General vs. Vertical AI: Why the future of LLMs and AI is dependent on the quality of locked training data and proprietary algorithms, and why How To SaaS built Inquisio
Imagine you are a neurosurgeon. You have a surgery coming up. Can you use ChatGPT or any other broad-based AI tool on the market to help you with your surgery?
Quick answer: Absolutely not.
Why? Because none of those tools have any of the following to give you a good output / support you need to complete your task:
They don’t have the training data of all the neurosurgeries to date.
They don’t have the context of each surgery type, patient data, specifics etc.
They don’t have the expertise to figure out how to solve detailed, complex problems with all kinds of edge cases.
Now, imagine an AI company that has access to the data of all the neurology procedures across all hospitals, categorized by type, with results and outcomes based on approach, along with patient data and specifics, plus they have the requisite processes / frameworks / context / expertise of those procedures. That company then uses that data to train its own AI / LLM.
Could this tool help a neurosurgeon? Yes.
This is the future of AI tools. While all the broad-based AI tools get the buzz, there is a massive opportunity for every business to build its own verticalized AI platform with locked training data sets and proprietary algorithms. The bigger your moat around the training data / algorithm, the more valuable it can be to a specific vertical / user / use case.
In the last year, major strides have been made by all the general AI platforms like ChatGPT, Gemini, Claude and even Deepseek. These platforms have been training on publicly available data from the general internet (YouTube, Reddit, Quora, Yelp etc.).
What is becoming clearer and clearer is that over time, these platforms will converge in their capabilities. It will be more like Google vs. Bing than one platform winning it all. Google will have its advantages with its lead in search and embedding Gemini into all its core services like Gmail and Drive, while ChatGPT will benefit from the Microsoft ecosystem. Each platform is vying for the General AI platform market share and each will capture a big chunk, along with the other tools in the space.
Meanwhile, the Vertical AI market is being largely overlooked and underestimated. Both Satya Nadella and Marc Benioff have recently stated that Agents will replace SaaS applications. As a concept, this sounds abstract. What does it really mean?
In most SaaS markets, 70-80% of features that are developed go unused by end users. This translates to billions of dollars in wasted feature development and engineering resources. The end user, in most cases, is trying to do a specific task. In a Vertical AI world, those tasks can be done through an agent instead of a user. When an agent exists, we just need to optimize development to enable the agent to get the answer:
Want a specific marketing report? Just ask the agent to produce it with the huge underlying data set inside Hubspot.
Want to figure out which sales reps are producing the most revenue and pipeline? Just ask the agent to produce a report with the data inside Salesforce.
The data is the product. You don’t need more features. Stronger proprietary data + AI Agent = Better retention and revenue.
Most businesses (SaaS, B2B, Services etc.), have some sort of proprietary data set (assuming you have the requisite permissions) that is locked or invisible to the general AI tools. This can be:
Data that sits inside your platform — this can be leveraged to create AI tools that enable them to do tasks faster.
Data that your platform outputs based on inputs from the user — this can be leveraged to create a proprietary algorithm / AI that allows for better outputs.
Data that your platform collects as part of working with customers — this can be leveraged to create better AI tools to help the customers self-serve / lower their cost base of working with you.
Beyond this is an important additional data set: your own proprietary process / insight into a market. This is outside of whatever data the client gives you. This is your core expertise that is embedded into your product or service offering. The more you are able to automate their process into a proprietary algorithm with proprietary data, the bigger the moat you will create for your business.
This last approach has its own set of risks:
You can cannibalize your own business by giving users a lower-cost, self-serve option to get a good enough outcome.
You distract your business by building an AI tool that may not have the impact you expect, while diverting resources away from core products / services that actually drive revenue.
The problem here is that one way or another, all businesses will be faced with this existential threat. Regardless of industry, someone in your market will create a solution like this to compete with your offering even if you don’t decide to invest in it. The cannibalization is going to happen anyway because that’s where the market is going.
This is why it is important to build the Vertical AI platform yourself, before the rest of the market catches up to your advantage of your proprietary data set and algorithm. The faster you move, the more impossible it will be for someone else to catch up.
I speak of all this because that is what we have been building How To SaaS for the last 6 months. (This example will also bring all of the above to life).
We do heavy marketing strategy consulting engagements with leading private equity investors. Think of us as a McKinsey for Marketing. Our three main solutions are:
Marketing Due Diligence: 2-week sprints to give PE investors a clear-cut idea of the marketing potential of a target investment.
