Full Transcript

For the folks that were here for what we're calling part one of this presentation, a lot of what we covered at that time — this is a couple months ago and that feels like a lifetime ago on the technology side — but in that webinar, we covered a lot of ChatGPT and some of the functionality that we have available for us there.

As things have progressed and platforms have evolved, a lot of our efforts and a lot of our activity has moved from ChatGPT over to Claude and the capabilities have expanded quite a bit too. For today's conversation I wanted to provide kind of a similar perspective, but using Claude and using some of the tools that we have available over there.

I've got the chat open here, so I'll dive right in and first I'm going to share my Claude screen here. This is a live demo, so as I'm sure you all know, live demos are the best way to make things not work when you're in the process of trying to work on them.

A lot of times when we talk about AI and when we speak about it in day to day, we're talking about large language models like ChatGPT, Gemini or Claude. And most of the time we think of this interface here, which is a chat interface.

It's a very capable and competent tool that we can use, but in a lot of ways it becomes almost a replacement for Google with a little bit more to it.

But if you haven't used Claude yet, there's a couple of things that can really increase what it can do for you. So first of all, you'll notice that I'm in a desktop app which looks a little bit different than the browser experience.

Now Cowork and Code both have a cloud component, but for the most part Cowork works on your local computer, so that increases what you can do safely because most of those edits and changes are happening locally.

I also have a personal account, but for that one I use a free plan. It's very common for me in conversations with folks that work at firms to pay for their own plan. And my suggestion would be to not do that.

Just like you wouldn't want an employee to have their own email account or to use their personal Gmail, there's going to be a lot of proprietary information shared in that account and you want to maintain the ability to turn that account off. The account should belong to the company.

It's very common for organizations who don't do that — the employees go and do it anyways because they see the value of what it brings to their role. So get a team plan or an enterprise plan, get everybody on it.

We can also see how many people are using it, how they're using it. This is advantageous because it helps you see who's using which product, what's the utilization rate.

When you have the desktop app, operate it in here within the settings. First of all, these profile settings matter. When it says your preferences will apply to all conversations within Anthropic guidelines, it's going to utilize the information that you share with it right here to inform what you're asking it to do.

When you go down to capabilities, we start to get into things that really impact the work that you do. Search reference chats allows it to see what you've done in the past. Generate memory from chat history lets it get to know you over time.

But this is where things start to get really interesting — under connectors it now allows you to make a direct connection to all of these other providers. Now we can make direct connections to our Office 365 account, Slack, Figma, Cloudflare, Circleback's note taking, and more.

We have really tight and fine-grained control here. We can turn it always allow, always needs approval, block everything. That's the other advantage of having a company account — the company account by default does not use your data for training.

Get these platforms set up, integrate them. This is going to make a world of difference. Under capabilities and skills, if you go to customize, you can add skills — inject your assistant with different specific capabilities.

We can technically work on our connectors from this view too. But with skills, we can actually inject our assistant with different specific capabilities. For example, if we want finance, install the finance skill and it becomes the next closest thing to your CPA.

Within Claude there is a legal plugin you can install. This does not turn Claude into an attorney, but it will have a pretty good body of content and decent instructions. It still needs a human in the loop.

Let me give you an example here. We're going to talk about using Client Intake.

Over here we have some client intake notes. These are fictitious, nothing private. Let's say your intake team takes a phone call, takes these notes and hits save.

What I'm going to do in Claude is take the notes file — this is kind of a mess — and give it instructions. I just finished an intake meeting with an injury client. Here's my notes. Create clean organized client intake summary.

It's going to reference those skills and actually do the work for me. It's going to give us instructions and show us how it's doing it.

I have it remember some of the things that I do. I do a lot of work in Markdown. We can ask it to put it out however we want.

I take the output and I tell it how to make it better and then I tell it to remember the feedback I'm giving.

The speed here is pretty great. If we build this into our workflow, we have an intake team who now needs to copy and paste this into case management. The organization has gone from chaos to really organized format. It identifies information we should have, gives us action items, case assessment notes, and a gap analysis.

Now I can give feedback on the document and say, okay, good start. But let's talk about scope of representation, add language about the client's obligation to cooperate.

There is. I think it's like 20 per chat. But now with Cowork, you can give it access to full directories. The theoretical limit is much higher. From the chat interface, I think it's 15 or 20.

