AI Will Not Figure It Out. That Is Your Job.

You are using AI. That's the right thing to do. You are enamored with how fast and intelligent it is. But you are getting inconsistent results. And you are worried it's your prompts.

You have been using AI. Getting inconsistent results. Pushing back when it ignores your instructions. Telling it the same thing ten times. And wondering if you are just doing it wrong.
 
You are not doing it wrong. You are misunderstanding what it is.
 
Michael Toguchi builds AI systems for universities, nonprofits, and some of the largest mission-driven organizations in the country. Stanford. University of California. Major associations with hundreds of thousands of members. Places where an inconsistent output or hallucinations are not just annoying, they have real consequences.
 
What he has found is that the AI problem is almost never an AI problem. It is a foundation problem. Organizations that struggle with AI outputs almost always have unclear goals, inconsistent data, and no governance before they touched that tool. AI does not fix the crack in your foundation. It runs on top of it. And if the foundation is broken, AI produces mediocre output at extraordinary speed.
 
In this episode of The Business Growth Architect Show: Founders of the Future, Beate and Michael dig into what it means to stop expecting AI to figure you out,  and start taking responsibility for what you are asking it to do. We talk about the comprehension shift that changes how you use every tool. The follow-up problem every business owner has and how AI can solve it at every level. When to build your own agent and when to hire a one-time build. And the data protection moves every founder should make before they go any further.
 
Michael also brings a framework most tech conversations skip entirely: the stewardship test. Did you design this system to help the people who will use it,  or to bypass them? That question, he says, tells you more about the health of your AI strategy than any compliance framework.
 
This episode is for the founder who uses AI every day and still does not fully trust it. Who has been frustrated by the inconsistency and is not sure whether the problem is the tool or the approach. The answer is in this conversation.

About Michael Toguchi

Michael Toguchi is the Chief Strategy Officer at eResources, where he leads platform direction for application management systems that streamline complex processes like scholarships, grants, admissions, and accessibility services. With over 25 years of experience driving digital transformation for universities, non-profits, foundations, and associations, Michael specializes in simplifying internal workflows to help mission-driven teams reduce manual work, scale sustainably, and strengthen compliance. His work powers organizations including Stanford, UC Davis, PG&E, the Roddenberry Foundation, and Google’s Certified Innovator Program. At the core of his mission is a commitment to building technology that enables teams to focus less on managing systems, and more on delivering meaningful impact.

Connect with Michael Toguchi

Websiste| LinkedIn| 

 #AIImplementation #MissionDrivenBusiness #AIStrategy #BusinessAutomation #HumanCentricAI #FoundersOfTheFuture #BusinessGrowth #FounderLife #AIForBusiness #BusinessPodcast

_____________________
We appreciate you, thank you for listening. Let us know in the comments what resonated in this episode, we want to hear from you.

Leave a comment, like, share with one person who needs to hear the message our guest shared.

Take our QUIZ and find out what your talent is worth in this market:  What's Your Talent Worth (http://WhatsYourTalentWorth.com)

Follow us on Instagram:
Check us out on Tik Tok:
Work With Us

Transcript

SPEAKER_00 0:00

AI has not at this point advanced uh to a stage where it can just be allowed to run amok by itself. You have to not only create the system and the workflow, you have to create the oversight, and these systems can either help people or they can harm people.

SPEAKER_02 0:16

So how do you make sense of it all and come up with a consistent product?

SPEAKER_00 0:20

One of the things that AI exposes is uh when an organization of any sort doesn't have good consistent data, you're you can end up with just some extremely, extremely poor product. Protecting that data is extremely important. It's still at this point a little experimental.

SPEAKER_02 0:36

Michael Dugucci from Tectonic with me. And Michael, my opening question to you is do we need to regulate the AI? Is that even a good thing? And if we regulate AI, who is responsible for doing so?

SPEAKER_00 0:53

Yeah, it's a it's a great question. Very timely for right now. Uh we've seen AI evolve at a breakneck pace this year. Um, you know, things that didn't even seem remotely possible uh even a few weeks ago are now happening, um, and everyone is just scrambling to try and keep up. And that creates this uh severe tension with uh institutions, organizations that want to have um thoughtful uh moral policies, uh they want to have governance that uh you know allows it to be implemented uh in in a good way. But it's hard to do that when uh what you're utilizing uh and how it's being used is changing on a day-to-day basis so quickly. So, I mean, from my side, I think the answer is yes, it needs to be regulated.

