The Leading in a Crisis Podcast
Interviews, stories and lessons learned from experienced crisis leaders. Email the show at Tom@leadinginacrisis.com.
Being an effective leader in a corporate or public crisis situation requires knowledge, tenacity, and influencing skills. Unfortunately, most of us don't get much training or real experience dealing with crisis situations. On this podcast, we will talk with people who have lived through major crisis events and we will tap their experience and stories from the front lines of crisis management.
Your host, Tom Mueller, is a veteran crisis manager and trainer with more than 30 years in the corporate communications and crisis fields. Tom currently works as an executive coach and crisis trainer with WPNT Communications, and as a contract public information officer and trainer through his personal company, Tom Mueller Communications LLC.
Your co-host, Marc Mullen, has over 20 years of experience as a communication strategist. He provides subject matter expertise in a number of communication specializations, including crisis communication plan development, response and recovery communications, emergency notifications and communications, organizational reviews, and after-action reports. He blogs at Blog | Marc Mullen
Our goal is to help you grow your knowledge and awareness so you can be better prepared to lead should a major crisis threaten your organization.
Music credit: Special thanks to Nick Longoria from Austin, Texas for creating the theme music for the podcast.
#crisis #crisismanagement #crisiscomms #crisiscommunications
#BusinessContinuity #LeadingThroughCrisis
#CrisisResponse
The Leading in a Crisis Podcast
EP84 Using AI to prepare for crisis, with Clearline Crisis
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Most companies don’t fail a crisis because they lack smart people. They fail because the “plan” is a dusty document, the spokesperson is untested, and the first real coordination happens at 11:30 on a Sunday night. We sit down with Mitch Cohen, founder of ClearLine Crisis and a 25-year crisis communications veteran, to talk about what it takes to build real readiness before reputational risk turns into an enterprise-threatening event.
We get specific about AI in crisis management: where it genuinely helps and where it can create dangerous false certainty. Mitch explains how ClearLine tackles the problem with two guardrail layers, a crisis communications doctrine that drives better decisions and a language layer designed to avoid generic AI voice. From there, we walk through how scenario planning questionnaires generate editable crisis playbooks across dozens of crisis types, and how audio-based spokesperson training can pressure test key messages with increasingly aggressive questions.
Reach ClearLinecrisis at clearlinecrisis.com
We'd love to hear from you. Email the show at Tom@leadinginacrisis.com.
Intro And New AI Tool
Tom MuellerHey everyone and welcome back to the Leading in a Crisis Podcast. So happy to have you with us again today. I'm Tom Mueller. Hey, on the podcast today, we're gonna take a deep dive into the world of AI once again. In the past, we've talked about a number of AI tools that are out there that help make life easier for corporate communications folks or organizations with smaller teams. Well, on today's episode, there's uh brand new AI communications uh and response. Well, it's really it's a commute, it's a response and planning tool uh that's been developed by a company called Clear Line Crisis. And the principal there is Mitch Cohen. Uh Mitch, welcome to the podcast.
Mitch CohenThanks for having me, Tom. It's a pleasure. I've been looking forward to this.
Tom MuellerMitch, take uh take just a minute and give our listeners a quick thumbnail of your background and career path.
Mitch CohenSure
Why Teams Start From Zero
Mitch Cohenthing. I've been uh I've been in crisis communications for over 25 years now. I have uh been on every side that you can imagine. Uh startups, in-house, agency, massive corporations, mid-size organizations all around all around the field. Um, most of my background has been in heavily regulated industries, uh, financial services and housing and um data and healthcare. And one thing I've noticed in my career over this entire quarter century has been that no matter the organization, whether they're a small startup or a massive corporation, when crisis hits, in a certain sense, everyone is always starting from zero. Even if there is a crisis communications plan, chances are it's on a share drive somewhere. It was last opened by someone, maybe uh six months, a year ago. The spokesperson may or may not still work at the organization. The regulatory stance may or may not be accurate. And uh the plan isn't so much a plan as it's a document. Uh, so I've seen this happen. I've been in on the receiving end of that 11:30 Sunday night call where suddenly we have to uh uh address a reputational crisis. Um but the preparedness piece has never been there. And I think the reason for that is because most comms teams, unless you have a dedicated crisis communications function, which many companies are not fortunate to have.
Tom MuellerNo, very unusual.