Now those changes are made live. It also created a Word Doc. It's taken that messy client intake note, turned it into an organized document, turned it into an engagement letter.

Now I want to transition to the next area of Claude. The desktop version. We have chat, but chat is limited in the browser. When we install the desktop app, we have Cowork and Code.

In Cowork we can choose a folder for it to work within. I'm going to give it access to this Discovery Docs folder. We have emails, a maintenance log, a lot of different files.

Cowork is going to happen on your local machine. It's still going to use the AI through the cloud, but the work is going to happen on this machine. That gives us more security.

Going back to our Cowork example, I shared the folder with all the discovery documents. What the AI was able to do is go through all those documents and put them into a Word doc.

Each one has its own content, just kind of a mess.

The other piece where AI can really clean up — turning documents into spreadsheets. Here we've got messy damages data in a text file. The data is not usable.

We're going to prompt this to clean up this data. When I think of clear communication with an AI, it needs context and it needs purpose.

We provide context by giving it access to files and then give it purpose — what do we want it to do?

It will simultaneously read the file and start pulling in its Excel skill. While this is running, I want to cover what OpenClaw is.

While that builds, we can see the spreadsheet result. 63 formulas, zero errors. It's not just translating content but building out formulas within the file.

If we click in here, we can see it isn't just translating the content, but building out formulas. Our example was off one text file. But if we had a folder full of these, we could do that too.

Cowork, using those connectors, can also connect to websites through a browser session. If you have a WordPress site, you can tell it what to change and it will go into WordPress and make those changes.

Anything you can use a browser for, you can set Cowork onto. Sometimes you'll have to log in and once you're logged in, it can do it.

OpenClaw doesn't have the guardrails that Claude or ChatGPT or Gemini have because it's open source software. You can download it, change it, add things to it on a localized device.

The interface is very simple. I use Telegram so I can chat with it through my phone. But it has a lot of security risks.

The way I encourage people to set it up is to get a Mac Mini that is a separate device that doesn't have anything on it other than the operating software.

I named mine LaClaw. I can ask it to do all sorts of things and give it skills.

What this is doing is building an entire website based on the transcript and deploying it to Vercel. Astro is a framework for building websites. Vercel is a platform for hosting them.

No, it does have Internet access — it needs it to reach the APIs. But it doesn't have access to your logins or files. I keep mine off my work network, on a separate network and its own computer.

It has its own GitHub login that can share code to me but not take code back — a one-way share.

OpenClaw can take initiative on a schedule. If you wanted IT to do a daily review of your website for SEO opportunities, it could do that every day at a set time.

If you wanted to take more risk, give it backend access and say make three suggestions daily — if you approve, it implements them.

I was able to share transcript from the sponsor section to OpenClaw. It built and published this — available on the Internet right now.

That covers chat, that covers Cowork. In Code, we can build things too, but it won't publish for us. Claude Code can work in the cloud as well.

Cowork leans heavily on customizations and plugins to enrich what it can do.

When we think of using AI, the chat interface is one part. The bigger and more powerful part is having AI build things and build tools we can use every day.

Those tools, even though they're built by AI, may or may not have AI tied into them. What we're seeing is the SaaS Apocalypse — services we used to pay subscriptions for are becoming easier for us to create ourselves.

If you have someone on your team who likes this sort of work, get them a company-owned instance of Claude or Claude Code. Start throwing your more complicated problems towards them.

The biggest mistake is just not using it. You have a limitless professional at your fingertips who can analyze what's working and what's not. The easiest way is Claude Cowork.

There's no reason not to do an analysis on your website. I'm a big advocate for empowering yourself. Ask the question and be skeptical — it will get some things wrong.

The biggest mistake is not trying to use it. I've had a couple clients install OpenClaw on computers on their network — not a good idea, but they were quick to remove them.

Start big picture — here's our website, our brand guide, who we are, what should we be doing, compare with what we are doing.

For more practical things, have Claude Cowork help you write responses to every review you've ever gotten. Also, building out social media calendars.

You might need to babysit it a little bit. It'll open up a tab in your browser. But it will write responses to all your Google reviews.

Again, double check those and make sure it's not obviously AI, but you're better off responding with something.

The analysis side is really unending. You can ask for high level or get into the nitty gritty — how can I improve my schema on this website?

