What commercial outcomes have you actually delivered?

The commercial outcomes start with a reality every procurement leader already knows: there aren't enough people to do the work properly. We have never met a procurement team with the manpower to manage every supplier relationship or sourcing event with the rigor and strategic focus they actually want to apply. The sheer volume of tasks leads to process inefficiencies, time pressure, and savings that go uncaptured. Not because anyone lacks the skill, because they lack the hours.

Deploying Natalie (and Ben, who we offer for indirect procurement) changes that equation. And the value shows up in a clear progression.

First, we establish an operational workflow for autonomous or buyer-in-the-loop negotiations. This alone is a step change. Suppliers that were getting a phone call once a year (if that) are now engaged through a structured, data-driven process.

Second, as Natalie and Ben take on more scope, they increase the effective capacity of procurement teams and reduce cycle times by cutting the overload on human buyers. Negotiations that took weeks to prepare and execute can be compressed significantly.

Third, manual workload drops as the AI Employees automate the repetitive administrative grind of preparing data-driven negotiations. Buyers get their time back for work that actually benefits from human judgment.

And as projects mature and previously disregarded negotiations begin to be handled with real analytical rigor, a trajectory for meaningful savings emerges. It builds as Natalie and Ben apply the same focus to mid- and long-tail negotiations that was previously reserved for strategic, high-value accounts only.

A few examples:

A leading European grocery retailer managing around $100 billion in revenue had a mid- to long-tail supplier base that wasn't getting the attention it needed.

Category managers didn't have the bandwidth to negotiate meaningfully with smaller branded suppliers more than once a year, and even then, preparation was thin. By deploying Natalie, they moved from reactive annual cycles to proactive, strategically timed negotiations, engaging when commodity prices dropped or competitive dynamics shifted, not when the calendar said so.

Twenty suppliers that previously received superficial annual attention now get comprehensive, data-driven negotiation coverage.

The retailer is already scaling Natalie to additional categories, targeting $10 billion-plus in broader tail spend.

Another example is from a regional CPG manufacturer (a dominant beer, beverage, and food producer in a single geography), achieved a 20% price reduction in a complex long-tail technical procurement category within the first month of deployment. For a business where procurement spend is 60 to 70 % of revenue and margins are tight, that kind of result in the first few weeks matters.

How does the system balance margin optimisation with supplier relationships?

This is the question everyone asks, and it's the right one. But it assumes a tension that doesn't actually exist the way people think it does.

Natalie doesn't squeeze suppliers for the sake of squeezing. She negotiates based on data. Commodity price movements, competitive benchmarks, volume economics, and market dynamics.

When she asks for a price reduction, there's a reason the supplier can see and verify.

So the “ask” is always grounded professionally, in data, not in emotion or through procurement being “tough”.

Most suppliers would rather deal with a counterpart who shows up prepared and transparent than a buyer who has not done their homework and behaves like a stereotypical “bully” from procurement who is there to beat up the supplier.

The guardrails matter here, too. Natalie operates within tone and conduct parameters set by the client, she's not sending aggressive emails at 2am or making ultimatums.

Her approach is to offer data-backed ideas and manage a structured back-and-forth conversation. If a negotiation hits a threshold that requires human judgment, say a long-term strategic supplier where the relationship carries weight beyond price, she escalates. She knows where her lane ends.

And there's a subtler point. Suppliers in the mid- and long-tail often have no real relationship with the buyer today. They get a PO once a year and maybe a brief call. Natalie actually increases engagement with these suppliers. For many suppliers, negotiating against Natalie is an upgrade from being ignored.

How does it handle volatility and perishability?

Natalie already incorporates commodity price movements and market intelligence into her negotiation timing and strategy. She knows when to engage because the data tells her conditions have shifted, not because a calendar reminder went off. That's core to how she operates today.

Perishability as a specific constraint, think shelf life-driven replenishment cycles or waste-adjusted pricing, isn't a primary use case right now. But the architecture supports it. Natalie's strategy engine is built to ingest whatever data drives the negotiation logic for a given category. If perishability windows or spoilage risk need to be factored into pricing and ordering decisions, that's simply a matter of expanding the elements we must configure for a deployment. And as we expand into fresh categories with our grocery clients, expect this to evolve.

What guardrails are in place to manage risk and compliance?

As grocery retailers begin to shift towards automating long-tail supplier negotiations, there are legitimate concerns about risk and compliance. At one of Europe's largest grocery retailers, the primary question was how agentic procurement technology could comply with company procurement guidelines and regulations, such as antitrust and cartel law. To conduct autonomous negotiations with external suppliers, we implemented guardrails to ensure compliance of our AI Employee Natalie.

Natalie has several layers of guardrails: legal, operational, technical, and user-in-the-loop.

Legal & Regulatory Guardrails: Above all else, Natalie has hard-coded constraints to comply with antitrust and cartel law. These are absolute prohibitions, active at all times. No exceptions, no edge cases, no grey areas. Furthermore, Natalie draws on the same body of antitrust compliance training that has been developed for human buyers over years, including real-world scenarios covering the full range of situations a negotiator actually encounters. Training agentic systems on these scenarios allows us to create a standardised, compliant, and auditable procurement process while also equipping Natalie to detect and alert for potentially critical supplier behaviour.

Operational Guardrails: Together with our clients, we set operational guardrails that define clear escalation paths for buyer intervention. These include specific price anchor scenarios, constraints on tone of voice to ensure a professional brand voice, and limiting the scope of all external communication solely to price negotiations while redirecting any attempts by suppliers to manipulate or broaden the negotiation.

Technical Guardrails: Our AI Employees are equipped with case-specific technical rules, including automated cross-referencing against source data and rule-based validation systems, that prevent them from straying off-track or hallucinating. For example, our European grocery retail client requires strict adherence to data-driven arguments. So all communication is continuously checked against the relevant data to ensure correctness. Natalie is inherently restricted from taking on tasks outside her capabilities or those not supported by available data. She knows what she doesn't know.

User-in-the-Loop Guardrails: For high-stakes roles like supplier negotiations, starting with user-in-the-loop negotiations assisted by Natalie is recommended. This positions the buyer as the ultimate decision-maker and point of control. Trust builds over time as the buyer watches Natalie operate within the guardrails, sees the audit trail, and gains confidence in the system's judgment. The goal isn't to remove humans from procurement. It's to let them focus on the work that actually requires human judgment rather than grinding through hundreds of tail-spend negotiations that follow predictable patterns.

Need more insights on agentic AI

Agentic AI is set to transform the food and grocery industry from supplier negotiations and procurement to merchandising and pricing, while reshaping how consumers shop through AI-driven purchasing decisions.

To help industry understand the complexity, the direction of travel, and current case studies, we have created a comprehensive report called ‘Agentic AI and the future of shopping’ that subscribers can now access.

Share your story with the food & grocery industry

We welcome contributions from retailers, brands, and solution providers shaping the future of food and grocery. If you have a compelling case study or perspective on technology, operations, or shopper behaviour, we’d be keen to hear from you.

Please get in touch with [email protected]