Step 6 of the standard sourcing process—sending RFQs, negotiating terms, and contracting—can be radically enhanced through AI-powered chatbots and analytics platforms. Below, we explore key insights from a pioneering study on generative AI in buyer–supplier negotiations and translate them into actionable guidance for sourcing professionals. This blogpost supports the online course Sourcing Process 2B.
Table of Contents
1. Why AI for Negotiations?
Procurement sits at the intersection of vast internal and external data flows: spend analytics, supplier performance metrics, market indexes, and more. Generative AI chatbots—trained on domain-specific data—can process this information in real time to conduct or support negotiations, freeing buyers to focus on strategy and relationship management .
A clear early adopter example is Walmart’s use of Pactum’s negotiation chatbot to manage tail‑end spend. By automating negotiations over non‑critical items, Walmart achieved an average 3 % savings and extended payment terms to 35 days without adding headcount .
2. Economic vs. Relational Outcomes: Competitive vs. Collaborative AI
Herold et al. ran three experiments (two with students, one with procurement professionals) in which a ChatGPT‑based chatbot (“ProcureBot Charlie”) negotiated on behalf of a buyer. They compared two prompting styles:
- Competitive approach (distributive, hard‑ball tactics)
- Collaborative approach (integrative, relationship‑building tactics)
Key findings:
- A competitively prompted chatbot secured larger price discounts, better payment terms, and shorter negotiation times—closely matching its 10 % discount target—across all spend types .
- A collaboratively prompted chatbot, however, generated higher levels of supplier trust, greater outcome satisfaction, and a stronger desire for future interaction .
These results underscore a fundamental trade‑off: economic efficiency versus relational capital.
3. Aligning AI Tactics with Kraljic Quadrants
To decide which AI “persona” to deploy, map the spend category to the Kraljic Portfolio Matrix:
- Non‑critical items (low profit impact, low supply risk):
Automate with a competitive AI agent to drive savings and speed, since the risk of supplier fallout is minimal. - Leverage items (high spend, many suppliers):
A competitive AI can exploit buying power, but you may choose a hybrid prompting style to preserve supplier relationships when necessary. - Bottleneck items (low spend, high risk):
Here, maintain supply security. A collaborative AI agent that emphasizes partnership can help secure continuity even at a slight premium. - Strategic items (high value, high risk):
These demand human‑led, integrative negotiations. AI serves best as an analytical coach, running “what‑if” scenarios on price, payment terms, and risk factors, rather than fully autonomous negotiation .
4. Practical AI Use Cases in Step 6
- Initial RFQ analysis: AI parses incoming bids, flags outliers, and ranks offers against your weighted criteria.
- Autonomous tail‑end negotiations: For thousands of low‑value contracts, AI can negotiate price and terms within pre‑set thresholds.
- Negotiation coaching: AI bots suggest counter‑offers, provide real‑time “should‑cost” insights, and even role‑play supplier responses to sharpen buyer readiness .
- Contract drafting and review: Natural‑language models draft standard contract clauses and flag deviations from preferred templates.
5. Implementing AI Responsibly
- Data readiness: Clean, structured data on past contracts, payment performance, and supplier profiles is essential.
- Prompt design: Carefully craft AI instructions—determine whether the bot should be “aggressive” or “collaborative,” and embed your organization’s guardrails.
- Human oversight: Always have a qualified buyer review AI proposals, especially for high‑risk or strategic negotiations.
- Change management: Train your procurement teams to trust and interpret AI outputs, and foster a culture of experimentation.
6. Pitfalls and Mitigation
- Over‑automation risk: Overreliance on competitive AI bots can erode supplier goodwill.
- Model bias: AI trained on historical data may perpetuate sub‑optimal negotiation tactics. Regularly update prompts and retrain models.
- Transparency: Disclose to suppliers when they’re interfacing with an AI agent—ethical norms and regulations increasingly demand it.
7. Looking Ahead
As AI models evolve, we can anticipate more nuanced negotiation behaviors—adaptive tactics that shift from competitive to collaborative within a single dialogue, for instance. The ultimate goal is a human‑AI partnership, where AI handles data‑intensive, repeatable aspects of Step 6, and buyers focus on strategy, stakeholder alignment, and sustaining long‑term supplier relationships.
Interesting reading from Pactum.
By integrating AI into your negotiation arsenal—guided by empirical insights and aligned to your sourcing strategy—you can accelerate Step 6, drive savings, and preserve the human touch that underpins trusted supplier partnerships.
Note: Illustration to blogpost “How AI Is Transforming Negotiation and Contracting (Step 6) in Strategic Sourcing” was created by SORA on April 18, 2025.