Transcription
In this video, I'm going to be breaking down exactly how to perform a $15,000 AI audit as an AI consultant. So, if you are someone who is wanting to make money in the AI revolution as a nontechnical person, then this is for you. And if you are a technical person and you're wanting to land bigger AI development projects, then this will also position you to do just that. So, let's dive right in.
See, the AI audit playbook is an ROI first diagnostic framework. Okay? Because 80% of companies that have tried AI up until this point, and MIT did this study, failed to deliver meaningful ROI for those businesses. And that's not AI's fault. Cuz this is what it typically looks like. The founder sees something, they get really hyped about this AI solution they've seen. They get the team really excited about it. They get the tool plugged into operations. They get a developer to build that very specific thing. Everyone hopes it works. And then reality sets in. Three months later, the tool is collecting dust. The team reverts back to manual because it's what they know best. And the company concludes that AI doesn't work for us.
However, the top 20% see well north of 100% ROI. They don't need to hire anyone else new and they have unlimited bottlenecks essentially being removed. Hundreds if not thousands by AI. Here is the simple difference between these two groups of people. Successful teams start with diagnostic and not technology. So they find the expensive problem. They then quantify the cost and then they prescribe the AI solution.
So, in this video, we're going to be breaking down my six-step framework. So, we define outcomes first. We then map the business. We then map the tasks. We then find the AI opportunities. We then prioritize them. And then we calculate the ROI associated with them.
So, it starts with step one. Define clear and measurable outcomes. So, we need AI or we want to automate our operations. These are not outcomes. These are vague wishes at very best. Okay? So, if you can't measure it, you can't improve it. And if you can't quantify it, you can't prove ROI. And if you can't prove ROI, you don't get paid or worst case, you have a very unhappy client at the end. So every outcome must include the current state, the target state, and the measurable delta between those two things. So here's what real outcomes look like. Three to four hour lead response times and get to under five minutes. A 20% conversion rate and get it to 24%. 12 hours a week on repetitive work, zero hours or one hour. Very easy to track. 0% after hours call coverage, 100% coverage. Real outcomes that can be tracked that can be measured. Okay.
Outcome forcing discovery questions. So throughout the AI audit, the key component of an audit is the interview. So this is where we are understanding from both stakeholders and employees what actually they do on a day-to-day basis from making decisions to getting the task complete. We have to ask very specific bleed, money questions through the interview process. These aren't casual chats. They're not your mates. You're not trying to get to know them. You are trying to uncover and unpack the key bottlenecks and hurdles within their business that you can solve with AI. So, stop asking clients what to automate. Start asking where they bleed money. So, what's the most expensive problem you're dealing with right now? What's the one thing if we could solve would improve revenue? Where are you losing money that you know about but you haven't fixed?
Every outcome. So, every one of these questions, every question you ask, you need to tie a dollar amount in. So, even if it's, oh, Sandra spends five hours a week putting our books together and then Sandra gets paid $50,000 a year, you can tie an amount to those hours that Sandra is putting into that task. So, every single thing that you identify, you need to associate a cost to it. Okay? Because it might not just be as simple as they're wasting money on software. More than likely it is hours being wasted by staff and therefore you have to associate their salaries and the hours wasted to that.
Then we're going to map out the business. Okay. So the wrong way to do this is talk to the CEO, get a high-level overview and start building. This is exactly the reason 80% of AI projects fail because the CEO goes, "I just want to do this," and there is no rhyme or reason. Then he thinks it's cool or he knows this is going to work, but the underpinnings are wrong. Okay. The right way is to interview the executive sponsor. So this could be the CEO is the person who wants to do the audit. Okay? So there's usually one person really leading the charge on this. You'll then conduct 10 plus interviews. It can be 20 plus. It depends on how big the company is. If it's smaller, it could be five. Bigger, it could be 25. But you need to speak with heads, stakeholders, and employees. So individual contributors. You need to know everything from idea and decision all the way through to action being taken and the results being delivered or not delivered, obviously, in some cases.
So a great technique for this is the yesterday morning technique. So never ask, "What do you do here?" Okay? Instead, ask, "Walk me through yesterday morning. What did you do first? Then what?" Get them to do it step by step. So this is what an actual business map looks like. So I tend to break every business down into three core elements, which is acquisition, delivery, and support. So we are grouping together the processes. So in this case, we've got lead generation and capture. We've got qualification, discovery, and closing. We then go to delivery, which is pre-delivery, setup, strategy, production, post-production, client delivery, and then we've also got our support. Done it. We want to highlight the major components that go from end to end to every business and then we also want to highlight any time, risk, sync, stuff like that within the organizations. And then I also tend to like to put a key summary of insights I've gained from these business maps.