Mitch CohenThe comms team generally has been built for visibility, right? To increase visibility, to increase credibility and authority. And it's a it's you know, I don't want to say an offshoot of marketing, but uh a kissing cousin of marketing. And it's a it's you know, and that's what the team is designed for, not to save the company in the worst week of its existence. So seeing this over and over and over again uh led me to think that there had to be a better way.
Tom MuellerOkay. Well, you know, we we're spending more and more time talking about AI tools out there because they're really, you know, the intention is, as I understand it, is to uh take on some of those mundane tasks that uh would then free up staff for more value-added, you know, planning uh type um functions uh or customer facing functions. So AI is intended to help, you know, sort of relieve some of the workload and automate things in the background.
Why Generic AI Fails In Crises
Tom MuellerBut now when you apply it to a crisis tool, what um, you know, tell us what your your thought process was going into that, Mitch, in terms of, hey, I've got all this crisis experience. Now I want to go, you know, beat my head against a wall and learn AI tools. Um you know, so and that's a a deep dive in itself. But um, so how did you why did you decide to do that?
Mitch CohenWhy is an entirely different question, Tom? And I don't know that I could answer that one. Uh, but um what I am uh I am fortunate enough that I'm a relative old timer when it comes to AI, particularly from a uh comms perspective. I worked, uh I ran public relations for a company called Black Knight, which was uh mortgage technology and data provider, uh, at one of the largest players in the field. Back in 2018, Black Knight acquired uh in a fit of seven acquisitions in two and a half years, acquired a company called Heavywater. And Heavywater was uh an AI machine learning firm. And this was back in 2018, which at the rate of uh the technology uh improving is the relative dark ages. Uh but it gave me a real strong foundation in what the technology could actually do as compared to what people assumed it could do or worried it could do. Um when it comes to crisis communications, the last thing in the world you want is that generic AI nonsense, you know, well researched, but you know, not of any value, not supported by any sort of uh experiential understanding. Um, you get what feels like a confident response, hey uh chat, let work me up a crisis communications plan. We're we're having a problem here. And you'll get stuff that'll come back and it'll sound extremely confident and will, you know, maybe even look the part, but it's not. And it there's there's uh it's it's not the sort of file that I would want to have in my crisis response. But I did start working with AI as a user shortly thereafter, and I grew up with the LLMs. The last 18 months or so before I launched Clearline, before I began the building of the platform in earnest, were spent eliminating that uh that AI voice and that AI uh false certainty.
Doctrine Guardrails For Better Outputs
Mitch CohenAnd what I did there was I built two separate layers for every prompt to go through. And this was for my use in my consultancy originally. And each of these two layers serves a a distinct purpose. The first is what we call a clear line doctrine, and it's a crisis communications philosophy. And it hold, you know, what a good holding statement should contain, what a stakeholder sequencing should look like, uh, you know, don't attack your accusers. There's a there's a an entire philosophy baked in. And it's a very, very uh comprehensive and complex series of instructions that then inform or began to inform every interaction I had with the AI.
Tom MuellerOkay, so you're basically taking your accumulated knowledge uh over your career, putting that into a series of rules and guidance that form sort of the basic operating structure for the AI to use.
Mitch CohenExactly. But it can't stop there because we need it to sound not like chat GPTs in the room, but like there's a senior crisis communications advisor in the room. So there's another
Making AI Sound Like An Advisor
Mitch Cohenlayer. Uh, and I have an MFA in fiction writing from the Iowa Writers Workshop, right? So I take I take language exceedingly seriously, right? And and I mean, you know, word choice, commas, everything means a lot to me. And AI slop, uh it personally offends me, you know, and and the and the fact of the matter is, and I don't want to cast any stones. But you know, if you go on LinkedIn these days and you just you read 10, 15, 20 posts, they all sound exactly the same now because they're all being pumped through one of the major LLMs and they're coming out and they sound uh interesting and they're generally pretty good stuff.
Tom MuellerI you know, I've been amazed, but but the language lyric.
Mitch CohenYes. So the second layer is language, and that took a lot, and it and it's something that I'm constantly honing. Uh, but like I said, 18 months spent building these layers that would wrap. When they became functional enough that I would use them myself, I started thinking, well, I know many of my peers. I mean, I I would have killed to have this, not literally, but I would have loved to have had this functionality at in any one of the many crises I've been involved in, which is, you know, from small reputational desk ups to um antitrust uh litigation, IP theft, an 18-month FTC challenge, uh $13 billion acquisition. Any one of these, if I had had this tool, would have been measurably easier.