SPEAKER_02 1:38

And who is responsible for that or who should be responsible for it? Because right now the Edelman trust barometer came out. And we now know that 70% of us don't trust anybody other than our close buddies and people that believe in what we believe in. So now am I going to think that the government that I don't trust, or 70% of American people don't trust, now we want them to regulate the AI. Who who who could even regulate AI and what does it mean to regulate AI? Maybe we should start there.

SPEAKER_00 2:13

Uh-huh. Yeah, in the I from my perspective, you're looking at shared responsibility for something this large that impacts so many people. As a Alexis de Tocqueville type person from my policy background, I do believe strongly in self-regulation and that concept. And so the companies that own these properties that are selling, you know, whether it's uh OpenAI or Claude and those kinds of things, like you need you need the folks that are building and selling these things to have some responsibility in the process to the extent that they are putting down some guardrails for their own tools and their own products. You mentioned the government. Obviously, you're right, we live in a time where people, in somewhat reasonably so, don't have a lot of faith and trust in the government to be able to equity equitably implement things like regulatory regimes for big things. But I think there has to be a role in the same way that you have over the years when industries get too large and they're too big to fail and they're impacting too much, you have to have some government presence. That's the only way to have some level of standardization and uniformity. Otherwise, in the United States, you would have, you know, 50 different states with 50 different sets of policies that it would be impossible to start. It would be disastrous. Yeah. It'd be disastrous. So you want, you need to have uh you need to have the government to have at least that top level role. They need to have uh the ability to, you know, if there are violations of the of the regulations, they need to be able to enforce those. And so I do think there is a place for the government. So if you have companies at least providing some level of self-regulation, you have the government providing their their guardrails, where is the the missing piece of that puzzle is the companies and the people themselves? That they're the ones that are closest to the problem, they're the ones that are choosing which AI tools they're utilizing, and they are the ones that are building systems. And these systems can either help people or they can harm people.

SPEAKER_02 4:06

What role does AI play in your business? And let's talk a little bit about Tectonic and maybe tell us a little bit about what the business does and why AI is so critical to what you're doing right now.

SPEAKER_00 4:19

Yeah, I mean, Tectonic is a strategy and transformation company. We work with mission-driven organizations, uh, some of the largest um associations and companies, as well as higher ed, like uh, you know, Stanford and University of California. We work across uh across the spectrum. And um, we help groups with their digital transformation. We help groups with uh technology and security and uh help desk, but we also help them with our orchestrate platform, which is for application systems, workflow automation, business processes. Uh, and I oversee that part of the business in which we do things like create scholarship systems to provide uh educational opportunities, to uh provide uh services for disability centers in higher education, allowing uh equitable access to resources for students with disabilities and solutions like that, where we're often trying to help any kind of team or multiple teams, whether it's in departments, uh, find ways to optimize what they're doing, to uh to streamline and to uh to just sort of integrate, collaborate, work together. And that all sounds like a bunch of business talk there, but there is uh there is a meaningful and moral part of it. And uh it goes back to the idea of design a good system that's meant to help serve people and in particular for the people that are using it, allow them to do impactful work, allow them to get time and meaning in their day as opposed to wasting their time, sort of doing manual tasks. Uh, and that's where AI has come in and become an important part of uh of ways that uh the best of AI is something that takes away manual tasks, that acts as a sort of force multiplier for everything that you're doing.

SPEAKER_02 6:03

So, what it sounds like to me is that you've been really at the forefront of figuring out how to utilize AI for business automation and operational excellence. You know, I I do a lot of operational stuff. That's my that's my sweet spot. I love processes and systems. But I also know that, you know, number one, it's very difficult to implement and design these systems. And then you using AI to do it. But then the AI goes rogue. The AI goes rogue a lot because it hallucinates based upon the unbelievable amount of mediocre inputs that goes into the AI. So so you don't even have quality control. I started with ChatGPT. I thought ChatGPT was amazing. Now ChatGPT is so mediocre that I'm I'm rolling my eyes every time I'm using it. Now I'm using Claude, and I was really excited with Claude at first. And now I'm my my love story with Claude is is releasing a little bit because of what happened. Now they're taking their shitty, crappy inputs over to Claude. Now that's becoming mediocre. So how do you make sense of it all and come up with a consistent consistent product? And then we haven't even talked about regulations and ethics.