On the aspirational side, if I had a footprint covering 57 small towns, I'd use Claude to build a website for every single one, include contact information, and start publishing sites.

Building co-branded sites with referral partners. Things that used to take a day or two — we built that site with OpenClaw in five minutes.

That changes our tactical approaches and opens up opportunities to flood the zone with marketing and positioning.

As far as LaFleur goes, we've shifted from being a digital marketing firm to more of a growth engine partner. We still do digital marketing, but we bring technology and skill set that allows us to identify challenges preventing growth.

Our company is really focused on helping clients grow sustainably. We're comfortable in the AI space. We build digital tools, software.

For example, with the Volkswagen emissions testing, we pulled a specific data set for direct outreach. We can do the same now at broader scale using AI.

We like to dive in, get our hands dirty, find out how the firm works, and then look for solutions.

For example, if there's a specific construction company that built apartments with faulty windows, we can build that list using public resources in ways that weren't available before.

That's at a high level what we do. We still do traditional digital marketing, but now we're building tools to help clients grow.

On our podcast, I bring thought leaders from the marketing and AI space together. We look to make it accessible with practical takeaways.

AI implementation is so much more about trying it and experimenting than worrying it's going to replace everything.

Go check out this podcast. They've talked about some really good topics.

That's right. Probably the best one.

Thanks. Thank you for the opportunity. I hope people walked away with something they can use.

I'll make sure this gets added to our YouTube channel. Thanks for watching.

Thanks, everyone.

Bye, guys.

All right, that sounds great.

Thank you guys so much. This has been wonderful. Chip and Chad, I love the questions coming in. I'm going to bring in a couple of our friends. One of them is Patrick from OnPoint Legal Nurse Consulting.

He's been a nurse for over 21 years now, and he has been traveling across the country working on a lot of cases for all of you guys.

Nice to see you, Ginger. It's been way too many weeks. But yes, I was traveling and had conferences and trials. I am a nurse, been a nurse for 21 years. My background was in the hospital ICU, cardiac ICU, and then home health, hospice, long term care.

What we do over at OnPoint is we help from intake all the way through trial. Sarah is our director for record retrieval. She gets all of the records organized. We can do medical chronologies, indexing or bookmarking.

Dawn is our other director. She does a lot of our medical malpractice cases. We have a fantastic database of experts. We also have three life care planners on staff.

The pain and suffering reports that John Romano always talks about — those are a non-economic damage report. It gives a different perspective, a nursing perspective.

We've relaunched our website. You can submit cases now through the website securely. And we finally have our database with AI so we can search for experts faster.

Thank you so much for popping in, Patrick. You're doing a wonderful job.

Our TBI seminar is coming up in a few short weeks in Sun Valley, Idaho. There's still time to join us.

For those of y'all have not attended, get ready. It is the most beautiful place. There's so much to do. Ginger and them and John and Nancy just do a great job of learning but also having fun and building relationships.

My name's Aaron, I'm with HMR. We are the medical funding group that's been with Connectionology since day one. We operate medical funding in about 39 states now.

With medical funding we get with providers in your area, we pay them their fees up front and then run no interest to the client. When there's a resolve, you guys pay us back.

We also do litigation funding. For all these experts, it gets expensive and we want to be there. That way you can hire the best experts.

We also do pre-settlement funding. Your clients are catastrophically injured, they need money for mortgage or car note. We've helped people save their homes.

This year we started new programs on post-settlement and post-verdict. We can help fund the law firm with that verdict. And we're doing lines of credit.

I'm going to put my information in the chat. Call me. I work seven days a week.

I think we should give away a complimentary registration today.

Stay tuned, we'll announce the winner at the end. Thank you so much, Aaron.

I'm going to bring in two of our friends. You may have had a chance to meet Teddy Wu a week or two ago.

Thank you for your kind introduction. So just a small slice to go over. First of all, we are very honored.

In Anytime AI, our mission is to transform access to justice for every plaintiff. I got my PhD from computer science in 2016, later joined IBM as an AI scientist.

Later I joined Pinterest as a principal AI sensor. We founded Anytime AI at the end of 2023.

This was a pivot talk I gave two years ago — AI is a mandate, not an option. We are proud of our team. 70 plus US patents, 200 publications.

