A version of this piece was first published in Voluntary Benefits Voice, the monthly publication of Voluntary Advantage.

The problem with innovation is that it's … well … change. Geoffrey Moore's early adopters, who once had years to build the bridge to the “early majority,” are now getting crammed into the same marketspace as AI compresses that timeline. APIs, AI-driven Q&A, analytics, and decision support have all become table stakes — and table stakes don't generate new wins on their own.

First, the leverage point is moving from who controls the data to how flexible the configuration is — products, rates, age bands, and eligibility rules increasingly need to exist in a shared, API- and agent-ready form rather than get rebuilt case by case.

Second, decision support is migrating well past the open enrollment window, becoming a year-round, conversational presence that shows up at the point of a trigger or claim rather than going dark once enrollment closes.

Third, software is starting to act, not just inform — once a plan's design and rules are agent-ready, the question of whether an employee's own assistant could compare, or even elect, coverage on their behalf stops being science fiction.

For employers, that shift is starting to show up less as a new set of vendor features to evaluate and more as a change in which questions are actually worth asking.

How this is likely to show up
The things that have differentiated a benefits vendor pitch — a polished portal, a custom integration, a smooth open enrollment experience — are becoming commodities across the board. That's a reasonable thing for an employer to welcome, since it means the baseline is rising everywhere at once. It also means the old evaluation criteria are losing their usefulness for telling vendors apart.

The more useful questions increasingly involve configuration and continuity: how flexible is a given vendor's underlying setup, really — versus how much of what looks purports to be “customization” is actually untracked exceptions living in someone's head? Does decision support for employees genuinely persist past the three-week open enrollment window, showing up at an actual claim or life event, or does it go dark the moment enrollment closes?

Configurability "in the loop"
Benefit program vendor changes are part of what's driving this. As they standardize their own configurations and connect more APIs to lower distribution and integration friction, programs become easier for consultants to compare cleanly against each other. Here again, it may end up being less the features a health management vendor offers, and more how data from other vendors may enrich and deepen the employee experience - which may in turn increase the number of employees who find it easier and thus more desirable to use the solution, than to go it alone.

There's a further-out question worth having on the radar, even if it's not yet a live decision: as AI becomes core to how employees get guided toward benefits decisions, it seems plausible that similar tools eventually get built to help employers evaluate programs, vendors, and even their own consultants. Being an early, informed participant in how that gets designed — with real humans still in the loop — looks like a stronger position than encountering it after the fact.

Part 2 of this series gets specific about a related cost question that's likely to become relevant at renewal — what “AI-powered” capability actually costs to run, and who's accountable when that cost, or a decision made by it, goes sideways.

~ Mark Head

© 2026. All Rights Reserved.

Set a 15-minute call

Set a 30-minute call