SPEAKER_00 7:18

No, I mean you've got it right. I mean, uh, one of the things that AI exposes is uh when an organization of any sort doesn't have good, consistent data, it becomes a bad input. And so if you task uh an agent to scan through your database to create letters or anything like that, um, you're you can end up with just some extremely, extremely poor product. AI has not at this point advanced uh to a stage where it can just be allowed to run amok by itself. And you know, for the kinds of organizations that we work, higher ed institutions that have very sensitive student data, you know, large membership organizations and their and their donor databases, same for nonprofits, protecting that data is extremely important. And so helping them craft a policy that ensures that an agent of any sort or any kind of built LLM isn't combing through that information, releasing it, you know, hallucinating and creating uh, you know, letters and outreach that they don't want. You have to not only create the system and the workflow, you have to create the oversight to ensure that it doesn't fail itself.

SPEAKER_02 8:19

Aaron Ross Powell So I think that the responsibility that you carry is is really great because we are building new systems on somewhat top of other systems that need to bring in pieces that we never had to think about before. So that's already a really tall order. But what I want to ask next is I want to ask a question about you. Because we talk about the show about spirituality, you know, really sort of being a driving force behind strategy. Some people say IA can be very spiritual and given the parameters. I've had some uh very deep conversations with AI about, you know, some of these things, you know, some of the programming is really good because it also has access to all that stuff. So how does your spir personal spiritual transformation interact with this very strategic and logical prompting?

SPEAKER_00 9:21

Yeah, that's a that's a that's a very interesting question. I mean, I as as a Christian, I try and have a servant's heart and try and apply that through work where I want to help organizations create, whether, you know, they do not have to be Christian themselves, but do they create a moral foundation that enables adoption, that creates trust with a team? And does it foster an environment of stewardship when they're using these kinds of tools to make sure that the systems are helpful and not harmful? And so that's something that I always try and apply on a daily basis, the idea of, again, that stewardship concept through Christianity, that we want to help that there's plenty of good working, hardworking, sharp people out there that are burned out and scared or that they don't know what to do with, you know, with AI, potentially having taking their job or making them learn and do things that they're not prepared for. How do we help those, how do we help those people? How do we design a system that helps them help other people and really feel like the work that we're doing is impactful? Uh, help them have meaningful, uh, you know, meaningful careers and things like that. And so that's that's sort of the approach that we take from or the approach that I take from uh from how spirituality sort of reverberates through the idea of these projects and engagements that we have.

SPEAKER_02 10:40

No, not everybody on this show has had this rock bottom transformational moment, but I think it's important that we talk about that sometimes your own simple moral compass rules, if you sell one for the for the lack of better word, can also be a pretty good foundation for for the work that you do. You had said to me, Michael, that you believe in simplicity. Do you want to talk a little bit more about that, what that means?

SPEAKER_00 11:09

Yeah, uh the idea that we what we do, um, and with the reason our team, which I think we have a great uh culture, and we've been around for uh much longer than many uh digital groups have been, um, we love problem solving and we love creativity. And so every challenge is different. Every, you know, organizations that are two or three people working out of their basement, guerrilla warfare style, as opposed to Fortune 50 companies, you're always finding a new way, a new set of people to connect with and help them. We like to take that complexity and find a way to simplify it. And then also help them, help the leaders that are there communicate about it, be transparent, be accountable. So, my our concept of simplicity is uh thinking about automation that especially in the framework of or the framing of AI, that there's all these legacy processes that people have built up over years, you know, so-and-so uses this tool and somebody keeps this in their email, and so on and so on and so forth. And, you know, for all of the unhealthiness of AI, at the very least, we can say it's exposing some of that, those broken processes. And so we're able to, in our conversations and our engagements, um, sort of break those things down and say, well, hey, this maybe this doesn't have to be as complex. We use simplifying as a way to sort of clear the noise for folks um and to sort of take away some of the manual labor that uh that burdens them and then create a good streamlined workflow that increases their collaborative environment.

SPEAKER_02 12:42

Can we talk about this maybe for a second? Because this is like something that I'm up against. So I like uh engineering processes and systems. I really do. And I find last week I've caught myself, you know, and I shared this with you when we were in the green room, that somebody called me on it and said, you know, why are you consistently improving processes and systems when you don't have validation that these processes and systems are producing the results? So can we talk about for our audience? So so somebody's heard this and said, fine, you know, we're 15 minutes into this conversation, then you know, AA, blah, blah, blah. Well, what's it gonna do for me? I'm a small business owner, I'm not a big institution. I need to figure out how I'm gonna make AI work for me. I wanna make it simple, but I don't know what that means. How do I know that what I'm using makes any sense? Can you give them some pointers?

SPEAKER_00 13:37

Yeah, I want to give them uh two pieces of potentially conflicting advice. Uh so that's uh it's always fun to start out that way.

