For both professional and personal reasons, I’ve been thinking a lot lately about the craft of summarizing ideas through writing.
A significant portion of my consulting portfolio involves creating what I collectively think of as summary documents: reports, evaluations, toolkits, and funding proposals. In these cases, my job isn’t to be original or innovative; people hire me to elevate their ideas with clarity and conviction.
In a typical project, I read and review documents provided by my client, conduct additional research and data collection as needed, and then write a new text that meets the requirements of a specific audience, such as a funder or an international evaluation committee.
At home, I’ve recently observed both of my children working on school assignments (think: research papers and projects) that require them to read multiple texts, pick out key arguments, and restate ideas concisely in their own words.
Watching them work, it dawned on me that the skills that I take for granted as part of my daily work don’t come naturally: they must be taught, practiced, and reinforced.
Or, because it’s 2026, we could simply outsource this work to AI.
But should we?
The empathy gap
About six months ago, I wrote about my journey from AI resister to cautious adopter. Toward the end of that article, I concluded that AI was ill-suited to the kind of strategic and developmental editing that I typically do, mainly because such work requires one to consider a text from multiple vantage points and make constant judgment calls about how a particular word, sentence, or idea will resonate with different kinds of readers.
To put it succinctly, the best writing requires empathy for the reader—for their perspective, background, and attention level.
AI does not possess empathy.
But what about the relatively straightforward work of summarizing text or drafting reports from pre-existing materials?
Like many other tasks, AI-generated summaries and compilations offer a tempting level of complacency and security. The speed and apparent efficiency might lead one to think that human writers will become obsolete, particularly for seemingly routine and “uncreative” kinds of writing.
Yet the more I experiment, the less convinced I am that AI is up to the task.
The human advantage
While I initially planned to turn this newsletter into a catalogue of AI-summary failures, I feel as though that trope has lost its appeal or efficacy. Many people today are leaning on AI to do this work, even with full knowledge that the results risk being inferior or ridden with errors.
Of course, there will be situations where AI-generated summaries are helpful—especially as a starting point for human editing. (Personally, I’ve found AI meeting minutes to be a considerable win to my productivity.)
I find it far more thought-provoking to consider some aspects of crafting summary documents that render human judgement and perspective essential.
Summary documents are not always straightforward or routine. Whether it’s generating a list of bullet points or crafting a longer report, summaries require judgement: what to include, what to leave out. If you are managing complex information, communicating controversial subject matter, or attempting to engage diverse stakeholders, then you’ll want a flexible and empathic writer in your corner.
The most compelling summary documents are driven by curiosity and infused with context. Nearly every time a client provides me with what they believe to be a comprehensive trove of material, I end up requesting additional information. Why? Because as a writing consultant (functioning simultaneously as your first outside reader) my job is to ask—and answer—any lingering questions before your grant reader or evaluation committee has a chance to do so. Even if contextual information doesn’t explicitly appear in the final document, it nonetheless informs the work and shapes the narrative.
The practice of writing summary documents will generate new insights. A few weeks ago, I was startled to read an interview with a university professor who noted that she hired fewer research assistants and instead relied on AI to do most of her literature reviews. While I recognize the pressures of extremist budget cuts to research, I was nonetheless surprised and saddened to read of this casualty to the pursuit of knowledge and training of new scientists.
By compiling and engaging with previously written material, choosing what to include, considering different viewpoints, and ultimately writing (and rewriting) a new text, you have the opportunity to consider established ideas in a different light. And if you do this work collectively —whether you’re leading an organization, a research group, or a writing team—you have even more chances to innovate.
This, in my mind, is one of the most rewarding aspects of writing reports—and the greatest loss if and when this work is outsourced to machines.
Copyright 2023 Intellerate Consulting. All rights reserved.