AI can make research, report drafting and analysis faster. Elly Dowding and Lee Coates OBE, directors at In Accord and the Accord Initiative say that for paraplanners, the important question is not simply what the technology can do, but how its use affects evidence, suitability and the chain of human judgement behind a recommendation.
Artificial intelligence is becoming part of the advice process rather than a separate technology project.
Meeting summaries, regulatory searches, report drafting, data analysis and client communications are obvious examples. For paraplanners, many of these uses are attractive because they sit precisely in the parts of the process where time is scarce and information volumes are high.
The risk is not that AI is used. The risk is that its contribution becomes invisible.
If a recommendation is supported by AI-generated analysis, research or drafting, the file still needs to demonstrate the same thing it always has: why the recommendation is suitable and where professional judgement was applied.
AI adds another link to the evidence chain
Advice already depends on a chain of trust. Client information feeds research. Research supports a recommendation. The recommendation is explained and evidenced. AI simply introduces another link into that chain.
That does not remove existing regulatory responsibilities. As Iga Sloan of Digital Regs highlighted in our recent Accord Talks discussion, the Mills Review leaves the existing framework in place. Consumer Duty, SMCR, operational resilience and data protection requirements continue to apply.
For firms, that means AI governance needs to be practical. Know where AI is being used. Start with the business problem. Assess vendors and data flows. Allocate ownership. Check outputs. Document the process.
For paraplanners, the final point is particularly important. AI output is not a copy-and-paste exercise. If analysis or wording finds its way into the advice file, somebody needs to understand it well enough to test it, challenge it and stand behind it.
The file should reveal the judgement
Good paraplanning has never been about assembling technically correct information. Its value lies in interpreting that information in the context of the client.
That becomes more important as AI improves. A system may identify patterns, compare solutions or produce a plausible rationale. But the suitability case still depends on whether the reasoning fits this client, with these objectives, constraints, behaviours and preferences.
This is where the distinction between data and understanding becomes useful. A fact-find may record that a client is a particular risk category. The file may also show a time horizon, income requirement and investment objective. None of those data points, on their own, establish that the proposed solution will be experienced by the client as suitable.
Andrew Storey of EV gave a useful example in our podcast discussion. Risk profiling is generally well established, but suitability has more than one dimension. Investment style and preferences need to be explored where relevant, rather than assuming a standard solution will fit everyone.
Capacity for loss makes the point even more clearly. Andrew described it as an ability, not a preference. Asking a client how much they can afford to lose is not necessarily meaningful; in relevant cases the capacity needs to be calculated, including the effect of market sequencing and the client’s wider plan, and then explained to them.
That has a direct implication for paraplanning. The research should not merely record an answer. It should show why the answer is credible.
Explainability becomes part of due diligence
As AI becomes more involved in research and drafting, explainability should become part of the due diligence standard.
Can the paraplanner explain why a particular source or tool was used? Can the underlying assumptions be identified? Can the output be checked against reliable evidence? Can somebody spot when a result looks plausible but is wrong? And, crucially, can the recommendation still be explained clearly to the client?
These questions are not arguments for creating another layer of paperwork. They are arguments for preserving a visible chain of reasoning.
A well-evidenced file should make it possible to see the progression from what the client needs and values, through the research and analysis, to the eventual recommendation. If AI contributed along the way, that should not weaken the chain. Properly governed, it may strengthen it by improving consistency, analysis and the time available for review.
More automation, not less responsibility
There is a temptation to assume that greater automation means less human involvement. In regulated advice, the opposite may be true.
The more a system can do, the more important it becomes to define what it should do, where it should stop and when a human needs to intervene. That is governance, but it is also professional judgement.
For paraplanners, this may be one of the most important consequences of AI. Routine production work can become faster, while the value of challenge, interpretation and context increases.
The future file may therefore contain more technology behind it, but it should not contain less reasoning. It should make human judgement easier to see.
Knowledge is becoming easier to generate. The paraplanner’s contribution is to help turn that knowledge into an evidenced recommendation that makes sense for the person behind the data.
These themes are explored in more depth in the latest Accord Talks podcast, The Human Advantage.
Elly Dowding and Lee Coates OBE are the directors at In Accord and the Accord Initiative, which provides free-to-access education, resources and compliance support to the financial advice sector.
Main image: background, white, paper, jj-ying-WmnsGyaFnCQ-unsplash