SPEAKER_01 13:43

Oh, oh, then my favorite guess conflicting. Two things can exist at the same time. And then they can.

SPEAKER_00 13:49

They can. So what one of the things we would always tell people is you need to have clear goals. And part of having a clear goal is understanding your ROI. What's the return on this? If I spend five hours doing it, does it save me any time? If I do this, will it help me gain more customers? There has to be a way, a metric to evaluate whether or not something is successful. And so in that way, you can look at, you know, AI, there's AI, as we talked about at the beginning, like it's a million different things. It's generative, it's helping you write things. It's agentic, it's helping you uh scan through a list and then crawl through something and help you find a list of prospective customers. There's a million different ways you can utilize these tools. So, but instead of getting overwhelmed with it, you have to pick the places where you're particularly, if you're a small business owner, where are the places where you think you need automation? Where are you spending more time on manual tasks than you wish to? You know, where are the where are the areas where you know customers or revenues are not meeting the expectations? And that's where you should be investing your time to figure out it where AI can help you. So that's sort of the first thing.

SPEAKER_02 14:52

I mean, I I I tell you what comes to mind. Follow-up, follow-up, follow-up. How do we utilize AI for follow-up? Now that I have you here, I'm gonna put you straight on the spot. Because that's the number one thing. Everybody says they need more clients, they need more people to come in, they need more revenue, but they don't get to the follow-up because they have all these other really busy, important things to do. How do we how do we how do we do that? Where do we start?

SPEAKER_00 15:22

Well, let's let's say you had a HubSpot, Salesforce, any sort of database, or even if it's just your email, if you want to let Claude co-work go crazy within your own email, it should be doing things like creating tasks for you. Follow, oh hey, Mike sent you a message a week ago, you haven't responded. So it that's the kind of thing that can create reminders, create tasks for you. Uh, can also write the follow-up for you if you want, and then you can do oversight of the message. If you start thinking about how to scale that, and I mentioned HubSpot or Salesforce or a tool like that, it can start creating things like, oh, hey, if somebody has clicked through on my email twice, but they haven't filled out my form or scheduled a demo or contacted me, then I can automate some type of response for them to create a follow-up, say, hey, I noticed that you were interested. Or you can get really a little bit even more detailed. Let's have uh a pro write a program that follows that. Oh, hey, I looked uh at Mike on LinkedIn. It seems like he was doing this. Let let me personalize or hyper-personalize my follow-up to mention something that he's writing about or talking about to see if that I can gain his interest that way. So I think there are some very easy use cases. I mean, when I say easy, uh there are there are applicable use cases for creating follow-ups both at for the smallest, most individual level within your email, uh in terms of integrations with uh your your CRM or marketing tools, and you can be off to the races in terms of taking taking some of that follow-up uh off your plate.

SPEAKER_02 16:53

So in you know, I actually uh tried to do that. So I'm gonna tell you exactly what I did. So I went to Claude and I said, okay, so here are all the things. And Claude and I, we got along great, and we had this really great way on how to do my email. And then I just said, Do you even have the capacity to do it? And then it says, No, I can't.

SPEAKER_00 17:21

I l I love it. I mean, you know, it's brutally honest and just uh it's also feels like it's just you might be going crazy kind of reading something like that. Yeah, I I think I think to create what I'm talking about, the example I was talking about, you have to be able to create your own agent, you know, sort of outside the framework of something like Claude. So that that's what we would normally counsel with build.

SPEAKER_02 17:44

I I I I think I sort of I sort of gather that. Can I create my own agent? Can I figure out how to do that? Do I need to hire somebody to do that? So how do I start doing something like this?

SPEAKER_00 17:56

Yeah, I mean, you you can make one of two choices. Um, choice A is you decide to do it yourself. And I I think there's some merit to that. Like it's there are it it is overwhelming how much is out there, but there are straightforward training resources on you know how to get started with agents and use utilize free or inexpensive tools to build those kinds of things. If somebody doesn't have the time for that, one of the things that's nice about the environment we're in is everyone is scrambling. Like there are people out there who have figured this out, and technology groups generally are used to big retainers, time and materials, things like that. This is shortcutting the process. People with no coding experience can figure these things out. And though, even and even those engineers that have, you may be able to find a one-time resource that's like, yes, I can I could build something like that for you in a day. You end up paying for it, but you're not doing an enterprise level implementation that's you know five, six figures worth of investment. Uh, it's a one-time thing uh to get yourself started on. And so I think both of those routes have merit, just depending on. I don't even think it's a technical capacity issue, because I do think a person who is savvy and smart and kind of motivated with critical thinking skills can figure out building agents, but you're gonna invest the same amount of you might invest 10, 12, 20 hours trying to watch videos and experiment for yourself. Some people might enjoy that process and then they, you know, can kind of own it. But others might say, no, that's not, that's not a good trade-off. That's not good ROI for me. I would rather find somebody that can help me with this.

