The psychology of self-deception can be potent, perhaps even insidious.  When you're looking at an established, often modest, software business that’s suddenly convinced it has found the philosopher's stone in the form of a new module, it’s almost quaint at first. But quaint turns dangerous when confidence outruns data, a phenomenon, I think, that the arrival of sophisticated AI tools has made considerably easier to sustain.

You see these a fair bit. Decent little businesses, ticking along in a niche, maybe making a comfortable living for the owner. They’ve got a solid, if unspectacular, customer base. Revenue is flat, or growing at a snail's pace, often for years. Then, when the conversation turns to a potential sale, out comes the "transformative module."

I recall one such case. Let's call the company "SteadySoft." They’d been in their vertical for nearly two decades. Reliable product, loyal customers, but sales had been bumping along the same low ceiling for the better part of a decade. We were doing our usual due diligence, understanding the core business, when the owner, let's call him Arthur, unveiled "ModuleX." His eyes lit up. This, he explained with fervent conviction, was the game-changer. This was going to take SteadySoft from a $2 million-a-year operation to a $10 million, maybe $20 million, player. The hockey-stick projection chart for ModuleX was a thing of beauty, a steep ascent into the stratosphere, leaving the gentle slopes of the core business far behind.

The pitch for ModuleX was passionate. Arthur genuinely believed it. He'd poured his heart, soul, and a good chunk of the company's (modest) R&D budget into it for the last couple of years. The problem wasn't his belief; the problem was how that belief was coloring everything, especially his valuation expectations for the whole company.

The Patterns of Self-Persuasion

When you dig into these "transformative module" narratives, you start to see familiar patterns of self-persuasion:

1. The Past is Another Country: Suddenly, the fifteen years of flat-lining sales are dismissed as ancient history. ModuleX is so revolutionary, apparently, that the company’s established performance, its ingrained sales processes, its existing customer inertia, are all irrelevant. "This changes everything," they'll say. Does it, really?
2. Aspirational Adoption Rates: When you ask how many existing customers have signed up for ModuleX, or even committed to a pilot, the answers often get a bit… fuzzy. "Oh, there's huge interest." "We've had lots of positive conversations." "A few are testing it." But firm commitments? Signed contracts for the new, premium-priced module? Often, they are conspicuously absent or very, very few. The projections are built on hope, not on hard purchase orders.
3. Underestimation of the "Lift": Selling a new, potentially complex, module into an existing, often change-averse, customer base is hard graft. It requires a different sales motion, perhaps different skills. The assumption that because customers like the old product, they'll automatically flock to the new, more expensive one, is a leap of faith. And if it's meant to attract new customers, then you're essentially launching a new product into the market, with all the attendant risks and costs, but often without a new product launch budget or team.
4. The "Sunk Cost" Fallacy, Glorified: Arthur had invested so much in ModuleX, emotionally and financially, that to admit it might not be transformative was unthinkable. It had to work. So, every polite nod from a customer during a demo was interpreted as a pending sale. Every "that's interesting" was mentally logged as "they're in."

It's not that new modules can't be valuable. They absolutely can. Incremental improvements, new features, add-ons: these are the lifeblood of a healthy software business. But the narrative of a single module suddenly catapulting a historically slow-growth company into hyperdrive needs an extraordinary level of scrutiny.

This is where, again, you realise the easiest person to fool is yourself. Arthur wasn't trying to deceive us. He had thoroughly convinced himself that ModuleX was the key to unlocking a valuation that the core business, on its own, simply couldn't justify. His pitch was less about the existing, proven (if modest) business, and more about the unproven, high-potential (but high-risk) future he’d pinned on this one development.

The easiest person to fool is yourself.

A More Rigorous Approach

Our job, then, isn’t to crush dreams, but to separate the dream from the P&L, the aspiration from the auditable. The mental checklist adapts:

1. Baseline Reality: What is the sustainable performance of the core business, stripping away the ModuleX projections for a moment? That’s your anchor.
2. Evidence of Module Traction: Not "interest," but actual sales, commitments, or at the very least, advanced-stage pilots with clear success metrics. How many, and at what price point?
3. The "Why Now?" for Transformation: If this module is so obviously brilliant, why wasn't something like it developed and successful years ago? What has fundamentally changed in the market or the company's capabilities to make this the moment?
4. Risk Assessment of the Module Itself: What if it takes twice as long to get traction? What if adoption is half of what's projected? What's the downside to the core business if resources continue to be poured into a module that doesn't fly?
5. Seller's Detachment (or Lack Thereof): Can Arthur discuss ModuleX with a degree of objectivity? Can he acknowledge the risks and uncertainties, or is any questioning met with unwavering, almost defensive, faith? This is a big tell.
6. Capability to Execute the "New": Does the existing team have the skillset, the bandwidth, and the sales & marketing oomph to actually make this new module take off in a transformative way? Often, the very stasis of the core business suggests limitations in these areas.

We've walked away from deals where the core business was okay, but the valuation was being driven sky-high by an unproven "transformative module." The sellers had looked at their creation, fallen in love with its potential, and priced the entire company based on that dream fully materializing. They’d forgotten the years of hard-won, steady, but unspectacular, reality that formed the bulk of the asset.

I find myself thinking about this in relation to the AI tools now available to almost everyone involved in a business transaction: buyer, seller, and intermediary alike. A large language model, asked to validate a growth narrative, will do so with remarkable fluency. Feed it the hockey-stick projections for ModuleX and ask whether the assumptions are reasonable, and it will, in all likelihood, find ways in which they could be. That is not dishonesty on the part of the tool; it is simply what a system trained to be helpful, and to generate coherent reasoning, naturally tends to do. The difficulty is that it can feel, from the inside, indistinguishable from genuine rigour. Arthur's mental checklist, already tilted by sunk cost and emotional investment, can now be validated by a system that will produce three convincing paragraphs in support of almost any position you choose to put to it. The antidote (applying the checklist questions to your own operational data, not to a tool that has none) remains precisely the same. But the temptation to outsource the scrutiny has rarely been greater.

Richard Feynman (Nobel Physicist) was right. It’s easy to construct a narrative that makes you feel good, that justifies your efforts, that paints a rosy picture. And when you’re selling your business, that narrative can become incredibly powerful, especially to yourself. The challenge for the buyer is to see the whole tapestry, not just the shiniest new thread, and to understand if the weaver has perhaps become a little too enchanted by their own design. In an age when AI can generate a polished investment thesis from a single bullish assumption, that challenge has not diminished; if anything, it has sharpened considerably.