We have been launching Total Teddy, like Anytime AI 2.0, in early February. So many people love this.

What makes us different is we truly build beyond speed. We design for loading homes, for medical malpractice, for complex personal injuries.

If you have any type of record that is 10,000, 20,000 pages, this is the choice.

Most important thing is about security. There's a new case from New York where someone targeted Claude and got subpoenaed. Everything in Claude had to be provided to the defense side.

Anytime AI from day one realized how important security is. Every document uploaded is encrypted end to end.

We have incorporated workflow and skills designed specifically for different legal practice areas.

AI is not going to produce 100% ready stuff. This is just a jump start.

If you're interested about enterprise, check out our website.

That's awesome, Teddy. John Romano's been a big fan of yours.

Put your website in the chat so everybody can check it out.

Now our good friend Ryan is coming in with Fox AE.

You guys are doing some of the best animations, demonstratives, graphics across the country.

Thank you, Ginger. We specialize in anything visual for your cases. Whether still-shot illustrations or full recreation animations for car crashes, industrial accidents, construction incidents.

This is an animation we did with John Romano, a construction incident where they installed a beam that was not the right size.

You have these workers up on top of this platform, and then we show the fall happening.

We show exactly what happened with the injuries. We can do animations that show only injuries, or the environment, or combine both.

We do have a stock website called Fox Prime with ready-to-go visuals. We're always adding to it.

Best way to get started is to jump on a Zoom call where we can strategize how to show your specific case.

Fox Prime is like Amazon Prime but for graphics.

Thank you.

Let's see. Any more questions? If not, we'll get into our winners.

I don't think anything else has come in.

All right. I'm going to bring in Ted with HMR to announce the coffee winner.

Our winner today of the Terrazzo is Valerie Farwell out of Austin, Texas. Congratulations.

Congratulations, Valerie.

Let's give away a winner for our TBI seminar in Sun Valley, April 20th through the 24th.

For the Idaho conference, it's Edward Edwin Zinman out of Sausalito, California. Congratulations.

Congratulations.

There's some incredible speakers coming for the TBI seminar. It's small and intimate.

We have five additional seminars this year. Santa Fe with Eric Fong, Denver at the Gaylord Hotel, Nashville programs, and Lake Como Italy.

Chad, I was going to tell you. Let's see if any questions come up from the attendees. He said feel free to make it conversational. So. Yeah, got it. That'd be perfect. All right. And then I will butt in probably in about 35 minutes. So around 3:10, about 3:35. And around 4 and Easter time, I'll pop in just to do our, our partners so they can talk for a minute about what they do.

And then I would say, Chip, at some point we're going to talk about what you guys do too, because I want you to kind of get into that as well and all the great things that you guys offer at the floor.

Okay. And what we. We go till 4:30. 4:30 is the wrap up.

Yeah.

So we could do Q and A for probably the last 20, 20 plus minutes.

Yeah.

Cool.

That'd be perfect.

That sounds good.

And if we go over a little bit, I, I'm here check chat. If you have to jump off, that's okay. But I'll be here for as long as you guys need me.

Cool.

Okay.

All right, I'm going to hit start and then I'm going to. And I'll send everybody a reminder again afterward in a minute. Okay.

Okay, let's.

And thank you guys for doing this super fun. All right, I see everybody is starting to join us. Welcome back, everyone, for another great webinar here at Connectionology. As you guys know, my name is Ginger and the executive Director of Connectionology and we are so happy that you're here with us for part two of this amazing discussion on AI and technology.

And I can't wait to hear about all of the things that Chip is going to bring you. Hope that you guys are going to have a lot of good questions to ask him because we are going to answer your questions throughout this webinar in real time. So we want to make this conversational. We want you to come away learning a lot of new things that will help you in your practice and beyond that. Also later, you're going to meet some of our wonderful partners over at Anytime AI.

You're going to meet our friends over at Fox AE, HMR, as well as OnPoint Legal Nurse Consulting. And then of course, you're going to hear all about LaFleur Marketing today, of course, where Chip has created this incredible company. You're going to hear about the wonderful things that they do. So for now, I'm going to go ahead and get started by introducing our wonderful moderator.

And we are so excited to have Chad Mance back with us. He's spoken on so many of our wonderful seminars across the country for almost nine years now. But he's been an incredible trialer in Savannah, Georgia, for even more than that. He's a personal injury lawyer. He helps a lot of the surrounding communities in Georgia with all of their needs and helps a lot of his clients. He's the founder of the Mance law firm.

