Some sectors were initially more immune than others; healthcare, law and finance being three prominent examples of industries that took a while to integrate AI. However, we are now seeing version 2.0 in the forms of HealthTech, LawTech, and FinTech. So if we focus purely on the tangible results of AI, then how can you accurately value the integration of AI into your startup?

How can we value AI?

Let’s zoom in and look at the direct output from AI. Assuming your first foray would be to buy rather than build, then let’s use ChatGPT – and perhaps the wider world of LLMs – as an example. Firstly, I want to really drill down that the quality of your output is only as good as your input. There’s no “right way” to prompt, but there are hallmarks of what a “good prompt” might look like in the results that an LLM produces. These are fidelity, and coherence.

Think of fidelity as the ability to be indistinguishable from real-world examples. It is realism. Perhaps it’s your AI chatbot deployed to onboard new customers that appears to be another (human) team member.

If fidelity is the train carriage, then coherence is the train. Coherence, as the name might suggest, is when the output of an LLM has a logical flow to it. It is the art of storytelling; think of this as a longer piece of data or text that “makes sense”. Perhaps this could be an AI-generated newsletter sent out to shareholders.

At this zoomed-in level, it can be easy to qualify “what good looks like”. This can lead to quick wins; for example, a coherent AI-generated marketing campaign that resonates with your audience or a high-fidelity customer support chatbot that resolves issues efficiently can be a resource-light game changer for your startup.

Traditional KPIs are still the strongest barometer for success

Quick wins are useful, but let’s zoom out. They can go by any other name, but traditional Key Performance Indicators (KPIs) are often a decent barometer of success for a startup. Chances are you already have a few; measures like sales figures, customer satisfaction through an NPS score… They are a gold standard in measuring how well your business is doing.

Now here’s the best bit: you don’t need to be reinventing KPIs for the purpose of valuing AI in your business. AI should be driving these existing KPIs forwards (if they are integrated into those workflows). All too often I see business owners – and this applies all the way up to enterprise level – trying to create new KPIs to fit the AI-driven narrative. Don’t do this. If you’re using AI in your sales funnel, keep measuring those sales figures. Have an AI chatbot? Continue to measure that NPS score. AI creates efficiency – it can streamline processes, reduce operational costs, and improve decision-making speed. It can automate repetitive tasks, and focus human efforts on more strategic initiatives. However, the end product should be that it becomes easier to move towards your existing targets.

A/B testing can also be a valuable method to measure AI’s impact. Compare the performance of AI-driven processes against traditional methods to establish a clear baseline for success and make data-driven decisions.

AI needs to be holistic

The best way to integrate AI into your startup is to have a holistic view. Consider not only the technology at work, but also the human and ethical implications. Ensure that the aim with AI is to compliment rather than to replace your workforce, especially within its current stage. Employees should feel empowered and supported by AI tools, which will create higher job satisfaction and productivity. Engaging and educating your team in the AI integration process, providing necessary training, and addressing any concerns can mitigate resistance and create a more positive attitude towards AI.

It’s also crucial to consider the ethical implications of integrating AI. It must be designed and implemented with fairness, transparency, and accountability in mind. There are (well-documented) potential biases in AI algorithms and so it’s critical to strive to maintain ethical standards in any AI practices. This will not only enhance the effectiveness of AI but also build trust and credibility with your stakeholders. Due diligence is key.

We’re on the side of AI, and so it will come as no surprise that I firmly believe that AI has a place in every startup. Adopting an AI native approach from the outset is, in my opinion, the best foundation for the majority of startups, and ensures that AI is not merely an add-on but a core part of your business strategy. Just be sure to be vigilant in what it is you are measuring when it comes to assessing the return on investment, and keep all of your stakeholders informed about where you intend to include LLMs in your workflows. And don’t create new goalposts, just see if you can move the ones you already have.