Practical AI has finally arrived, and tools such as Claude, ChatGPT, Perplexity and others can now assist us with technology transfer and commercialisation. There is a lot to keep track of and to learn but here are some key things to know:
- You can significantly enhance your productivity. AI tools can help your daily work in many ways, but thoughtful prompts, clear goals and well-thought-out workflows yield much better results than vague prompts and loose instructions.
- Different AI tools have different strengths. Some, such as ChatGPT are better at creative tasks, whereas Claude is better at data synthesis and others are better at web searching. They are all rapidly evolving – capabilities that seemed impossible a year ago are now standard features, with major advancements arriving every few months. The best way to keep abreast of changes is to use these tools regularly.
- AI is not a substitute for expertise. The tools amplify but don’t replace the knowledge and experience of practitioners. The best results are obtained when they are guided by professionals who understand both the technology and the commercialisation pathways.
- Technology transfer professionals must understand how researchers use AI in their work. In particular, AI-assisted research outputs can affect intellectual property protection since some jurisdictions only allow human inventors.
There are hundreds of specialised AI models that have been optimised for specific functions like prior art searches, market analysis, or targeted customer discovery. But, there are too many for all to be successful, and many don’t provide much (if any) advantage over using the comprehensive foundational models. There is also an incredible rate of change as each vendor strives to gain an edge over their competitors.
I personally use Claude, ChatGPT and Perplexity for my work. Perplexity for quick research with web references, ChatGPT and Claude for general queries, general research and coding. I haven’t used the latest Gemini 2.5 from Google yet, but it is supposed to be similar in performance to ChatGPT and Claude. These tools are all being updated constantly. New competitors such as Grok and Deepseek continue to enter the market and are reaching a very similar level of performance.
There are also many specialised applications built on top of these tools offering support for prior art searching, market analysis, customer discovery, etc. But there are too many of these tools for them all to succeed. My recommendation is to choose at least one of the major tools like ChatGPT, Claude, Perplexity or similar and maybe one or two specialised tools. I would avoid long-term subscriptions and be prepared to switch if your tools start lagging behind.
Most universities already have a policy around the use of AI tools within the institution. Ensure that AI usage is documented, and you have a process to identify potential IP issues before disclosures proceed down the commercialisation path. It would be prudent to include a field regarding AI use in invention disclosure forms, sooner rather than later.
Finally, remember that technology transfer ultimately succeeds through human relationships. Reserve AI for background work, but keep communications with researchers, industry partners, and investors authentically human. The personal touch remains irreplaceable in building the trust needed for successful commercialisation.
Here are some useful websites to look at:
- Best AI Tools – directory of AI tools (one of many)
- Stanford HAI AI Index Report – yearly report on the state of AI
- AUTM AI Special Interest Group – announcement (membership required for details)
- MIT Guidance for use of Generative AI – policy on use of AI