So let's look at a real-world example of breaking some of these processes down into specific tasks. Okay, so the VP of the sales department said, "Our reps are selling all the time." We spoke to the reps and they all said, "Between 9 and 11, we have to manually do the prospecting." So they have to search LinkedIn for prospects. They cross-reference for duplicates and then they manually copy and paste the fields into the CRM. They were wasting two hours per day building their lead lists for the day. And when it comes to sales and marketing departments, solving stuff like this is even more valuable to companies because those are the two departments that actually have an ROI attached to their role. So both marketing and sales departments are essentially performance roles. So if you give them back two hours of time, they can go and sell more people. So it's even more important to solve these tasks. So we would break this down. So say they open CRM, they fill the last 24 hours, they check ICP matches, they read submission details, they draft a response, they send an email, send follow-up. This is what a task would look like if you broke it down step by step.
So let's look at the lead generation one that we just looked at. So the SDR gets a target profile. It's strategic work. They should be doing this. Then they search LinkedIn manually. They find prospects. They open HubSpot. They check for duplicates. Then they manually enter the information. All of this stuff now can just be done by AI. They do not need to be doing any of this. AI and workflows can do every single one of these things. They can then get added to a sequence which again trigger through a workflow and then they start doing their outreach and they start jumping on calls. So this two-hour block of time these SDRs now have back.
This is another example, okay, for researching deceased property owners, uh, for a company to purchase these properties essentially. So they would pull uh county tax records. They would filter by value, the taxes owed, individuals versus company, ancestry databases, obituary databases, searching public records. Then they would have to cross-reference common names, identify them, and then find the documents and then move on to the next lead. And they were doing this many times per day. So steps eight and nine were 15 minutes per lead, just these two. So, if we solve this and they're doing even 10 leads a day, which they were doing a lot more than, but even if they did 10 a day, this would save an, what, an hour and a half on just that. Crazy, right? This is why AI makes such a big impact. Make sure smash that subscribe button, hit the notification bell if you love this type of content and you want to make money with AI.
Next is identify what AI can do and what AI cannot do. And it's really, really important. So, we want to break down and look at each of these individual tasks. So, step one is, is the input data structured? So is it a form, email, a database? This is good. If it's a vague verbal request, that's bad. Is the output predictable? So is it a standardized response, a classification, or is it just some creative strategy? Again, vague, right? Are decisions rule-based? If-then logic, or is it just a judgment call from an individual using experience? And then is it repeated often enough? Is it a daily, bi-daily, weekly task? Okay? Or is it like a quarterly one-off, something once? If the answer to all these four is yes, this is something AI can do and AI can handle 80% plus of the work and you can leave humans do the last 20%, the judgment calls, which is the important bit. Right?
So this is really, really important because people try and use AI for things that they shouldn't. So this is where AI shouldn't be used. So conflict resolution, anything to do with clients, staff, HR, anything like that. You can use it to streamline your job, but you never want it to do anything that a human should be doing. This requires empathy, nuance, and human judgment. Okay. Strategic business decisions. Don't let AI start making judgment calls for your business. It's great to have as a, a bit of a co-pilot to, you know, throw your ideas and bounce against the wall, but ultimately it will never have as much experience and context about your business and the industry and the environment you've worked in for all these years. That is where you come in. That's your value. Highly unstructured request. So AI is great at structuring data, but it's kind of a totally different process. So if the data always just comes in, it's totally crazy and unstructured, AI does a bad job at this. You would have to build a separate engine to actually take the unstructured data and make it structured. And then rare and low-frequency tasks, unless you have a big company like, let's say you have to do quarterly earnings reports because you're an IPOed company, that would still be valuable and yet it is low frequency. But for most companies, they're not doing very high-value, low-frequency tasks like that. So if it's low value and low frequency, just don't even bother, just have a human continue to do it.
