The buyer stopped waiting for you.

The idea of the trusted advisor is eroding, and AI is accelerating it. This isn’t new. It started well before generative AI existed. During COVID, pharmaceutical reps lost in-office access to doctors, who used the disruption as a reason to stop taking meetings they’d already grown tired of. That’s a real part of why so much pharmaceutical advertising now targets patients directly on TV: reps lost the room, so the industry rerouted around them.

Professional services is exposed to the same pattern, because consultative, relationship-based selling has always depended on early access. Buyers increasingly don’t want that access from a rep when AI can do the same job faster and without a sales agenda attached.

The 67% figure is already a year old, and the trend line only points one direction. The 69% figure is the one worth sitting with: buyers aren’t skipping human judgment, they’re arriving with an AI-generated plan already in hand and asking you to validate it. Sometimes that plan is genuinely useful. Sometimes it’s confidently wrong. Either way, you’re no longer opening the conversation. You’re auditing someone else’s homework, and deciding in real time whether to follow it or rebuild it.

This is standard sales logic, not a new insight: the earlier you enter a relationship, the more trust you build, and the longer and healthier that account tends to be. What’s changed is where “earlier” is actually possible. Buyers are leaning on AI specifically because it lets them avoid the moment they’d otherwise have to talk to a rep. The natural entry point keeps sliding later in the journey, not because sellers got worse at their jobs, but because the door buyers used to walk through on their own no longer requires a person on the other side of it.

None of these are the only ways to move earlier, but they’re the three worth building discipline around. Referrals are the known quantity: almost certainly the strongest revenue source in the room already, and the question is how to generate more of it deliberately instead of hoping it happens. The AI-to-AI approach is closer to “if you can’t beat them, join them”. If buyers are going to lean on AI regardless, the opportunity is influencing what that AI has to work with. Positioning is the branding and storytelling question underneath both: this is moving fast, and there’s now real data on what works and what doesn’t.

The common model is what most technology consulting firms already live inside: revenue built on fixed projects with a defined start and end. Once the implementation is done, the relationship naturally goes quiet. The usual fix is newsletters and check-in emails, but to a client who’s already finished the work, that’s noise. It shows up as unsubscribe rates, not engagement.

The engineered model borrows a page from private equity. PE firms treat client-portfolio diversification as a non-negotiable early priority for the companies they acquire, precisely because smaller businesses tend to get comfortable inside a narrow client base and a narrow project type. That comfort is what caps growth. One way they break the pattern: take on smaller, easier-to-close projects across a broader mix of verticals. Yes, churn is higher on those. But each one is a live opportunity to ask for an introduction while you’re still top of mind, something that becomes far harder once the project has wrapped and the relationship has gone cold. It’s an uncomfortable ask to build into an active project. It’s a necessary one.

A second lever alongside this: converting a completed project into a retained, annualized offering, such as performance optimization, analytics, or some form of managed service, so you stay in front of the client instead of resetting to zero. That ongoing presence is what keeps the referral conversation alive.

Referrals are the direct answer to AI’s influence, because personal trust operates at the very first moment of the buyer’s journey, before AI ever enters the picture. It’s the strongest lever here, and also the hardest, because it requires real structural change to how the business generates and manages its pipeline, not just a new campaign.

If buyers are going to open ChatGPT or Copilot to shape a strategy, a budget, or a set of expectations before they ever talk to a person, that’s the moment worth trying to influence, not after the fact. Most companies hand AI the same thing they hand everyone else: brochureware and generic thought leadership. The opportunity is giving it something genuinely useful to work with instead.

The RFP builder is a concrete version of that: helping a buyer build the exact document they’d otherwise ask their AI to draft from scratch. Technology companies don’t typically help clients build a competitive RFP against themselves, but if the tool is robust enough to actually use, and the output is something the buyer can download and feed back into their own AI, you’ve done two things at once. You’ve made yourself useful earlier than a typical cold outreach ever could, and you’ve stayed embedded in the buyer’s process as it continues past that first interaction.