SPEAKER_02 19:31

Yeah, I think you're absolutely correct, is that um we have to, you know, make the evaluation or the calculation, what do you value more time or money? And then if your time is worth whatever, $250,000, $500 an hour, and you're spending 10, 20 hours on something like that, it is the equivalent of you spending $10,000 on having somebody really develop develop this for you. So as we're coming to wrap this up, uh is there anything you want to tell our audience that they should not be doing when it comes to AI implementation, integration, uh safekeeping, or ethical violations with the AI that we you if you were to tell them, what would you tell them?

SPEAKER_00 20:16

Uh, that's a good one. That I would think that we lead with the idea that a good transformation is something that creates time, the most precious resource, and creates meaning and stewardship. Uh, and some of these concepts that I've mentioned that provide that moral compass or fo philosophical foundation. I don't really think in terms of ethical AI, just I think in terms of like the human connection. And, you know, if you're again, when you're designing a system, like are you designing it to help your staff? Are you designing it to help the people that are going to interact with it? Or are you building something that ultimately is going to be, you know, harmful for those that that come in contact with it? So I do think that every time I talk about AI at this point, I always end up in uh data security policy and governance that everybody wants to be innovative. Everybody wants to try the latest and greatest thing, but you have to be careful. Like we were talking about should I let Claude comb through uh, you know, my entire email inbox? Uh, should I let people have access to all of my finance and contact records that allow, you know, it it's powerful. It may allow them to do some things that you didn't think were possible and help you, but it's still at this point a little experimental, and you have to make sure you you know what you're getting into uh to safeguard um your data, your, you know, your contacts data and things like that. So I mean, those are two of the things I'd say.

SPEAKER_02 21:40

Yeah, I like that a lot. I think that uh one of the ways I'm thinking about this is that it would be really advisable for people to have a personal email and a business email as an entrepreneur, because then you have different prompts for the different types of messages that are coming into your inboxes that require different type of safeguarding. Because if you send estimates and things via the business email, now you have a much higher exposure than you do when you forward your Southwest travel itinerary to your mother that you're coming, that you're coming home, right? That that is uh that is a different different type of a risk. So uh where can people find you and what type of clients do you serve?

SPEAKER_00 22:28

Oh, we like to serve mission-driven groups. Um, and however that's once someone wants to define it, it could be a nonprofit, it could be a member association, it could be um a school or educational group or foundation. So that we tend to work with groups that are you know mid-sized to small, um, but some of the largest uh organizations in those categories, we work with them as well. Yeah, it could be it it could be a two-week, uh $10,000, $20,000 project, or it could be a two-year, you know, six, seven figure. Like it that's what I said. Like we we really enjoy our culture of problem solving and um find real meaning in helping people serve. Uh, and that can be at a lot of scales. So I mean, there's something really fun about getting a group off the ground, and there's something really fun about helping someone that's already been there and done that. You can find me on LinkedIn. There's not a lot of Mike Taguchis that float around the uh that space. And our website is uh teamtectonic.com, where you can find out about our AI services, our our orchestrate platform. We do uh amazing uh marketing and branding as well, and uh have IT security. So um, yeah, that's those would be the two places.

SPEAKER_02 23:40

Beautiful. Excellent. Well, thank you so much, Mike, for being on the show today. I really enjoyed talking to you. My favorite language, tech language.

SPEAKER_00 23:48

Well, thank you so much. It's uh a pleasure to be on here and uh look forward to talking to you at some point.

SPEAKER_02 23:53

Thank you so much. And that is it for us for today. Please take really a close listen to some of the things we talked about here about AI. Uh it is still a little bit like the Wild West, so you must safeguard your information and what you put out there at all times. And if you have any need and assistance in having your AI projects come to fruition, please do reach out to Mike. And if you can think of anybody who can benefit from this episode, please forward them the show. And I thank you. I appreciate you. I thank you for your time and for your trust. And until next time, and goodbye. That's it for this episode of the Business Growth Architect Show, Founders of the Future. If you're done playing small and ready to build the future on your terms, subscribe, share, and help us reach more trailblazers like you. And if you're serious about creating, growing, and scaling a business that's aligned with who you are, schedule your uncovery session at uncoverysession.com. Lead with vision. Move with purpose. Create your future.