Tom MuellerSo, Mitch, when you're talking to potential customers and clients about this tool, uh, what's you know, what's the sales pitch or the USP for for this tool then to you know to make it something people are worth, you know, spending uh a subscription fee every month for, or maybe a much larger um cash you know fee for larger organizations? What's what's in that?
Mitch CohenYeah, let me let me back up just a
Reputation Risk As Enterprise Risk
Mitch Cohenbit. Uh so my my basic thesis is that reputation risk is enterprise risk. It's just never been managed that way or priced that way, or uh uh every other level of risk, operational, legal, financial, uh compliance, has an infrastructure built to manage it. The reputation layer is intuition, emails, word docs, uh teams of you know, incredibly talented people doing a lot of work in a lot of different places. But there's no infrastructure. And everyone is essentially starting from zero again, each time a new crisis occurs. Uh, there's the reactive method of hiring a firm, incredibly talented people to come in and see you through the crisis and get you to the other side. But that's reactive, and they're coming in and they're figuring out your business as they're helping you through the crisis. Um what I wanted to do was I wanted to create a readiness and response layer that could serve as
Scenario Readiness And Playbooks
Mitch Cohenthat infrastructure. And that's what Clearline is. On the front end, on the readiness side, it allows for an organization to run through various scenarios. Um, and they're guided by questionnaires, right? Uh asking questions about the given scenario. And there are 137 scenarios in the library now across uh more than 15 different industries. And so they cover data breaches, they cover natural disaster, they cover vendor supply chain interruption, but also stuff like executive misconduct or labor disputes or um, you know, it it's a it's a robust library and it continues to grow as we build out more scenarios. Uh but they the scenarios are there to walk the client through the questionnaires and then create a playbook, a crisis playbook based upon their answers to that questionnaire as well as the company profile within ClearLine, which is rather robust and has you know close to 40 different fields that provide insight and input into the artifacts. So when it produces the playbook, it's producing it from the facts it knows about the company, about the industry it's in, about its executive team, about the areas uh of the of the country or the globe in which it operates, so on and so forth. Uh there's another piece, and and what what it does is it produces a playbook, which is fully editable and intended to be. It's a work product that is meant to then be taken by the professionals, fine-tuned, and then sent out. Uh, always a human in the loop with clear line.
Audio Spokesperson Training Pressure Test
Mitch CohenUm but the there is also then a spokesperson training component, which allows you to uh train your spokesperson. Uh and what it is, it's an audio-based interview mini media training where the reporter, such as it is, keeps asking increasingly aggressive questions uh to test. Those are graded and uh pointers are given, and then it can be done again to bring a uh spokesperson up. So that's that's readiness.
Tom MuellerOkay, that's readiness. Let me pause you right there, we'll circle back, but I'm so curious about the spokesperson training uh piece of that. Uh, because essentially what you're building here now is the opportunity uh for individuals to go in and practice their responses to a media interview, right? So the the tool will ask questions and I give my answers. Is is that all written or is that verbal?
Mitch CohenIt's verbal.
unknownOkay.
Mitch CohenIt's verbal. You could do it written, but I think it it takes the uh the pressure testing out of it. Absolutely. Uh when when you do it verbally, and and here's the great thing too, it's not only uh aligned to those uh crises scenarios that have been worked through, but you can you can run an interview prep on anything. So if there's a board briefing or a regular media interview coming up, um, you can use the system to train and fine-tune the spokesperson uh for whatever the engagement may be. But yes, it's audio and it goes back and forth and it it it hears the responses, it get and then it grades based upon whether key messages have been conveyed, whether uh, you know, there were holes in the answer that left room for another question, and it tries to uh create better spokespeople.
Tom MuellerYeah, that's fascinating. Okay. Well, I don't like that very much because I do a lot of media training, and uh I don't really want AI taking that job. But uh, you know, even you think about internal communications, you know, if you're gonna go to a town hall meeting somewhere uh and you know there's gonna be employee QA, uh again, another opportunity. Now, this tool isn't built specifically for that, but it wouldn't be a stretch to um, you know, to incorporate some kind of an internal communications aspect to it as well. So that's all very cool. All right.
