Beware AI’s snake-oil salesperson

30 Jan 2024

Since ChatGPT was launched in November 2022, software vendors have shifted into overdrive to capitalize on the hype of generative AI. Business leaders have been flooded with offers – but are any options watertight?

Comments about AI’s potential impact, for good and bad, have fueled interest, understandably. For instance, Alphabet’s CEO, Sundar Pichai, claimed it would be “more profound than electricity or fire”.

That may be so, in the future. But right now, we are at the low end of the maturity scale. And AI sales pitches seem all too familiar: “technology easily and quickly transforms your broken processes and catapults performance to new heights”. 

However, peel back the label, and many overly optimistic solutions improve existing organization tools that – still – lack integrated data. The promises invoke Skynet, but the reality delivers only shinier spreadsheets. Or they erode differentiating service and customer trust through blind automation. Is that the transformation you want?

How, then, do business leaders spot these AI snake oil salespeople capitalizing on hype? 

First, scrutinize the targeted processes. Do they perform well already without extensive manual intervention? If not, new technologies will replicate existing dysfunction rather than revolutionize it. Optimize the process; don’t just automate a poor one.

Next, examine if the proposed tools enable integrated decisions using available data assets. If, instead, they create more disjointed datasets, prepare for additional complications. Siloed systems with siloed machine learning solve little.  

This integration matters because many current AI initiatives accelerate processes that are not worth accelerating. The most significant opportunity currently lies not in fixing broken processes but in discovering new sources of value. New technologies should integrate, not further fragment.

Integration remains crucial because of the long-standing divide between leadership intentions and operational reality. Leaders often believe processes perform better than they actually do, not realizing how much manual heavy lifting occurs below the surface. So, when sold new tools promising automation, they see an upside that only exists in their imagination, not practical daily execution.  

 

Repackaged optimization – not transformation

The same psychology applies to consumer technology – for example, the iPhone appears magical but obscures vast infrastructure enabling it. Business leaders must peer behind the curtain of their operations before determining if AI can genuinely help.

After reviewing integration, determine if the promised capabilities require bona fide AI or optimize long-standing tools. Can existing software like Excel – which, rather depressingly, remains the go-to presentation tool in global boardrooms – deliver similar insights once feeds connect to richer datasets? Many solutions remain glorified dashboards, not predictive engines.

AI promises a revolution but, at this initial stage, offers faster horses, not airplanes. Expect old automation to be repackaged before seismic shifts to autonomous intelligence.  

Review existing process maturity before adopting newly marketed AI tools. Perfected current capabilities often outperform half-baked bleeding edge functionality. Translate demonstrated performance, not far-fetched fantasies.  

Incremental adoption balancing ambition and reality proves most prudent. If simplistic, expect fast imitation dulling any edge. If truly transformational, brace for a bumpy multi-year ride. Study the road ahead and chart a balanced course before hitting the gas.

 

Beyond the hype

To spot snake oil salespeople, start by asking for specific use cases already delivering value at other similar organizations. Vendors often provide glossy vision statements with little backing evidence. Press them to get granular on the actual return on investment.

Also, carefully evaluate AI’s alignment with your differentiation and value proposition. Will service suffer from over-automation? What human touchpoints matter most to customer trust and loyalty? AI should empower workers to enhance their experience, not replace them outright.  

Repackaged optimization risks eroding the special sauce of relationship-based, high-touch service models. Leaders must dictate strategy and values while AI enables execution. Machines can’t determine what makes your business stand apart or why customers choose you. But they can undermine those sources of affinity.  

Today’s AI hype cycle resembles past technology bubbles where vendors capitalized on executive fear and uncertainty. Consider the dot-com frenzy or the Y2K bug hysteria. 

For the latter, as the year 2000 approached, consultants warned that computer systems would crash, given their coding tracked years in two digits. They sold comprehensive “remediation” despite the unproven risk, pocketing fortunes. When the calendar flipped with no meaningful bugs, those who had bought into the scare story will likely have felt shamed.

The AI frenzy brews partly from the same psychology: predicting an improbable but technically possible future of runaway algorithms usurping human judgment. This captures the attention and budgets of those selling protective tools and services. 

Present AI capabilities remain nascent – not the immediate existential threat some portray. But that storyline sells, so it persists. Therefore, leaders should approach AI with optimism but guard against overhype playing on fears. 

Beware those quick-fix AI peddlers less concerned with your competitiveness than their consultancy fees. Neurotic leaders make easy marks, eagerly handing over checks conjured from imagination, not rigorous diligence.

Instead, adopt a studied, prudent strategy that acknowledges AI’s constraints as much as its possibilities. Set ambitious destinations, then chart incremental, measured courses accounting for culture, change management, and capability building.  

The promised land awaits but will only welcome the readiest. Those falling prey to profit-seeking scaremongers will find themselves outsmarted by more clear-eyed competitors. Much like the old fable, the tortoise, motivated by reality, not fantasy, ultimately crosses the finish line ahead of tech-drunk hares.

Steps to see through the hype

1. Scrutinize integration capabilities and targeted process maturity

2. Determine if promises require true AI or just optimization

3. Review the impact on differentiation and value proposition

4. Demand quantifiable evidence like use cases proving value

5. Adopt solutions incrementally with prudent piloting

 

• Author(s)

• 

---