The sales manager had seen this movie before.
A new platform landed with big promises: "AI-powered coaching," "realistic simulations," "instant feedback." The demo was slick. Reps nodded along. A small pilot went well enough.
Six months later, the team was back in the same room, talking about the same performance gaps, and a renewal decision no one felt strongly about.
What went wrong? Not the idea of AI roleplay. The problem started much earlier, inside the RFP due to the questions never asked.
The promise behind AI roleplay isn't features. It's moments.
Feature checklists are easy to fill out: scoring, dashboards, multilanguage, integrations. But the real promise of AI roleplay and simulation is something else: better conversations in decisive moments.
A rep staying composed through a complex negotiation.
A supervisor deescalating a furious customer without breaking policy.
A manager delivering hard feedback without losing trust.
An interviewer recognizing potential, and staying anchored to fair, consistent criteria.
If your RFP doesn't start from those moments, you risk buying something that looks powerful but never really changes what happens in the field, in the contact center, or in performance reviews.
Four uncomfortable questions most teams avoid
Stronger RFPs often come from questions that feel uncomfortable at first, especially when multiple functions are involved.
Are we accidentally buying four-point solutions?
Roleplay, interview simulations, coaching workflows, and formal assessments often live on separate tracks. If you don't decide whether you want a single engine or a set of tools, vendors will happily sell into each silo.
Who really holds the keys, the vendor or us?
Business changes outpace content. If the only way to update scenarios is to file a ticket, your "AI practice" becomes a static catalog. The deeper question is how fast your own admins and SMEs can respond when reality shifts.
What will we say if a decision is challenged?
Once AI touches hiring or promotion, the conversation inevitably shifts to "How did we get this score?" and "Can we defend this decision?" If your RFP doesn't ask vendors how they support defensible decisions, you inherit that homework later.
Where does the data live, and who sets the rules?
These tools touch far more than generic learning data. They capture customer voices, employee language, candidate interviews, and policies. That raises questions about retention, residency, access, and model training that a simple "Is it secure?" cannot answer.
A better way to organize the conversation
Underneath all of this is a simple idea: your RFP should read less like a feature catalog and more like a story about how you want practice, coaching, and decisions to work in your organization.
That story can then be translated into structured sections: outcomes and use cases, authoring and admin control, coaching experience, bias and defensibility, security and responsible AI, integrations and analytics, so each function can see itself in the evaluation.
Turning reflection into an RFP you can defend
CGS Immersive has helped teams across Sales, CX, Talent, L&D, Compliance, and IT turn these uncomfortable questions into concrete benchmarks, and the Simulation, AI Roleplay, Interview & Coaching Checklist is one way of sharing those benchmarks with you.
Inside, you'll find:
A question set organized around the four core use cases.
Dedicated sections for admin control, bias and fairness, and responsible AI.
A weighted scoring model that lets different stakeholders score independently and still land on a decision you can stand behind.
If your next move in this space needs to serve more than one team and withstand more than one kind of scrutiny, starting with that kind of structure can make the difference between another "nice pilot" and a platform that truly changes conversations.
Download the Simulation, AI Roleplay, Interview & Coaching Platform RFI/RFP Checklist
