Lance Owide, general manager of B2B, Commerce | Image credit: Lance Owide
Editor’s Note: This interview with Lance Owide, general manager of B2B at Commerce.com Inc., explores how emerging agentic AI systems are reshaping digital commerce for distributors and manufacturers.
Unlike traditional rule-based automation, agentic AI can reason across complex data and act autonomously, enabling faster quoting, personalized guidance, and dynamic decision-making at scale. Owide explains how companies can build the right data foundation, integrate AI with existing platforms, and measure return on investment (ROI) — while maintaining trust, compliance, and security.
What distinguishes agentic from other AI?
Digital Commerce 360: How does agentic AI differ from traditional ecommerce automation, and why does it matter for distributors and manufacturers?
Lance Owide: Where automation saves time on repetitive tasks, agentic AI drives growth by creating better customer experiences and unlocking new revenue opportunities.
While traditional automation relies on static, rule-based workflows, like sending reorder notifications when stock drops below a threshold, or auto-approve orders under a certain value, these systems are reactive. They require manual configuration, don’t adapt easily to changing buyer behavior, and must be manually programmed to support the complexity found in modern B2B environments.
By contrast, agentic AI is goal-oriented, context-aware, and capable of acting autonomously. They can understand intent, respond dynamically to buyer behavior, and optimize over time without human intervention. This makes them particularly well-suited to B2B contexts, where selling processes involve long-tail catalogs, contract pricing, approval hierarchies, multi-channel fulfillment logic, and much more.
For distributors and manufacturers, the difference is transformative. Agentic AI can serve as an intelligent, always-on sales assistant, guiding customers through product discovery, generating quotes, recommending bundles, and even submitting orders. Unlike traditional automation, which may still require sales reps to manually process quote requests or decipher purchase orders, agentic systems can parse unstructured inputs, generate tailored responses, and only escalate to humans when thresholds are exceeded. This frees teams to focus on strategic work, while customers receive faster, more personalized service.
Agentic AI systems continuously learn. They observe which recommendations convert, which quotes are accepted, and where drop-off occurs, adapting their behavior to improve results over time. For distributors, where margins are tight and customer expectations are rising, this ability to intelligently orchestrate interactions and optimize workflows at scale offers a major competitive advantage. It’s not just about efficiency, it’s about moving from automation to autonomy, and unlocking a new tier of operational and commercial intelligence across the business.
How to lay a foundation for agentic AI
DC 360: What data foundations are needed — product, customer, pricing, inventory — to make agentic AI dependable instead of risky?
Owide: Agentic AI is only as strong as the data you feed it. If your foundation is fragmented or inconsistent, you create risks like wrong price quotes, wrong product recommendations, or poor customer interactions.
One of the most powerful aspects of agentic AI is its ability to ingest and interpret unstructured data, emails, PDFs, chat logs, product descriptions, spec sheets, and more, and turn that into actionable insight. Unlike traditional data management tools, which rely on predefined fields and rigid schemas, AI can understand context, infer intent, and extract meaning from messy or incomplete inputs. Rather than requiring everything to be perfectly structured, AI thrives on volume and variety. The more data we can give it, across formats, systems, and sources, the better it gets at pattern recognition, decision-making, and personalization. This makes agentic AI more adaptable to real-world complexity, where perfect data is rare, but insight is everywhere.
That’s why distributors and manufacturers need to get four pillars right:
- Product data: doesn’t need to be perfectly clean, but it must be deep and descriptive. Agents need access to technical specs, compatibility rules, rich attributes, product relationships, and even unstructured content like user reviews, FAQs, and documentation to make confident recommendations, substitutions, or comparisons, even when the buyer’s input is vague or incomplete.
- Customer data: unifies everything from account hierarchies and buyer roles to historical orders and entitlements. This allows the agent to interpret intent accurately, even if the request comes in via PDF or a casually worded email, and still respects contract terms, negotiated pricing, and buyer permissions.
