What AI Is Really Bringing to the Sales Process

Q: AI in sales is everywhere right now. Every tool claims to be "AI-powered." Where do you see the landscape today?

It's true — the market is flooded. And to be fair, a lot of these tools are genuinely useful. You have AI that helps reps write better emails, polish pitch decks, generate product brochures, personalize outreach at scale. Then on the other side, you have tools that scrape the web to find new contacts, enrich leads, build prospect lists. These are real capabilities, and they save time.

But if you step back and look at the big picture, you'll notice something: almost all of these tools operate at the edges of the sales process. They improve engagement — the quality of what you say — or they improve discovery — who you say it to. Neither of them helps you understand what's actually happening inside your pipeline right now.

Q: What do you mean by that? Isn't that what the CRM is for?

The CRM is a system of record. It captures snapshots — a stage, a close date, a dollar amount, a log of activities. But a deal isn't a snapshot. It's a process that unfolds over weeks or months. It has dynamics: it accelerates, it stalls, it shifts direction. The CRM doesn't capture any of that. It tells you where a deal is. It can't tell you where it's going.

And that's the gap. Sales teams have plenty of data about what happened. What they're missing is an understanding of the dynamics of their pipeline — which deals have healthy momentum, which are quietly stalling, how a rep's portfolio is trending, whether the team as a whole is moving in the right direction. That's the problem Dynamiks was built to solve.

Q: So how do you define "Impact" in the Dynamiks world?

Impact is what we call the elementary unit of sales. When something happens on a deal — a meeting, a proposal, a new stakeholder entering the conversation — it changes the deal's trajectory. Some events accelerate a deal significantly. Others barely register. Impact measures that change: the actual effect of what's happening on the likelihood of a deal closing.

Most sales organizations measure effort: how many calls, how many emails, how many meetings. We measure impact: how the deal's dynamics are shifting as a result of what's unfolding. It's a fundamentally different lens. Instead of asking "are my reps busy enough?" you start asking "are my deals moving in the right direction?" That changes everything.

Q: And "Momentum" — that's the other key concept?

Exactly. Momentum is about how a deal is evolving right now. Is stakeholder engagement expanding or contracting? Is the buying committee growing or going quiet? Are interactions accelerating or decaying?

Traditional pipeline management gives you a snapshot — a stage, a close date, a dollar amount. But deals are dynamic. They accelerate, they stall, they shift direction. Momentum captures that motion. It tells a rep or a manager not just where a deal is, but where it's heading. That's a fundamentally different kind of visibility, and it's one that simply doesn't exist in a CRM.

Q: What kind of AI makes this possible? This doesn't sound like a typical LLM application.

It's not. And that's an important distinction. Most AI in sales today is powered by large language models — great for generating text, summarizing calls, drafting content. But understanding the dynamics of a deal over time is a different class of problem. It requires modeling how deals evolve, detecting patterns across hundreds or thousands of historical outcomes, and learning what healthy deal dynamics look like versus what precedes a stall or a loss.

We use deep reinforcement learning — the same family of techniques used to master Go, optimize chip design, and control complex systems. It's uniquely suited to this because it's designed to learn from dynamic processes with delayed outcomes. A deal that closes in three months carries information about the patterns that unfolded weeks earlier. That's exactly the kind of temporal, sequential learning problem reinforcement learning was built for.

Q: Can you give a concrete example of what this looks like for a rep?

Sure. Imagine a rep has 40 open opportunities. In a traditional setup, they'd look at their CRM dashboard, see stages and close dates, and prioritize based on gut feeling and whatever came up in their last pipeline review.

With Dynamiks, they see something different. They see the dynamics of each deal. This one has strong momentum — engagement is building, the pattern resembles deals that have closed before. That one is losing steam — interactions are fading, the buying committee has gone quiet, the trajectory matches deals that historically went dark. And it's not just deal-level: they can see their own momentum as a rep — is their overall portfolio trending up or down? — and managers can see the same thing at the team level.

That's a completely different conversation in a pipeline review. Instead of "what stage is this deal in?" you're asking "what's the momentum here, and does the pattern look healthy?"

Q: How does this fit into the tools sales teams already use?

We integrate directly with Salesforce and HubSpot, and the primary interface for reps is Slack. There's no new tool to learn, no new dashboard to check. The insights surface where the work already happens.

This matters philosophically too. We believe in augmentation over automation. The goal isn't to replace the rep's judgment — it's to give them visibility they've never had before. The best salespeople have incredible intuition about their deals. What we do is give that intuition a quantitative foundation. You don't have to guess whether a deal is healthy. You can see it.

Q: There's a lot of skepticism about AI in enterprise sales. How do you address the "black box" concern?

I'd actually flip that question. The current process is the black box. When a sales manager asks "why did we lose that deal?" the honest answer is usually "we're not sure." When a VP of Sales builds a forecast, it's based on rep self-reporting and stage-based probabilities that haven't been validated in years.

What we're doing is making the invisible visible. Impact shows you the real forces at work in your deals. Momentum shows you where they're heading. That's not adding opacity — it's removing it. For the first time, you have an empirical, data-driven view of the dynamics of your pipeline.

Q: What does success look like for a team using Dynamiks?

Three things. First, clarity — reps and managers finally see the real dynamics of their pipeline, not just static snapshots. Second, consistency — you stop relying on a few star performers' instincts and start giving the whole team the same depth of visibility into how their deals and their portfolio are actually moving. Third, learning — every deal that closes or doesn't close makes the system smarter. The model learns from the team's collective experience in a way no individual rep ever could.

Most sales tools help you do more. We help you see what's really happening.

Q: What's next for Dynamiks?

What we're building is a multi-agent system that leverages the power of large language models and temporal difference learning to model and optimize the sales process. Today, we deliver Impact and Momentum insights — what's driving your deals, and how they're evolving. The roadmap goes deeper into strategic intelligence, team-level patterns, and coaching insights. But the vision is clear: we're laying the first stones of the self-driving pipeline.

Nicolas Maquaire is the Co-Founder and CEO of Dynamiks.ai, where AI sales agents are created, trained, and deployed to operate across your pipeline. Based in San Francisco and Paris, Nicolas previously founded EntropySoft, which was acquired by Salesforce.