Using ChatGPT and Bard in your company

Industrialization 5.0 is here and there are new opportunities but also challenges

Industrialization 5.0 does not only mean pattern recognition and automated knowledge generation, it also means that new tasks fall to humans when ChatBot technology is used. New roles arise for human employees when you consistently question everything, construct special cases, and make use of contextual knowledge logic and emotionality.

Humans will still be in demand, even if Sascha Lobo thinks: “But we can already see the outlines of a change that is likely to be as large and intense as industrialization.” Bots like ChatGPT are enabling AI in vest pockets. But just as with algorithms, it’s a matter of tracking down what’s beyond the patterns and contextualizing that – with humans – when using the bot with AI.

For leaders, that means:

  • Do our services actually still aim at the right solutions?
  • What solutions should our company offer?
  • What problems do we see for which we offer solutions?
  • What are others doing – can we do it differently?
  • Do we have enough data that we can evaluate to answer these questions?

It is the culture that decides

Getting to the bottom of such issues to strengthen the company’s raison d’être and ensure market success means cultivating new forms of collaboration. Once problems and solutions to an (assumed) customer need are identified in a data-driven manner – we soon need employee talent. But in their different roles they can offer as talents, not as a person with a task in a fixed organizational chart.

Data mining is no different than what bots like ChatGPT and Bard do. With two differences: it’s generic data from all the knowledge available online, this currently “ends” in 2021, and the bot crawls for probabilities and speaks and understands natural language.

So what a chatbot like ChatGPT gives us are statements based on the probability of which terms are empirically associated with which others. The low-threshold operation and multidimensionality of the possible solutions and answers enables a high degree of self-control of a company.

If it was previously necessary – going beyond the use of Google – to have one’s own data programmed into an application with a specific purpose; here everything can now be designed oneself. In particular, if one’s own data is also read in. This is still a bit complicated at the moment, but will be possible via API interfaces. ChatGPT can show much more than anecdotal knowledge, such as how a poem would sound in the style of Heinrich Heine. Such a bot can specifically “do” complex tasks. For example, a challenging learning unit can be created for a self-learning program in the company. True/false answers can be created; knowledge units created can be used as a basis for discussion for deeper understanding and reflection.

Actively using a chatbot is also becoming mandatory because, like Bard, they will increasingly specify their sources, should soon function in real time, will be integrated with clouds and Google Sheets and Adobe – and so on. Still, bots provide statistical sense-making, but don’t cover the exceptional cases and can’t illuminate complex relationships. As the founder of the Swiss bot “Open Assistent” (which, by the way, also draws on data from the “big guys,” in addition to data created in-house), Jannic Kilcher says of this kind of support:

“[It] lacks many things: a long-term memory. Or the ability to make logical inferences. It might be able to learn those from more data. However, there’s not that much high-quality data on the web anymore. What’s still missing is direct interaction with the Internet – and with the real world . ..”. He notes, on the other hand, that he still sees a role for humans: “Even if they were given and machines thus became human-like intelligent, I would not expect mass unemployment, an Armageddon. Rather, I would expect a change in society similar to that brought about by the smartphone. We will become more productive, do things differently, some jobs will certainly become superfluous . . .” (NZZ, 5/3/2023 – translated from German).

Humanoid work remains in demand

Humans will form the machine-human interface, in that they will be able to ask the right questions and consider special cases in context, and they will be able to reflect on the environment. This means that additional reflection and consulting skills are needed to make the best use of AI. In particular, it will also be a matter of translating and compiling what is collected and done with the bots.

Open questions

Although bots are already running and new applications are emerging by leaps and bounds, accompanying circumstances need to be considered and ultimately brought to consensus social regulation – or at least social agreement. This concerns, for example

  • Regulatory issues of a legal and economic nature, such as ownership of ideas or responsibility for damages;
  • Transparency about the origin of data;
  • Responsibility for content generated and, where appropriate, consequences thereof;
  • Attributability: the disclosure of underlying algorithms and, where appropriate, certifications for “truth” or a solution generated to a specific one.
  • Privacy and energy issues

Fields of application at a glance

The range of applications for chatbots is growing with the inherent machine learning, but also with data uploaded by the user. Their added value is therefore constantly increasing.

  1. Marketing & PR: Chatbots can create PR and social media templates, suggest strategies based on personas, and more. Trial and error is the order of the day.
  2. Sales & Customer Service: You can upload your own files via an API interface and have specific questions answered accordingly. However, check what their existing sales software can already do and how they interface with ChatGPT. As always: when using customer data, check the security of the data. Since this is hosted in the US, you need to find coherent ways here.
  3. Production, Quality & Processes: You can, especially also when you start “training” the bots with your own data and questions, understand ChatGPT and Bard as a complement to quality management and for process optimization or work-flow creation. It’s the same as for all other areas: Experiment, challenge the bot while scouting good input questions (“prompts”) and benefit from your bot.
  4. Personnel administration, HR & people management: You can make up a lot of ground, especially in the area of onboarding and offboarding and with regard to the processes stored for this, because the bot learns with you. Consider – again under the premise of what problems you actually want to solve – where you have advantages when working with bots. In particular, statements from surveys, failure data, and other information could also lead to AI-powered statements that add value to you.
  5. Workforce Development & Organizational Learning: learning from practice and via sharing are hip practices to move forward together. Use formats like peer groups and stakeholder knowledge to learn what workforce development might look like as well. Create exercises to use in staff development.
  6. Knowledge Management & Decision Making:Immense value is added by applications where knowledge is ordered and brought forth. Knowledge insists on information. Depending on the questioning technique, here you can playfully use the bot as a tool for knowledge transfer. You can create completely self-learning units and invite employees to exemplary knowledge transfer; the file for this then becomes your own knowledge management.

It’s about People and People decide

Forget that you are entering new organizational territory with the use of ChatGPT as well as with the possible editing of your own data sets in the app, which must be complemented by an appropriate culture of appreciation, shared learning, and a thinking in terms of roles instead of tasks and functions. Old-style ways of working will stifle creativity rather than help it flourish.

For the application areas outlined, explore what the bot can deliver to you. Also be aware that what it says will differ each time. So it’s not about absolute truth, but approximations as probabilities that can help them in their business. Carefully distinguish what the bot can do for you and where custom implemented AI applications can serve you.

Besser Zusammenarbeiten - besseres Onboarding mit Wissenstransfers - unlösbar scheinende Probleme mit Moderation zu einer Lösung führen - Speaker - Inspirationsquelle für Geschäftsleitungen

Comments are closed.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More