It’s a timely question for the practices DocVA works with every day. Many have already brought in virtual medical staff to handle scheduling, insurance verification, and documentation, and are now wondering whether AI belongs in that mix too, and if so, where it starts. Gaurav’s answer is refreshingly practical, and it has as much to do with sequencing and measurement as it does with the technology itself.

Q: Healthcare practices are hearing constantly that they need to “adopt AI.” Where should that conversation actually start?

Gaurav: It should start with readiness, not tools. Most practices we talk to have already tried to bolt on an AI feature somewhere, a chatbot, a transcription tool, and it either stalls or creates more work than it saves. The real starting point is understanding what data you have, what workflows are actually broken, and what a successful outcome looks like in numbers. That’s the assessment phase and skipping it is the single biggest reason AI projects fail to scale.

It should start with readiness, not tools.

Gaurav Murghai

Q: Many practices have already brought in virtual staff, people like the medical assistants and front-desk support DocVA places, to handle administrative load. How does AI fit alongside that kind of human staffing model rather than replace it?

Gaurav: That’s exactly the right frame. The practices doing this well aren’t asking AI to replace a person, they’re asking where AI can remove friction from the work that person is already doing. A virtual medical assistant handling insurance verification or patient scheduling is making judgment calls constantly, reading context, knowing when something needs escalation. AI is very good at the repetitive layer underneath that: pulling records together, flagging missing documentation, drafting a first pass on a prior authorization. The human stays in the loop for judgment and patient relationship, the AI takes the grunt work off their plate. When we design implementations, we’re almost always designing for that handoff, not for full automation.

The human stays in the loop for judgment and patient relationship, the AI takes the grunt work off their plate.

Gaurav Murghai

Q: You mentioned measurable outcomes a couple of times. What does that look like in practice?

Gaurav: It has to be something a practice owner or operations lead can point to and defend. Hours of staff time recovered per week, reduction in documentation lag, fewer dropped follow-ups, that kind of thing. This is really the core of how we operate at Softobiz: we take a client from strategy and readiness assessment all the way through implementation to outcomes that are actually measured, not just assumed. If you can’t tie an AI initiative to a number, it’s very hard to keep it funded past the first budget cycle.

If you can’t tie an AI initiative to a number, it’s very hard to keep it funded past the first budget cycle.

Gaurav Murghai

Q: Any advice for a practice owner who feels behind and doesn’t know where to begin?

Gaurav: Don’t start by shopping for tools. Start by mapping where your team’s time actually goes and where the bottlenecks are, whether that’s administrative staff, virtual support, or clinical documentation. Once you can see that clearly, the right starting point for AI usually becomes obvious, and it’s rarely where people initially assumed it would be.

Gaurav’s full background and Softobiz’s approach to enterprise AI transformation are available at softobiz.com .