Are you nervous over what McKinsey and Gartner say about gen AI? Artificial intelligence is going extremely fast now. Yesterday’s newsletter of McKinsey and recent experts webinars by Gartner are clear about where our society goes and which companies will be the winners.
The good news: it is believed that Gen AI will really turn the markets ‘only’ by 2028 (in a window that spans between 2027 to 2029). So, there is still three years to go. ‘Non-AI servicing’ will quickly decline ‘only’ as of three years from now. The bad news: the inevitable learning curve started a while ago.
McKinsey’s vision
McKinsey observed that high-performing companies are using gen AI in multiple business areas (on average in three areas), not merely one or two. While that typically relates to marketing, sales, and product or service development, the outperformance is where these companies also deploy gen AI solutions in a wider field of their organisation: to include risk, legal and compliance ( ! ), in strategy and corporate finance, or in supply chain and inventory management.
Yep: high performers happen to have supporting corporate functions already in the learning curve of AI service delivery. Notably, this includes legal departments. It is probably a mindset-thing in their organisational DNA that defines their high-performance. Where departments beyond marketing, sales or product development embrace a ‘growth mindset’; an attitude of ‘can do’ across the organisation makes them high-performant.
Gartner’s view
But there is a deeply fascinating aspect, and every legal department should go through it. Or rather, will inevitably go through it. Gartner identified several stages of implementing AI. It only starts with a foundational model (like the generic ChatGPT).
With a ‘fixed mindset’ of scepticism about (the risks associated with) AI, an organisation will never arrive at a fully operational AI servicing. Corporate functions will see that AI only starts with experimenting, with getting ‘consistently good answers’. To elevate service delivery to a higher level (of ‘seamless UX’ – a super smooth user experience), Gartner tells us that the AI journey goes from:
- Consistently good AI answers (identify pitfalls and navigating them adequately), through
- AI answers that fit the customer context (learning how to improve AI deployment), to
- AI answers that the customer helped shape (increased customer engagement – for a legal department collaborating with their internal clients), to
- AI answers for the internal client that are intuitive to get to and use.
This learning happens in real life practice, as it is specific to the industry, the business’s particular context, and the stakeholders involved. In other words, implementing AI services effectively and efficiently requires a learning curve, and it is better to step in now.
Consistently getting good answers requires AI prompt engineering, as much as each of the three subsequent stages of developing an AI-driven delivery of (legal) services.
What matters now?
McKinsey states it very clearly: high performers are paying more attention to gen-AI-related risks. They have probably, more likely than others, experienced every negative consequence from gen AI already.
Yep, high performers are not necessarily the high-risk-takers. They are just faster in figuring out how AI accelerates their business in all relevant respects. High performers are more likely than others to consider risks (!), as well as other legal and reputation-defining aspects, to be crucial to their gen AI use. Even more fascinating (from a legal department’s point-of-view) is that high performing companies are more proactive in mitigating risks than others.
Gartner’s other predictions
More than ever, it is crucial to step into the learning curve, because AI requires more than ‘just’ AI service delivery. The learning curve will bring self-assessment insights on:
- Better quality-organised data (and elevate Weagree CLM metadata to a more mature level)
- Foundational models as a service (integrating ChatGPT, Microsoft Copilot with Weagree)
- You’ll know what you need (specific to your industry, your business’s particular context, and your internal stakeholders)
- Domain models as a service (use of AI models specific to the legal domain)
It is not very likely that companies will develop their own in-house AI models, as there is likely not enough sample data inhouse to achieve this in non-core business fields (let alone the ‘training’ effort that would be required).
Weagree’s strategy is therefore to offer powerful API-integration options. Weagree can easily be connected to ChatGPT. Weagree’s extraordinary user-friendly user interface provides you the seamless UX to optimise your contracting processes.