Marketing Strategy Consulting: 3-4 month engagements with deep dives into new and existing portfolio companies of PE investors / founder-led companies to help them figure out how to scale pipeline and revenue from marketing efforts and how much budget is required in order to get to their growth aspirations.
Fractional CMO Services: 3-6 month engagements where we get operationally involved with the team, the CEO and the board to ensure the day-to-day and month-to-month operations are going smoothly until a new CMO is found. We also support the hiring process for the CMO and help onboard the person into their new role.
We’ve built a meaningful business with these three solutions over the last 6 years, with over 15 employees, strong YoY growth and high EBITDA margins.
Something I figured out early on is that the variable to growing our business is the number of Private Equity investors we meet on a weekly, monthly and quarterly basis. The more investors we meet, the more we get to educate them on the value of marketing, and the more deals we close.
Sidebar: This is something that I think a lot of companies would be well served to figure out. What is the one key metric / activity that, if you increased, would impact your revenue metrics across the board? Each business usually has one north star metric that can change everything.
To solve this problem, so far, we’ve invested heavily in thought leadership.
I’ve written two books (Post-Acquisition Marketing and Exit-Ready Marketing) that are heavily focused on how Private Equity and Founders can leverage marketing to scale enterprise value.
We run the Private Equity Value Creation podcast that we use to interview PE investors, advisors, investment bankers, lawyers and more. The podcast is one of the top 3 PE podcasts out there at the moment.
I speak regularly at Private Equity conferences to educate the market on the value of marketing as they think about increasing the valuations they are able to sell the businesses at.
I publish daily content on LinkedIn and YouTube to generate awareness of our offerings on social media.
One of the biggest challenges we’ve encountered in working with Private Equity firms is that the most common scenario where we are brought in to support an engagement is after an LOI is signed. The reason for this is that most PE funds are not able to invest too much money into outside service providers until they are confident that the deal they are investing in is going to go through.
Meanwhile, private equity firms still have a need Pre-LOI to do a bunch of analysis.
They need to understand the market and build their initial market thesis.
They meet tons of companies and need to evaluate potential investments based on limited information.
They have to build internal investment committee memos when they are preparing to submit an LOI to get approval from their GPs and LPs.
Yet, they don’t really have a solution that can help them figure out the marketing potential of an investment at this stage, even though it is one of the most critical levers for value creation. As it currently stands, they try to piece together disjointed data from Semrush, Ahrefs and other platforms as best as they can. But these platforms are:
Not built for the private equity use case
Lack complete context of what the investor needs to build an IC memo for a transaction
Have an inaccurate data set that needs a bunch of expertise to be transformed into something usable
This is why we built Inquisio. It is a self-serve, AI-powered, data-driven Marketing Due Diligence platform that gives investors the ability to understand the marketing potential of an investment in seconds. The kinds of insights you get instantly with Inquisio:
Understanding overall marketing budget required with benchmarking on program spend and headcount spend.
High-level organizational design, structure and headcount to benchmark the target investment’s overall team.
Understanding how marketing program spend should be deployed across channels and campaigns, along with expected ROI.
Benchmarking KPIs like conversion rates and CAC payback periods to figure out how efficient the target investment’s marketing activities are.
Competitor data to understand the overall market landscape and how to succeed in this space.
Overall ROI expectations for different levels of marketing spend based on target YoY growth rates and EBITDA expectations.
These kinds of data points and outputs are virtually invisible to most platforms because they do not have the requisite context of what PE firms actually need to make investment decisions.
Given the context of Vertical AI described earlier, we at How To SaaS were uniquely positioned to create this platform:
We have been working with hundreds of leading private investors, including firms like TA, HG, STG, Updata, Genstar and many others.
We have analyzed and made marketing strategy recommendations on thousands of different product lines and business units, across different industries, geographies, verticals, markets, average deal sizes, sales cycles, business models, sales complexities and much more.
We have developed our own proprietary process to rapidly assess and make recommendations on how to scale pipeline and revenue for all kinds of companies in a way that can be transformed into a proprietary algorithm for the Inquisio platform to use.
We have developed our own proprietary metrics and benchmarks to measure and scorecard companies across all domains to understand their marketing potential.
With these advantages, Inquisio is our answer to the existential Vertical AI problem in the market where Private Equity and Marketing intersect. Our hope is that Inquisio becomes part of the operating system of all Private Equity firms as they analyze, perform diligence and invest in companies at every stage of their lifecycle.
If you are a PE investor interested in a free 30-day trial, schedule a call here.
