Pricing and Agentic AI: Towards a new strategic frontier
Elie Abitbol Senior Manager Pricing & Sales Excellence chez Converteo
As a Senior Manager within Converteo’s Pricing & Sales Excellence practice, Elie Abitbol supports our clients in transforming and optimizing their pricing strategy. He leads expert teams at the intersection of business strategy and advanced modeling, carrying the “AI & Pricing” vision to ensure the sustainable business impact of new models.
Key Takeaways
- Pricing is entering an era of autonomous execution, not just analytical optimization. The shift from predictive AI to agentic AI is transforming pricing from an analytical exercise into a real-time action capability. The challenge is no longer to find the “right price,” but to be able to adjust and execute it at the speed of the market. This repositions pricing as a strategic operational function at the heart of commercial competitiveness.
- The democratization of data is shifting internal decision-making power. Access to insights via natural language eliminates dependence on technical teams. Pricing and business departments become direct producers of analysis, which shifts value from tools to the quality of the business questions being asked.
- Competition is becoming agent versus agent. The emergence of buyers equipped with agents capable of automating tenders and continuously challenging prices requires a symmetrical technological response from sellers. Pricing is becoming a field of algorithmic confrontation.
What’s Really Changing: From Analytical Pricing to Pricing That Acts
We are leaving the era of data-assisted pricing and entering the era of agent-operated pricing.
Historically, Big Data and Machine Learning focused on prediction. Today, generative AI brings interaction and data understanding, while agentic AI introduces decision-making and action capabilities. The major disruption is therefore functional: pricing is no longer just an aid to human decision-making but is becoming a system capable of executing decisions under supervision. As the current situation highlights, price loses its value if you are not able to execute it at the speed of the market. The challenge is no longer to find the theoretical “right price,” but to adjust it in real-time.
"Price is nothing if you are not able to execute it at the speed of the market."
This evolution has major structural consequences:
- Acceleration of cycles: A radical reduction in sales lead times.
- End of data silos: Unstructured formats like PDFs or emails become directly usable.
- Agentic Commerce: The emergence of partially automated exchanges where competition becomes an “agent versus agent” confrontation. As buyers equip themselves with agents to automate tenders, sellers must respond with symmetrical technology.
The Obstacles and Paradoxes: Why Everyone Sees the AI Revolution… But Hesitates to Move
Despite technological maturity, several obstacles are slowing adoption. The first is the “data wall”: insufficient quality that undermines trust in algorithms.
There is also a central paradox: technology allows for autonomy, but organizations are not ready to delegate decision-making power. Defining the acceptable level of autonomy thus becomes a political and governance issue rather than a technical one.
Finally, the fear of AI hallucinations is blocking its use in critical processes, even when the potential return on investment (ROI) is high.
"We are moving from an AI that predicts to an AI that acts."
The Recommended Method: A Transformation Through Action
To break the inaction, a pragmatic transformation framework is recommended:
- Start small: Data must be made visible quickly, even if it is imperfect. Using simple dashboards exposes reality and creates a dynamic of collective improvement by making inconsistencies visible.
- Put people at the center: The involvement of business teams (sales, pricing, operations) is crucial from the start. AI must be perceived as an amplifier of human decisions, not as a replacement for field expertise.
- Learn before industrializing: It is necessary to freely test use cases via POCs (Proof of Concept) to understand the system’s limits before freezing processes.
- Move fast, but keep control: Prioritize cases with tangible ROI and gradually industrialize with the right partners, avoiding overly heavy or premature programs.
Concrete Cases: When AI is Already Creating Value
The impact of AI can already be measured through concrete applications:
- Promotional optimization: In industry, modeling elasticity has made it possible to eliminate 20% of promotions with zero ROI without a drop in revenue, shifting from a volume-based logic to a strategic allocation of discounts.
- Tender processing: For RFPs containing thousands of references, AI agents now read unstructured PDFs to suggest product matches. This drastically reduces response time and improves the conversion rate.
- Data analysis copilot: By allowing executives to generate analyses via natural language queries, AI eliminates dependence on technical teams and shifts decision-making power to the business departments.