- Pricing data: must be structured enough for rule-based enforcement. The agent doesn’t need perfect formatting, but it does need access to customer-specific price lists, discount logic, and margin thresholds, so it can apply pricing confidently without human review, and without risking over-discounting.
- Inventory data needs to reflect real-time availability, across warehouses, suppliers, and fulfillment methods. It needs reliable inventory and lead-time data to generate accurate promises and alternatives.
In short, agentic AI doesn’t necessarily require clean data, it requires complete, connected, and interpretable data across both structured systems and unstructured content. The more context we give it, from specs to sentiment, the more intelligently and autonomously it can operate.
Some tools already help the largest online sellers share product data with the likes of Google and Microsoft and have built new partnerships with AI tools like perplexity. For those looking to share product data directly with the AI search engines, exploring the tools available is a particularly good first step.
Managing complex B2B2 purchasing behaviors
DC 360: How can agentic AI manage complex B2B purchasing behaviors, such as contract pricing, approval workflows, and negotiated orders?
Owide: The beauty of agentic AI is that it can reason across complicated B2B commerce operations. Traditional automation struggles when you move beyond a simple catalog price and credit card checkout. But agentic AI can adapt to contract-specific pricing, recognize when an order needs approval before submission, and even handle negotiated or configured orders dynamically.
For approval workflows, agentic systems understand organizational hierarchies and role-based permissions. They can route purchase requests to the appropriate approvers, track status across multiple decision-makers, and follow escalation logic when timelines slip or exceptions arise. This orchestration happens dynamically, adapting to workflow complexity without requiring rigid, predefined paths. The AI can also learn from behavioral patterns over time, such as who typically approves what and when, and optimize routing accordingly.
When handling negotiated or custom orders, agentic AI can extract key terms from past correspondence, identify relevant history, and generate tailored proposals or quotes that align with buyer expectations. It can flag potential conflicts, such as discrepancies with standard terms, and prepare draft communications or quotes that a sales rep can quickly review and approve. This enables fast response times and a highly personalized experience, even when the buyer’s request falls outside the boundaries of standard catalogs or pricing.
How to understand when it’s working
DC 360: What measurable outcomes should be expected — faster quoting, higher conversion, reduced manual touches — and how is ROI tracked?
Owide: With agentic AI, you’re unlocking measurable business outcomes like improved efficiency, revenue growth and customer satisfaction because of improved service, and customer experience. AI also leads to faster quoting cycles, higher order conversion, and a significant reduction in manual touches on repetitive tasks like pricing checks or order re-entry.
Tracking ROI starts with baseline metrics. For quoting, measure cycle times before and after AI deployment; for conversion, look at win rates and abandoned quotes that are now close. For manual touches, track how many orders flow through without human intervention. Beyond that, monitor margin consistency, error rates, and even NPS or customer satisfaction scores because agents should be making the experience smoother for buyers.
ROI shows up both in cost savings and in the top line. Distributors and manufacturers who deploy agentic commerce well see the dual effect of a leaner sales process and happier customers who come back more often.
How agentic AI will change sales roles
DC 360: How will agentic AI reshape sales roles — inside reps, field sales, independent agents — and where should humans stay in the loop?
Owide: Agentic AI will reshape sales roles. Inside reps will spend less time on data entry and quoting and more time on initiative-taking outreach. Field sales will walk into meetings already armed with insights into pricing, availability, and customer history. Independent agents will get digital co-pilots that help them serve accounts faster without losing their personal relationship.
Humans remain essential in a couple critical places: building trust and handling exceptions. Agents can suggest the right offer, but customers still want a person to stand behind it. And when a deal involves negotiation, risk, strategic value, or other nuance, salespeople need to be in the loop to guide the process. In other words, AI manages the heavy lifting, but humans close the distance between transaction and relationship.
Risks to be aware of
DC 360: What risks around customer trust, compliance, and data security need to address before deploying AI agents in ecommerce?
