If you’re a writer or editor, or anyone else who works with words, and you’ve attended a conference lately, you can attest that the hottest topic at that conference was AI. For that matter, if you’re a designer, project manager, doctor, engineer, therapist, farmer — the list goes on — you can count on AI as a central theme of any industry event.

AI was, of course, front and center at this year’s Bid and Proposal Conference (BPC), hosted by the Association of Proposal Management Professionals. Dragonfly is always excited to attend APMP’s annual flagship function, where we get to reunite with old friends, make new connections, and gain fresh insights into the proposal world. This year’s conference, held May 13-16 in Denver, gave us some hope.

From automation to accountability

We knew going into it that AI would dominate discussions and breakout sessions; AI was the talk of the town at last year’s BPC in Nashville too. In Nashville, though, it felt like the proposal industry was in a hard sprint toward an AI-only world.

“Your job won’t be replaced by AI, but you will be replaced by someone who knows how to use AI” was a popular phrase that floated across sessions in 2025. After all, the entire proposal process — from the bid/no-bid analysis, through the first draft and SME review, to revisions and editing and design — could possibly be handled by AI. The proposal team of the future would look very different and would certainly be much smaller.

This year, in Denver, there was a noticeable shift. Proposal professionals, many of whom have spent the last 12 months testing what AI can and can’t do, are pausing. They’re backing off their AI-only predictions.

In a session from Deloitte’s Gregory Roberts and Arlo Solutions’ Sarah Bednarz, the focus was on real people preparing proposal content for effective AI use. A key theme was the continued need for a “human in the loop” as data becomes the main differentiator. Roberts also noted how this is showing up operationally, with human check-ins moving from quarterly to monthly and now even biweekly.

These themes showed up in session after session, and all this talk of real people was a welcome change in tone. When you hear enough of your colleagues say “humans still need to be in the loop,” it becomes clear that it’s not just a platitude meant to make humans feel better. It’s a genuine discernment of the landscape.

The cost of ‘saving time’

There’s still talk that AI will save everyone time. In one session we attended, a panel of proposal professionals excitedly declared that they can crank out a first draft in minutes instead of days, which sounds great on the surface.

When asked how this is possible, though, they described the necessary prep work: Find an AI platform that fits your situation; this will not be an off-the-shelf product, but rather something customizable. Work with the platform’s vendor to set it up according to your needs. Feed your content library (if you have one) and other data into the platform. Test it thoroughly. Maintain it over time based on user feedback.

For us, this raises an obvious question: Is all that money and all that work worth the time you save creating that first draft? Granted, at Dragonfly, we’re word people. We love to write, so putting that time into hammering out a first draft by hand feels natural and rewarding. We derive no joy from building the perfect bot to do that work for us.

So, what’s next?

Every organization is in a different place in its AI journey. At Dragonfly, we continue to explore how to ethically implement AI into our daily workflow. One way we’re doing this is through our newly assembled AI Council.

“We choose to lead AI rather than be led by it,” reads the AI Council’s Purpose. “By becoming expert operators and stewards of AI, we free ourselves to do what only humans can: Provide editorial judgment, be creative, build trust, and show up for our clients as strategic partners.”

When evaluating growth opportunities, whether through AI or other tools, we always come back to the following questions:

• Is AI actually the right solution?
• Would automation or another process change work better instead?
• Does this need a human touch? If so, what level of involvement?
• Is using AI in this case worth the environmental/resource cost?
• Are we choosing the least energy-intensive tool/model that fits the task?
• Does this improve our team’s ability to focus on higher-value work?
• Will this genuinely save time or just create more checking/rework?

We hear people at other organizations — even the organizations that were putting all their eggs in the AI basket a year ago — slowing down and asking themselves some of those questions too. This should be a relief to humans everywhere, because it means the bots haven’t come for us just yet.

Written by Grace Teater and Dave Nelsen