And he's known as the community's law firm for the injured. He provides a lot of aggressive legal representation in counties across Chatham, Fulton, and DeKalb counties. His mission has always been to help use the law to make society's crooked places straight.

He helps train a lot of lawyers nationwide at a lot of top events like National Trialers Summit, the Mass Torts Made Perfect, State Bar of Georgia convention, as well as a lot of programs here at Connectionology. And those topics range from jury selection, cross examination, trucking, loss strategy, and beyond. And he's also been regularly featured on VH1, TV1, WTOC, as well as WSAV, including the Atlanta Journal Constitution.

And so, Chad, we couldn't be more thrilled that you were here with us today. And I can't wait to see again what Chip is going to teach us.

Thank you, Ginger. Well, thank you all for having me. This is a great day because the intersection between marketing and artificial intelligence is one that is inevitable for all of us. It's an inexorable part of what we do. In order for us to be able to scale and to do things better, faster, cheaper, and with more impact, folks like Chip become inextricable to what we're doing. And so, without further ado, I want to introduce Chip LaFleur.

Chip LaFleur is a visionary leader, author, and technologist in digital marketing. Best known for writing Digital Marketing for Law Firms: The Secrets to Getting More Clients and Better Cases. Chip is the founder and CEO of LaFleur Marketing, a Grand Rapids-based growth engine partner nationally recognized for its data-driven, human-centered strategies that help law firms, healthcare organizations, and financial institutions grow with purpose and precision.

A recognized voice in the evolution of AI marketing, Chip has led the development of proprietary AI tools such as the AI Knowledge Base, Longest Tail AI, Clear Board, and others — platforms that transform data into actionable insight and creative direction. Through LaFleur One Inc., he continues to pioneer the integration of analytics, automation, and creativity to elevate how professional service organizations connect with clients and communities.

In addition to his engagements with Connectionology, Chip is a frequent speaker and industry educator, presenting for organizations including the American Marketing Association, GR Tech Week, Spartan Innovations, University of Michigan, and the State Bar of Michigan. He talks about exploring topics like AI adoption, digital transformation, and brand humanization.

A lifelong learner and technologist, Chip is driven by a belief that creativity, data, and purpose-driven leadership can transform organizations and most importantly, lives. Without further ado, I want to introduce Chip LaFleur who will take us into the world of AI and marketing.

Chips,

Thanks so much, Chad. I really appreciate it. I feel like my first to-do after this is to shorten my bio significantly, but I really appreciate you running through all of that.

Be sensible about what you share. Don't upload highly confidential documentation to ChatGPT or Claude. There's opportunity in Cowork to do more with company-owned information. Claude does offer a BAA for enterprise clients if they require it.

As it pertains to security, there's a tool called OpenClaw. It's a set of software that lives on a computer and will do things proactively for you. But that is also a security nightmare because if you install it on your daily computer, it has access to the accounts and files on that computer.

OpenClaw is open source software. You can install it and then download plugins and all bets are off. If you hear that someone installed OpenClaw on a company computer, they should take that off.

I look at it as a spectrum — on the far side you have things like OpenClaw which need very specific and explicit controls. Then you have things like Claude and ChatGPT where you need to check what they're using your data for, but company accounts give you coverage.

I think it answers the question. And for all of the participants, if you have a question, if you type it in the chat box, I will gently interrupt Chip and pose the question.

I'm not an attorney, so I'm limited. I think the courts are going back and forth on what is copyrightable. There's a ton of litigation we are yet to experience. You need to use your judgment.

Matt Ennis pointed out that attorneys may utilize such confidential tools, but a client who utilizes AI tools without attorney supervision may not be protected and deemed waived. This was from United States vs. Hepner, issued in February.

It's going to be harder and harder to not use these tools. There are ways to use them more confidentially. One is an enterprise agreement with the provider. Anthropic will provide a BAA for enterprise clients.

There are also localized environments. We have a client who built their own machine on premises with a 5090 graphics card for local AI processing.

In that scenario, they use their local AI to help with intake. It automatically transcribes phone calls, goes into their local model, does full transcription and highlights name, address, email, incident details, and populates those into case management.