What does Rob Smith mean by 'human Band-Aids' in retail organizations?

Rob defines 'human Band-Aids' as manual, unwritten workflows used by staff to fix gaps in data or technology. He argues that these processes must be codified and written down before they can be effectively automated by AI.

Expert Perspectives

Episode 119

OSF Digital: Why AI exposes every weakness in your ecommerce data

Q: What is the Google Universal Commerce Protocol (UCP)?

Rob Smith explains that UCP is a standardized protocol allowing AI tools to connect directly to a retailer's website to pull catalog data and create orders, essentially making agentic shopping technically feasible.

Q: How does AI change the way customers search for products?

Search is shifting from simple keywords to complex 'intent-based' queries. Instead of searching for a 'jacket,' customers provide context like location, temperature, and preference, requiring retailers to provide much richer product data to AI agents.

Podcast

Rob Smith

SVP of Commerce

In this episode we talked about:

Why AI reveals data and process weaknesses that humans have been quietly covering for years
How to audit your product pages and why most retailers have only one-third of the content AI agents require
The difference between platform limitations and poor implementation or UX strategy
Why commoditized commerce platforms shift competitive advantage to content, data, and discovery strategy
How to determine your discovery-to-owned-experience ratio based on your category and brand model
What the “human Band-Aid” problem is and why undocumented workflows block enterprise AI adoption
Why AI-first works for startups but requires process codification inside large organizations

🎧 Listen now on Apple Podcasts, Spotify, or YouTube

Episode highlights:

[02:39] The evolution from early ecommerce to agentic commerce

[05:57] What Google’s universal commerce protocol means for retailers

[12:11] Replatforming myths and performance issues

[16:13] The data, content, and context retailers are missing for AI-driven discovery

[22:55] Where AI shopping experiences will live: search engines vs. brand sites

[27:18] Why internal workflows and “human band-aids” block true AI adoption

Rob’s Bottom Line: AI doesn't just create new opportunities; it exposes the structural and data weaknesses that have been hidden for years. If you want to win in the era of agentic shopping, you have to move beyond "human Band-Aids" and codify your processes and product data for machines to understand—not just for humans to skim.

Rob Smith — Transcript

The Ecommerce Toolbox: Expert Perspectives • Human-Reviewed Transcript

[00:00:00] Rob Smith: I think what retail and commerce are missing, though, is just how much more data is necessary to answer the kind of questions that AI is asking, or customers will ask via AI, right? So, we move from this really keyword-driven search. The conversation's so much more nuanced.

[00:01:00] Rob Smith: I was a web developer, right? When the original wave of web development appeared... I have the same feeling, by the way, now with AI. Everything else in between has just been a bit incremental. I started an agency... we picked one platform that we were gonna focus on completely, and that ended up being Demandware... which became Salesforce Commerce Cloud.

[00:03:20] Rob Smith: We've gone through a few cycles. We had the early days of the web... then mobile commerce... then the wave of social commerce. But now, agentic shopping has suddenly become both technically feasible and quite straightforward to implement. Every retailer I speak to is excited, nervous, and also the traffic is still a small percentage of revenue. The future is not everywhere yet.

[00:05:49] Rob Smith: The announcement from Google was that they brought out something called UCP, Universal Commerce Protocol. OpenAI has ACP. These are ways their AI tooling can connect to your website and create an order effectively. Google CEO, Sundar Pichai, went to NRF to make that announcement with Walmart. That was a big statement of intent.

[00:09:27] Rob Smith: When AI hits the detail of an organization, they have an operational readiness problem. AI exposes all of their weaknesses that have always been there, but they've been hidden by humans. Either the customer is getting a bad experience, or the store colleague is having a hard time. All of these are humans suffering at the hands of data issues.

[00:11:42] Rob Smith: We've become really boring in ecommerce. Every website looks the same. We forgot that customers actually need to find things out. I see plenty of sites where my questions are not being answered about the product. Context is missing. If you replatform, you take your business with you to the new platform, and it underperforms because it wasn't the technology in the first place.

[00:16:16] Rob Smith: What retail is missing is how much more data is necessary. We move from keyword-driven search like 'winter snow jacket' to nuanced intent: 'I'm looking for a jacket for skiing at this specific place, I want to be warm.' AI reads the long descriptions we were told to cut for conversion. This bridge content is coming back in force.

[00:23:17] Rob Smith: High-frequency business like grocery will be app-based, but commodity businesses like batteries will be heavily disrupted by search tools. Customers need to know who they are and why they're here. Amazon is building its own Rufus AI because it is threatened by this. Someone not opening Amazon for a search is mildly terrifying for them.

[00:25:26] Rob Smith: Companies have so many 'human Band-Aids' inside them. I talked to a contact center and asked if they follow the written process. They said, 'No, we just onboard people and they figure it out.' Unless you codify that process properly, it can't be AI-driven. AI requires you to remove the Band-Aid and write the rules down.