AI with Humans will Substitute Humans with AI: Why CEOs Must Build Substitute Businesses.

Karim Lakhani , Co-founder and Chair at D3, at Harvard wrote a seminal article, “AI Won’t Replace Humans — But Humans With AI Will Replace Humans Without AI.??

To add on to what Karim says, AI centric businesses built by humans will replace humans that will use AI to augment work.?

The business that is AI at the center, designed and managed by humans is a fundamental substitute to businesses that are humans at the center supported by AI.??This is true generally, but it is urgent for labor/ professional intensive businesses.

I think most CEO’s in impacted industries know it, but are hesitant to make the tough political choices. 45% of CEOs believe that their company will not be viable in 10 years on the current path due to AI (PwC). At the end of the article, I outline the common red flags CEOs should watch out for when responding to Generative AI.

The last 10 years were the era of asset-light businesses . The next 10 years will be the era of employee-light businesses and of AI intensive substitutes. And, most labor intensive companies are not ready for it.

Case in point: Klarna showed us recently, tackling 2.3 million conversations, two-thirds of Klarna’s customer service chats are estimated to drive a $40 million in profit improvement (driven by 700 fewer support staff) to Klarna in 2024.

A new vendor is selling AI agents to Klarna is substituting or reducing the footprint of a customer support vendor. Those support vendors like Teleperformance and Concentrix saw their stocks fall after Klarna’s news. A startup in Cerebral Valley funded by Sequoia or Andressen Horowitz is going to pick a business model that replaces humans every day over AI that augments humans. This story is just the start.?

Reason: Whenever a long term technology shift happens, the instinct and the trap for large enterprises is to think of all the ways this shift could enhance their existing operations.

The questions that CEOs and boards typically ask are:

• How can we protect and improve the current business infusing them with Generative AI?
• What is the role of this Generative AI in making my employees and current operations more efficient, data driven and insightful to serve my customers better??

These questions take you down the road of what HBS professor Mike Tushman calls, "complementary innovation." This is an improvement or addition to an existing product or service. It bolsters the firm's competitive advantage and may increase marketshare.

Much of the discourse and investments around Generative AI and AI in general by the large incumbent enterprises is focused on augmentation. While that continues to be a great approach, not investing in parallel in AI centric substitutes is a mistake.?

The challenge with complementary innovation is that it is a short-term answer to a long term technology shift. The alternative is to generate substitutive business - products and services that fundamentally change or challenge the basis of competitive advantage thereby reducing the customer preference to existing offerings and shifting them to new offerings.

Every client is now looking at AI based models that deliver the Klarna style impact when looking at the vendors that people intensive activities.

When faced with technology shifts, the playbook is kinda the same: Create some outposts in Silicon Valley. Create a Corporate Venture Capital fund to learn from what's happening in startups. Find an Entrepreneur in Residence. Run some innovation challenges. Build partnerships with tech and AI companies, and shoot out some press releases.?Meet once a week to take stock, and spend rest of the time on business as usual.

Case in point: Daniel Julien, CEO of Teleperformance, said: “Even the most high-tech or the most AI-involved companies are clients of Teleperformance. We chose that there is a complementarity and not separation ,” Julien told CNBC.

Why would Klarna maintain its spend with Teleperformance or another call center vendor, when it has an AI team or a startup that can deliver efficiency without sustained labor costs??

To pre-empt, CEO’s have to ask the uncomfortable substitutive business questions:?

• What does an AI centric offering look like to serve our customers needs??
• What is a new business that we can create that can serve my customers that could compete and substitute my current business model?

The ideal solution for Teleperformance would have been to build the business that is now substituting their current offerings.?So, why did they not?

StraightTalk CEOs to Drive Substitutive Businesses:?

As a CEO, it's your responsibility to ensure that your team is either creating substitutes for your business or actively seeking out substitutes being developed in the market.?

This team should not report into your Business Units (BUs) and report directly to the CEO. BU leaders are incentivized to maintain the current book of business, minimize friction, and run a tight ship. While that's their job, the CEO's job is to build for growth.?

Embrace the possibility that the new substitute businesses may compete for the same clients – and it will be disconcerting when two of your units are facing off against each other. But, better than a new type of competitor that you can't compete with.

The role of the CEO is critical to manage this creative conflict .

As Tushman, Binns and Smith say in their article, Ambidextrous CEO , “Innovation’s only friend is the CEO”. The article goes on to say,

“Firms thrive when senior teams embrace the tension between old and new and foster a state of constant creative conflict at the top.” When CEO’s cede this responsibility to business units or to a Chief Innovation Officer, the new business loses their biggest friend while negotiating the large firm - results are disastrous

CEOs should create a substitutive business unit, outside the core, to build the new AI centric business. Here are some key red flags to look out for they might be messing up:?

1. If you, the CEO, are not spending dedicated time learning, exploring, experimenting with Generative. Your fingers on the keyboard. Pondering about what to ask, how to use generative AI, how to use it for your work everyday - and reflecting on an AI-centric vision of your company.
2. If you believe substitutes can be built for your current business without any internal or market friction.
3. When you, as the CEO, do not spend real work time with the builders but primarily interact with the leaders and managers.
4. If you exaggerate risks using anecdotal examples like Air Canada AI agents as a reason to tread very slowly into Generative AI.
5. Instead of majority AI engineers and vertical experts, your team dealing with Generative AI has more leaders and middle managers that push PowerPoint.
6. When the people building the AI centric substitutes are successful internal leaders best at navigating internal complexity not at building new businesses.
7. If the Generative AI investment you just approved has all been allocated to BUs and shared service organizations and none to building substitutes.
8. When thinking about AI centric substitutes, you mostly source ideas of applications to build and not foundational systems and infrastructure you will need to build 100s of such ideas fast.
9. If your substitute business builders spend a significant amount of time managing internal bureaucracy and not building and working with clients.

Building substitutes to your beloved business is not a comfortable process, but it is necessary.

Startups that IPO typically take 5-6 years on an average to get there. I guarantee that that pace will be faster for AI companies building substitutes. You don’t have much time. Get going. You don't want to be that CEOs on whose watch, the obsolescence started.?

Roll up your sleeves, build that team. It's on you.