How to AI-ify a Business in 2023

AI is changing business at unprecedented levels.?????· 9 min read

In early December of 2022, ChatGPT has changed my perspective on AI and business. I’m electrified. And I may not be the only one.?

As a fascinated casual user of DALL-E before, I had sensed some of this coming. But I was very impressed and remain deeply inspired after I asked ChatGPT my very first question. It was related to our business. And I received a detailed answer that was both insightful and correct. Just a few moments later, ChatGPT wrote code for me, which could have been used in a connector for our software product.

As the CEO of a software SMB that works with customers in the Forbes Global 2000, I have not only been thinking a lot about what our business looks like in an AI-first world. I am also thinking about what AI means for our customers – what AI means for businesses in practically every industry.

We will be living in a business world that is pervaded by AI sooner than we thought just three months ago. AI-ification of business won’t start tomorrow. It is already reaching next level today. And it is applicable to every business we as a company are working with. All knowledge work will be touched by AI. And those who fail to AI-ify will be left behind in the long run.

We tried before

For years, AI has been on the R&D agenda in our company. We have looked at a lot. We have tried a lot. We have built several meaningful prototypes over the years. In our field of automating complex migrations, this has covered data profiling, mapping, project planning, project governance, etc. But frankly, none of the ideas that we tried with previous generations of AI stuck.

Some analytics came out of it, yes. A bit of big data, yes. A lot of data stuff, of course! But genuine AI, nope.?And we did not even think about everyday use by individuals in our company at the time.

Then came December of 2022. And all of a sudden, AI possibilities for us widened way beyond research and product development.

What AI can now do

With the pre-2022 AI that we had looked at – including “traditional” machine and reinforcement learning – we were supposed to train the AI from scratch and with our own data in most cases. Applications of AI at that stage were mostly pattern recognition, clustering, and matching. They were mostly discriminative or what I would call passive.?In some cases, even the term AI was a stretch.

Enter generative AI. Image generation with DALL-E gave me a hands-on foretaste. With this personal experience I had a first inkling, and a strong one, that what was traditionally seen as creative craftsmanship dropped in importance significantly. What, on the other hand, gains drastically in importance is inspiration and creative drive. Taste of and selection by humans still strongly matter.

I could sense that this would soon expand to fields beyond graphical design and the creative industry. Think legal expertise, which is a lot of text processing and generation. Its whole foundation is text-based in codes of law, written verdicts, and commentary. Think medical diagnosis – for general practitioners and specialists.

Then, all of a sudden, ChatGPT made it real and expanded the possibilities into innumerable fields. Text generation at that level is fascinating in itself. Think the tedious task of documenting anything. Think creating whitepapers. Think having summaries of specialist articles and books in any field at your fingertips.?

Yet this is just the surface of language-based transformers. It goes on with idea generation – brainstorming with a machine peer on pretty much any topic. There is writing and improving code. All of a sudden, this can boost productivity in any software development team, in any marketing department … for anyone who writes emails or works with any written ideas ever.

This is just what is available to the public in these very early days. Although we have technically already seen some progress in version numbering – GPT-3 or rather GPT-3.5 at OpenAI, Midjourney version 3 going into 4, etc. – and ChatGPT builds on the progression at OpenAI, we are basically looking at ChatGPT in its early beta (!) version today. The Jan 30 Version came out just before I started writing this.

Full disclosure, by the way: This text is still written in the old-fashioned way, word for word by a human being. This human being, nevertheless, has ChatGPT open as a kind of thesaurus while he is writing this.

As icing on the cake, take a moment to think about the fact that we are already at a point where AI is improving AI. Not only do I assume that every human working on foundational AI models is leveraging everyday AI capabilities for enhanced coding and writing of all kinds of texts. But an AI approach called AlphaTensor has found more efficient matrix-multiplication algorithms in 2022, which themselves are fundamental to the operational underpinnings of AI. These findings can and will increase the performance of pretty much any AI out there.

I’m tempted, but this is not the place to go into further topics like consciousness in AI, artificial general intelligence, AI embedded into physical everyday robots, and societal implications. Although I believe any of them have the potential to be huge.?

Back to business.

AI meets Reality of Business

The possibilities are mind-blowing and seem endless. You have 100.000 employees in your organization? Then 100.000 of them can already use AI meaningfully today – in a variety of ways across divisions and roles.

