There is understandably a great deal of talk about AI in the BA community. One topic of regular conversation is how AI tools can help BAs be more efficient and effective. A second is how business analysis can help ensure that organisations deploy AI in ways that are appropriate, ethical and effective. Both of these are important topics, and worthy of articles of their own, but there is a third which appears to be getting far less “air time”.
AI Use by Outside Actors
Most conversations within organisations seem to focus on internal AI use. This is a type of AI usage that is, theoretically at least, controllable to an extent. What seems to be discussed less often is how policies and procedures might need to change in response to the way that external actors are using AI.
At first glance, this might seem like a distraction—and indeed perhaps this is why it is not attracting attention. Yet further analysis would likely show that there are some fairly significant areas where AI usage might have an impact. For example:
- Customer Agents: Imagine a customer uses an unattended AI ‘agent’ to connect to an insurance company’s ‘live chat’ to make a small, routine change. The agent clears security, and provides an instruction which is actioned. If the customer later challenges that instruction, how should the organisation respond? A lot will likely depend on whether it has a policy in place around whether or not it permits a customer to correspond via agentic AI. If it hasn’t considered this in advance, there will likely be a costly decision to be made the first time an issue crops up.
- Complaints Made Easy: Let’s face it, we’ve probably all had situations where we’ve wanted to complain about poor service, but life gets in the way and we move on. What about when customers decide to use AI to craft a formal complaint, citing case law and precedents? Sure, the AI may well cite a mixture of real and fake law (“hallucinations”), but has the complaints team been trained on how to respond? Is there a risk complaints will increase because of their ease, and the processing time will increase, due to the complexity of interpreting them. If the process itself doesn’t change, this seems likely…
- Suppliers Responding to Requests for Proposal (RFPs): RFPs (or Invitations to Tender, ITTs) are common in procurement, and typically utilise a series of questions backed by a scoring mechanism to help determine the (theoretical) “best fit”. Suppliers will likely use AI to start responding to these. Yet, this might mean there are a greater number of responses… and what about “hallucinations”? Sure, it’s easy to say that it’s the “supplier’s responsibility”, but I suspect many reading this have seen situations long before the widespread use of AI where a supplier “stretched” the truth about a capability. Add in AI hyperbole, and that risk feels manifestly larger. Procurement processes may need to change to place even greater weight to evidencing key capabilities.
Using Business Analysis: AI Stakeholder 360
This is just a small list of examples, I’m sure you can think of others relevant for your situation… and this is before we even start to think about bad actors and cybersecurity implications!
Stakeholder identification and analysis has always been a key part of business analysis, and when it comes to AI, this is no different. As well as undertaking it in a traditional way, it is worth asking the questions: “how might this stakeholder formally use AI in their role” and “how might they informally use AI”. This second question represents the reality that humans are resourceful, and even if their organisation has no formal AI policy (which it absolutely should!) they may still decide to use AI tools under the radar. This has its own risks associated, but to deny it happens would be to deny reality.
In essence, it is worth conducting a 360 degree stakeholder analysis, identifying who is (or might use) AI and how. Then, the key impacts and associated processes can be prioritised. There will be times when no action is needed, some where action is needed in the medium term, and likely some where pain is being seen already.
With so much else taking up organisational time, conducting an exercise like this might seem optional. But organisations will have to deal with customers and other stakeholders using AI. If this isn’t considered in advance, there’s a risk that any responses will be “urgent”, knee-jerk type reactions. And that is usually best avoided!
What are your views? Please add a comment below, and let’s keep the conversation flowing!
If you’ve enjoyed this article don’t forget to follow me on LinkedIn and check out BA Digest, the quarterly BA magazine.
About the author:
Adrian Reed is Principal Consultant at Blackmetric Business Solutions, an organisation that offers Business Analysis consulting and training solutions. Adrian is a keen advocate of the analysis profession, and is constantly looking for ways of promoting the value that good analysis can bring.
To find out more about the training and consulting services offered at Blackmetric, please visit www.blackmetric.com
© Blackmetric Business Solutions, published exclusively on www.adrianreed.co.uk. May not be republished on any other site without permission.