Once we have these, we're then prioritizing and so this is called an opportunity matrix. So we have quick wins, which is high business value and easy to implement, usually low cost too. Then we've got big swings. So this is where we go to next, which is still very high business value, but it also has a higher cost associated to it. Then we have the nice-to-haves, which is low value but also low cost. And then we've got de-prioritize, which is high difficulty to do but low value. So we never do these. We always focus every engagement on doing these as fast as possible, usually within 90 days. Sometimes we'll even throw in some of the nice-to-haves if you know they want them. And then what we're really using the quick wins for is to catapult us into the bigger projects under the big swing section.
So then, once we've identified them, we want to actually calculate the ROI. And this is what I said to you earlier, why it's important that we know how many hours and the salaries of these people who are doing these tasks because then we can associate ROI. So two hours per day from these SDRs, eight SDRs, right? So that's 16 hours per day. 260 days a year, $40 an hour, it's $166,000 a year being wasted. So when you present that and you say, "Hey, it's going to cost you $60,000, $50,000, which is under 30% to actually deliver a solution that these SDRs never have to do that stuff ever again. It's totally autonomous. It is a no-brainer for a business. And the bigger the business, the more of a no-brainer it is."
So let's look at some real-world results for the property management company that we've done some work for. So lead response time could still be done manually, was three to four hours. Is now under 18 seconds. It could be even shorter. It could essentially be instantly. Total game changer. Increased happiness of clients massively, too, and prospects. Conversion rate 20% to 24%. That's just great. Instant ROI, $161K in additional revenue. Tenant communication, slow, manual, you know, Doris is off sick, it wasn't happening. Minus 68% time. The other big thing to track here is CSAT scores. So customer satisfaction scores, right? And this is huge. So especially if you have a recurring service. So property management is a great example of that. The higher quality your communication is, the lower the response times, the higher your satisfaction of your clients. So that's just all around good for your business. Your net promoter scores go through the roof. And then maintenance coordination. So manual tracking, minus, and then go to minus 58% on the time. It was $12,500 saved. Also with maintenance tracking, what we ended up doing was certain issues, they didn't need to call out, you know, the plumber to fix the thing under the sink. We could actually just say, "Hey, if you look under the sink, does it just look like it's coming out there? You might just be able to screw it with your hand and it tightens it." So, there was a genuine ROI saved, not only a lack of complexity, but an ROI. So, it was my, they saved $215,500. The ROI in year one was 166% and they had a 9.2-month payback period on all the implementations done. And if you are any serious business owner, you would know that anything under a 12-month payback period on operational changes or anything like this is immense and people would jump for opportunities like this. And this is what AI can deliver.
So how to scope out an audit. So for small businesses under 50 employees, you're going to focus on the quick wins and then you can use this as a springboard into the big swings because essentially you've saved them or made them enough money to pay for the next step, which is great. You've got immediate ROI emphasis. Quick wins can usually be done in under 90 days. You've got tactical recommendations. So this is what we're going to do to have very immediate impact. And we've got fast execution timelines. When you're dealing with mid-market and enterprise, 100, 300, 1000 person companies, it's a lot of strategy and education, cross-team alignment, road maps before implementation because there has to be a lot more adoption. And something that very few people talk about is change management. Okay. So one of the, the telltale signs of AI solutions not working within companies is the lack of employee adoption within or buy-in, basically, within the engagement that you've had. So if within the auditing process, they feel that essentially they're going to be replaced by enabling you to build solutions for jobs that they do, and then essentially what's going to happen is when you deliver the AI solution, it's not going to work because if the employees don't want to use it, they're not going to use it. So change management is hugely important. It's something almost nobody talks about.
So this is the breakdown of a two-week opportunity assessment. So first week is all to do with interviews. We're identifying the waste. We're mapping the workflows and we're mapping the tasks too. Week two is all about building the opportunity matrices, validating it with all the stakeholders. So getting their input and saying, "Hey, is this right? Like, this is what we identified. Is this what you want to focus on?" We then pull together the top three recommendations and then we build out the 90-day road map. Typical pricing anywhere from $5, $10, $15,000 for an audit on the lower end. We've done audits upwards of $50,000 on hundreds of person companies. Again, totally numb.
So building the automation is the easy part. There is thousands of people that went down the technical route when AI first came onto the scene. But deciding where to put AI actually determines its success. And that's what businesses actually want. They want successful deployments. So if you like this video, make sure to smash that subscribe button and hit the notification bell. If you want to start your own AI consulting firm, first link in the description. If you want us to do an AI audit on your business, second link in the description. And I think you're going to love this video.