Vendor Contracts: The 3-Year Lock-In is the Industry Default
With half of all benefits contracts spanning three years, the stakes and long-term value of the outcome become much higher for buyers and vendors alike.
In This Post:
Most information on a vendor’s profile describes the present.
A capability is either offered, or it is not, and both answers can be tested in a demo, or confirmed on a reference call.
However, when a vendor marks a capability as “on the roadmap,” it is announcing a plan to sell something it does not have yet.
That creates a problem for any vendor due diligence checklist: the entry records intent, and intent is the one thing an evaluation cannot test.
A Roadmap Is a Forecast
Shortlister allows vendors to mark capabilities as offered, not offered, or “on the roadmap.”
Although the benefits vendor build queues can reveal useful market themes in aggregate, a single entry cannot be treated as a delivery record because the product and profile may change.
In fact, software vendors often draw that boundary themselves.
Microsoft’s Dynamics 365 release plan warns that delivery timelines may change and projected features may never be released. Oracle states that its general product direction is not a delivery commitment, whereas Workday advises customers to base purchases on currently available services, as future functionality may change or fail to arrive as planned.
Those statements define how product roadmaps are meant to be used. They communicate direction while preserving the company’s ability to revise its plans.
Vendor vetting should apply the same definition. Until the buyer can verify a capability, the roadmap remains a forecast.
A Missed Date Is the Industry Norm
None of this suggests that the vendor is being dishonest.
A roadmap date can be sincere on the day it is written and still be wrong, because companies are consistently poor at predicting how long their own projects will take.
Teams estimate future work under uncertainty, often before all dependencies or technical problems are known.
For example, Oxford’s Bent Flyvbjerg has built a database of more than 16,000 projects across 136 countries. By his count, only 8.5% finished on time and on budget, and 0.5% delivered the promised benefits.
His research connects much of that gap to optimism bias, which encourages teams to plan around the expected outcome rather than the performance of comparable projects.
AI Has Its Own Roadmap Problem
Software has a long history with the same problem.
In fact, the industry coined the word vaporware in the 1980s for products announced well before they were ready, with Microsoft’s delayed release of Windows helping popularize the term.
And the pattern remains today, particularly when it comes to AI.
Gartner estimates that only about 130 of the thousands of vendors selling “agentic AI” have built anything that earns the label, and predicts that more than 40% of agentic AI projects will be canceled by the end of 2027.
That matters in HR technology, where AI capabilities increasingly appear across HRIS, benefits administration, and other workplace technology.
Shortlister’s own data shows how quickly AI has moved into vendor messaging.
Mentions of AI in vendor proposals have increased 340%, putting more pressure on buyers to separate capabilities that exist today from those still being built.
Regulators Police the Present Tense
Roadmap language gets more complicated when optimism crosses into a claim that the company cannot support.
Regulators have been particularly active around claims about capabilities that supposedly exist today.
In March 2024, the SEC charged two investment advisers with false or misleading statements about their use of AI. In January 2025, it settled charges against Presto Automation over claims about how its AI ordering product was supposed to work.
That being said, future product announcements are not automatically exempt from scrutiny either.
U.S. antitrust law has long recognized that false product announcements can become a problem when they are used to influence customers and disadvantage competitors.
But buyers operate under a much lower threshold than regulators.
“Planned” Hides More Than It Reveals
For buyers, the central problem is the information gap.
“On the roadmap” compresses several separate buying questions into one label.
It records the vendor’s intention but does not show how far the work has progressed, when customers will receive it, who will have access, or whether the quoted price includes the capability.
Two vendors can therefore mark the same requirement as “on the roadmap” while describing very different situations. One may already have the feature in pilot with a general release scheduled. Another may have little more than an internal plan.
On a vendor profile, those two answers can look identical.
That information gap is what buyers need to close through a vendor due diligence checklist before giving a planned capability credit in an evaluation.
Why a Target Date Is Not Enough
A target date closes only one part of the information gap.
“Q1 release” still does not tell the buyer whether that means a limited pilot or general availability, or what happens if the date slips.
A vendor risk assessment questionnaire should therefore ask more than whether a capability is planned:
- What stage has the capability reached: planned, in development, pilot, or generally available?
- What is the current target date, and what exactly is expected to happen on that date?
- Which products and customers will have access?
- Will the capability be included in the proposed price or require an additional purchase?
- What evidence can the vendor provide today?
Recording these answers in a vendor due diligence checklist gives the buyer something much more useful than “planned.”
They separate intention, timing, access, price, and accountability rather than assuming a single roadmap status can answer all five.
A Roadmap Can Guide the Conversation
Roadmaps are still helpful because they can help a buyer understand where a product is heading, particularly when two vendors meet the current requirements equally well.
It becomes risky only when future functionality receives the same credit as something that already exists.
Written by Ivana Radevska
Senior Content Writer at Shortlister
Sources
- Shortlister Proprietary RFP Data, Q2 2026
- Dynamics 365 Business Central Release Plan, Microsoft
- Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027
- Oracle Health and Life Sciences Product Roadmap
- SEC Charges Two Investment Advisers Over AI Statements
- SEC Charges Presto Automation Over AI Product Statements
HRIS Systems
Browse our curated list of vendors to find the best solution for your needs.
Stay Informed
Subscribe to our newsletter for the latest trends, expert tips, and workplace insights!
Related Posts
The Benefits Vendor Build Queue Is Compliance, Not Features
Many benefits vendors already offer the expected product features. Shortlister data shows that security reviews and health care accreditation are moving to the top of their priority lists.
“On Our Roadmap” Is the Least Reliable Line in a Vendor Profile
A roadmap is written in the future tense but scored in the present. That mismatch allows a planned capability to influence a buying decision before anyone knows whether it will exist under the customer’s contract.
The Shrinking Shortlist
Shortlister data shows that the average RFP now includes fewer than six vendors. The vendor selection process has moved upstream, where research and AI increasingly decide who reaches the formal evaluation.