Owide: Businesses need to think critically about three primary risk areas around trust, compliance, and data security.
If an AI agent gives the wrong answer on pricing, availability, or policy, it erodes confidence fast. Customers also need transparency regarding whether they are dealing with a human or an AI. Guardrails and human-in-the-loop escalation are essential to keep trust intact.
Many industries face strict regulations such as financial disclosures, safety standards, or export controls. An agent that improvises here could put a distributor in violation. We need AI guardrails that reference approved content and respect role-based permissions. We also need a paper trail to show how an answer was generated.
Finally, data security is key because agents need access to sensitive information like pricing rules, contracts, inventory, even customer PII, all of which are high value targets for attackers. Organizations already have security standards for ERPs and CRMs. AI needs to have access controls, encryption, and monitoring that matches or exceeds those.
Integrating agentic AI into ERPs and CRMs
DC 360: How is agentic AI integrated into existing enterprise resource planning (ERP), customer relationship management (CRM), and ecommerce platforms without disrupting current workflows?
Owide: Integrating agentic AI into ERP, CRM, and ecommerce doesn’t mean ripping and replacing what you already have. The goal is to augment, not disrupt, your current workflows. A few guiding principles:
- Agentic AI should connect through standard APIs, middleware, or iPaaS tools, so it reads and writes to ERP, CRM, and ecommerce without introducing custom code. That way, your “system of record” remains untouched.
- Instead of pushing bulk data jobs, the AI listens to events, like a new quote request in CRM, a stock change in ERP, or a cart update in ecommerce, and acts in real time. That reduces latency and keeps humans in the loop.
- Early deployments should mirror existing workflows: for example, the AI drafts a quote in CRM, but a rep approves it. Over time, as trust grows, you can dial up autonomy.
- AI works best when product, pricing, customer, and inventory data is normalized across systems. Building this foundation makes AI dependable and prevents conflicts between ERP vs CRM vs. ecommerce records.
- Start with narrow use cases and only expand to more complex tasks once integration proves stable.
The key is that agentic AI sits on top of the stack so you can adapt it without rewriting your ERP or retraining your teams.
Agentic AI use cases
DC 360: What are the best early use cases — such as guided product discovery, automated reordering, or supplier sourcing — that can deliver quick wins?
Owide: It’s important to start with using cases where the agent is augmenting existing workflows, not reinventing them. That way, the organization gets quick wins in customer experience and efficiency, while building trust in the tech.
- Guided product discovery. AI agents act like digital sales rep: surfacing compatible products, suggesting substitutes when items are out of stock, and tailoring recommendations to customer specs. This can quickly reduce cart abandonment and increase average order value with minimal integration work.
- Automated reordering and replenishment. The agent notices consumption patterns and prompts the buyer, delivering measurable efficiency by cutting manual touches and avoiding stockouts.
- Supplier and sourcing assistance. For distributors juggling multiple suppliers, an agent can automatically compare lead times, pricing, and availability, then tee up the best option, creating early gains from reducing time-to-quote and improving margins.
- Customer support triage. AI agents can resolve routine inquiries like order status, invoice copies, stock checks, freeing reps to focus on higher-value conversations.
- Quote configuration and approvals. Agents guide buyers through complex configurations and pre-validate against pricing/contract rules before sending them for human approval, which cuts down cycle time and errors.
Sign up
Sign up for a complimentary subscription to Digital Commerce 360 B2B News. It covers technology and business trends in the growing B2B ecommerce industry. Contact Mark Brohan, senior vice president of B2B and Market Research, at mark@digitalcommerce360.com. Follow him on Twitter @markbrohan. Follow us on LinkedIn, X (formerly Twitter), Facebook and YouTube.
FavoriteThe post Q&A: How agentic AI Is redefining B2B digital commerce appeared first on Digital Commerce 360.
from Digital Commerce 360 https://ift.tt/FBdyuWa
0 Comments