that exact number, is your shadow. Measured, dated,

00:14:41.264 --> 00:14:44.597
and ready for the one page brief you learned to write last episode.

00:14:45.107 --> 00:14:48.823
Fair warning from someone who has run this exact exercise more than once,

00:14:49.107 --> 00:14:50.803
the number will be bigger than you think.

00:14:51.052 --> 00:14:53.792
It always is. Every single time.

00:14:54.441 --> 00:14:57.844
Next episode, we take on a fight I promised in the very first show,

00:14:58.250 --> 00:14:59.766
the Great Refresh debate.

00:15:00.323 --> 00:15:03.074
Three years? Four? Five or more?

00:15:03.886 --> 00:15:09.222
The vendors have brochures, the CFO has opinions, and I have actual failure rate data.

00:15:09.704 --> 00:15:13.023
And one of those three is going to win. We'll find out which.

00:15:13.795 --> 00:15:18.050
That's today's episode. The flashlight is yours now. Go point it somewhere uncomfortable.

00:15:18.712 --> 00:15:23.240
I'm Bill Van Nort. This is the Operational ITAM Podcast. Like the shadows,

00:15:23.240 --> 00:15:26.537
speed the front door, and I'll see you next week. Take care.