Live Response And Regulatory Deadlines
Mitch CohenWell, next step is response, right? If and when a crisis event happens, and and they can be crises of any size, existential or you know, minor reputational dust ups. Uh, but when they come, you declare a crisis. And if there is an existing training scenario playbook, it is loaded in and the system updates that with facts on the ground and then provides immediate actions that should be taken, provides a draft of the holding statement and provides the stakeholder uh notification sequencing, which is where something that I'm pretty proud of comes into play. Um, the architecture of ClearLine allows for regulatory requirements to be plugged in as content, right? So, for example, when there's a data breach, there are 54 different jurisdictional reporting requirements depending upon where the company does business. Clearline knows from that profile where the company does business, they know from that profile what industry they're in, and they know from the regulatory library that plugs in at the top which regulations apply. And so when an active crisis is declared, one of the key components is the stakeholder sequencing. And it follows if you know part of our philosophy is always internal before external and legal holds aside. Uh, but it it goes through uh the stakeholder sequencing, including those regulators, and it lists out which regulator, what statute, and what's the deadline, and then it starts ticking down. And as you get closer, it starts, you know, it changes color and then it becomes red. And so it the system is used, uh, it becomes sort of a system of record for all of those various communications components that go out, and that includes the internal communications from the CEO to the employees, as well as the holding statement that that's readied, uh, and any of the regulatory notifications that go out, client communications, so on. And so that's during response. And every time anything is crafted, edited, uh, or changed in any way, or or veers from what has been uh pushed down as uh the standard that it must adhere to, any change is recorded, time and date stamped, and becomes part of the official record of the communications response. So later, when a regulator comes in or an investigator comes in, or you know, even just uh, you know, trying to internal audit to do best practices, rather than going through Tom's email and Mitch's email and the G drive and and how many things, you know, and and who who was it that approved, you know, the and rather than trying to piece all of that together, you just go to clear line and the lifecycle strip just has it all there, ready to ready to be exported for the auditor, right?
Tom MuellerOkay, so wow, you've got a complete record of your uh response communications here, essentially. Incredibly complex tool that you're building in all the the regulatory notification requirements as well as doing spokesperson's training, tracking all of this.
Audit Trail And System Of Record
Tom MuellerI'm sitting here, Mitch, wondering, well, who's who's leading this from the company side? Is it the communications team? Is it the crisis management team? Um, you know, when you're talking to companies, uh where would this live in the corporate world?
Mitch CohenThe titles change, the departments change, but who owns this is who owns risk generally. Whoever owns and to be more specific, whoever owns reputational risk. Like we said, most comms teams are built for increasing visibility and awareness. Uh risk is part of the calculus, but it's not their job. Um I would see the users and their and the the thing about Clearline is that it's it's fully team enabled. So
Who Owns It Inside The Company
Mitch Cohenthere are five seats for any any organization that. That signs up, has five seats, and you can add. So you can bring in your entire internal team. You can have somebody from legal on there. You can have somebody from compliance. Obviously, your comms person if you have one. Um, but the buyer is whoever that chief risk officer is, or you know, whoever owns risk. The user is across the organization. But I would see comms being the folks who would benefit most from it, especially given the fact that you have that interview training based upon any interview you want. You can generate executive briefings based upon a contact in the CR in the built-in CRM.
Tom MuellerSo it's good to live with whoever's managing reputation risks. And, you know, uh training and learning a new tool is always a challenge for an organization, right? And uh gosh, I went through this so many times working for a very large oil company, um, you know, trying to get people trained up on a tool and then trying to keep that training refreshed over time, you know, people change, different jobs, new people coming in. So what does the sort of onboarding process look
Onboarding And Organizational Memory
Tom Muellerlike? Is that fairly simple or is that uh, you know, as I call an opportunity for personal growth over time?
Mitch CohenUm, it is fairly simple. Uh, although when you get into that parent-child multi-organizational uh structure of the of ClearLine, it becomes more complex. Two things are at work there. The onboarding process itself is uh interview-based. So it's asking you questions and it's and it's pointing out why the answers matter, right? So it's it's asking questions, things to think about, and why why it matters. So it's a it's a a very self-served kind of process. Um there's also an organizational memory layer. Now it's not uh using data to train itself, but it's it does learn more about the organization. So when over time the system that you're using in month 12 is demonstrably more attuned to your organization than the one that you rolled out in month one. But the one that you rolled out in month one has walked you through this very simple onboarding process as well as uh the scenario questionnaires. There are two main groups that we're focusing on here, and one are larger enterprises, uh insurance carriers, private equity firms, um law firms and holding companies, even agencies that with larger clientele. Um but then the there's the group that the original product was first built for, and that is the small to mid-sized business uh in America. And there are somewhere north of 150,000 of those that are large enough to know that they have reputational risk exposure, but not large enough to afford the $25,000 a month crisis retainer to be in a state of preparedness and could barely afford it, even though it's a it's an existential cost that you must pay at the time of crisis, right? So what what do they do? They improvise, they don't do anything, they just you know, and there's I I don't know what the actual stat is because I I've learned that there's some doubt to it, but there isn't a very large share of companies that suffer a breach that don't make it out the other side. And it's a lack of preparation and a lack of a coordinated and sophisticated response that
Pricing For SMBs And Enterprises
Mitch Cohenputs them in that position. What I was trying to do was create a tool that could democratize the access to crisis communications, you know, for as wide a group as possible, and that group still exists. So we're selling to the large enterprises and those implementations probably a couple of weeks. But we're also selling to that vast you know sea of small to mid-sized businesses that could benefit from this, and who signing up is as easy as swiping a credit card and uh onboarding is as easy as going through an interview.