Yet, as always in business, there is way more to consider than the naked technology and tools. AI meeting the reality of business raises a lot of questions. Let me share a stream-of-consciousness set of them that are pouring out of my work reality. Feel free to skim through them at your own desire:

• What information do I want our team to feed to an external AI? Is it okay for them to paste our product code into ChatGPT? How do we prevent sensitive customer data from slipping into the external AI?
• Do I want to know if any text produced at work has been generated by AI or enhanced by it? Will people really seem smarter than they are?
• Can I insist on an improved level of internal and external communication – in grammar and style??
• How confident do I feel with an AI-generated business contract? How about an AI-reviewed one?
• To what degree is a person responsible for AI-generated output? To what degree does a person or several have to be responsible and in charge of AI-generated output?
• How can we embed modern AI into our product? Do we have to rebuild all or parts of our product? What additional modules and general functionality have become possible??
• Shall I insist on better code documentation, because the AI can do it quickly? Or is it less necessary, because the AI can explain code on the fly?
• To what extent will AI really improve productivity? Can our team become twice as productive with the same headcount? How can this be measured? What will it cost?
• How do we create a culture for AI adoption? How do we stay on top of a rapidly evolving wealth of AI tools – general and specialized – for all areas of business?
• Who needs to learn what level of prompt engineering skills?
• Will every single employee slip into a management role – as the manager of their personal AI team?
• What example do I set personally? How can I continue to become more productive with AI? Can I free more of my mental space to shift it to the most creative and inspiring of my responsibilities?

This is just the tip of the iceberg. There are deep legal, data protection, cultural, IP, strategic, organizational, HR, even moral implications to consider. Maybe there will be CAIOs in the future. Maybe there will be regular prompt engineering trainings. Probably AI interfaces will evolve beyond simple text prompts. Exciting times.

While by far not all of these questions have definite answers in our company, yet, let me move on to a structure for the continuing thought process around the AI-ification of a business.

Levels of AI-ification for a Business

With the evolution of AI and the broad capabilities it is unfolding, the possibilities of AI-ification now extend across several – I would argue: all – levels in business. They range from fundamental disruption of a business to enhancing everyday tasks of any employee.?

Let’s take a walk through the different levels. This may provide a simple yet systematic structure to address AI-ification in your business:

Disruption

Some industries and some businesses are being disrupted at their core by AI already. More will follow. Not all will be able to re-invent themselves to even maintain their industry. Service businesses will be impacted heavily. Not everyone in graphical design, for example, will survive this shift. And there will be more industries with fundamental upheaval and massive consolidation.

(Re)Building your Core Offering

Some businesses will have to re-think and re-build their business and offering from scratch to survive and thrive in their industry. The industry will stay relevant, but the big players will be reshuffled. Some new businesses will break into the market, as they build their offering on AI from the beginning. This does not mean they will necessarily have to build their own AI. Some, probably many, will build new businesses on existing foundational AI models.

Expanding your Portfolio

Maybe your core portfolio survives with minor adjustments or needs to be (re)built from scratch. In any case, AI-based additional offerings may expand your portfolio and even become the new stars, the new cash cows. Also, don’t underestimate the power of additional AI-based parts added to existing solutions.

Industry-Specific Standard Applications

Then there are industry-specific AI applications commonly used within an industry. Regardless if they are their own native models or an add-on to a general AI model, these applications digitally embody (if you allow the oxymoron here) industry-specific activities and processes. X-ray image recognition and diagnosis is one classical example of technology that could and should hardly work for one clinic only. It shall rather be used by many while possibly being supplied by only a few specialist AI providers.

Keep an eye out for such emerging applications in your industry. Or build one.

Next-Gen … AI-Gen Generic Business Applications?

Software applications that digitally embody rather generic business processes and are all built on a solid AI foundation. Such business processes are summarized under the abbreviations of CRM, HCM, ERP, SCM, and the like. Some experts expect a disruption of this market itself. But regardless, there will be providers for software that supports such processes.??

The key for you will be to watch what is happening in that market and what to adopt. The successful adopters of the best AI-driven providers will reap the benefits in these processes. There are certainly some players who have put themselves in a very promising position for now. We are watching closely.

Personal Productivity

Adoption of generally available AI technology on a personal level. Ensuring the means and creating the culture to enable that are key. This may not even give you a competitive advantage in the long run, because it will be the new status quo very soon. But it is essential. And it is essential across the org chart, on all levels, including top management.

Your teams are using digital devices in their everyday business today instead of paper files and letters. In the same way they will be AI-ified tomorrow. The shift will be much quicker. And it is important to not fall behind.

This last level was the biggest eye-opener for me at the end of 2022.

So, you certainly don’t have to be an AI business to benefit from AI. You don’t have to be an IT business to benefit from it. Not only will all knowledge work be touched by AI, as mentioned before, but most likely will all of it be transformed by AI.

Only the Beginning

This is work in progress. This topic and this market is bustling and blooming. And this is only the beginning. conemis and I are still experimenting and learning ourselves.

If you are interested in this topic, I invite you to join the conversation and share your experiences. I am happy to read your thoughts in the comments. And let me know if there are topics that you would like to read more about – in AI, in business, in innovation.

More will be coming for all of us. A lot more.

From the Series "A Moment for Innovation". First published on?Medium.