Tom MuellerAnd so you've got uh multi-tiered pricing for that. It really it's a subscription model for the small medium-sized businesses, right? It's a monthly fee of uh, I think I saw on your website $299 a month.
Mitch CohenThat's the yeah, that's for the readiness function. Uh, for readiness and response, it's a bit more. Okay. It's but it's still meant to be uh something that a business line manager could justify or that a small business owner could uh absorb without having to, you know, fire someone.
Tom MuellerRight. Well, yeah, to your point, relative to what you'd pay a PR agency or a crisis agency to come in and support you, or just to even be on retainer, you know, a fee at this level is pretty affordable, um, especially when you think about the preparedness um that it brings and helping your organization prepare for multi-scenarios and laying out so many different options for you to respond. So, yeah, very uh very interesting. Well, congratulations on on pulling this whole enterprise together, Mitch. And it sounds like it's been a lot of work for you. I I do have one question for you around confidentiality, though.
Privacy Promises And Encryption
Tom MuellerYou know, because every tool built on AI, you know, that that little thought lingers in the back of your head is like, how much personal information can we share with this uh and still feel confident that it's you know it's not exposing us to some you know, some other nefarious thing going on out there in the ether world. So how do you reassure people about the uh confidentiality of their company information using a tool like this?
Mitch CohenUh well, I personally and have always operated from uh a stance of maximum security paranoia, right? Always. Uh that's me personally, everything I've done in my own uh my own consultancy and in my personal life as well. Um, but and Clearline follows that same philosophy. There is absolutely no client data that is used to train any AI model. Okay, so that's first of all. Uh the only uh any transfer of information or or prompting is ephemeral. Uh the question or the information that is fed via API that goes through those the doctrinal and the language layers to to you know to make it uh to prevent hallucinations and to uh uh uh institute the proper guardrails. None of that is going to we use anthropic. None of that is going to anthropic for training purposes. Uh it's all ephemeral, it's all encrypted, uh at rest, as well as anytime it's in transit. And the build of the of the platform itself is such that each tenant's uh security is sacrosanct, and that's the only place where work product is stored. So the security maintaining the work product and keeping it uh for eyes only of the organization is always top of mind. And nothing is used for training, even the organizational memory, as I mentioned, isn't a matter of using it for training, it's just having a larger database of information to pull from when it's generating in that ephemeral state.
Tom MuellerAll right. So a strong reassurance there around the confidentiality of data using the AI tool. All right, well, that's always good to hear. We're always going to ask that question because you know it's it's on everybody's mind.
Mitch CohenAnd to be quite honest, uh, you know, uh, I I don't I don't trust, I I wouldn't trust an AI without that proper guardrails or you know, you you have to verify, trust to verify, someone said. And and and that's and that's that's what you need to do.
Tom MuellerAll right. Well, Mitch, thanks so much for coming on to the podcast. Uh uh and congratulations for you know kickstarting this new company and you know, taking advantage of the latest technology to help small medium companies as well as the big boys um to help be better prepared to deal with crisis. So thank you, Mitch, for pre-being on with us.
Mitch CohenWell, thank you, Tom, for having me. It was uh a pleasure and uh I really appreciate it.
Closing And Contact Info
Tom MuellerAll right. So if you're looking to learn more about this crisis tool, you can uh reach out to them at clearlinecrisis.com and uh you will find Mitch there on that website somewhere. And that's gonna do it for this episode of the Leading in a Crisis Podcast. Thanks so much for joining us. Hey, if you want to drop me a line, you can drop an email to Tom at leadinginacrisis.com, and we look forward to hearing from you. We'll see you again soon for another episode.