Connected Intelligence, Agentic Reality
Retail’s new operating system from the CDAO Summit Boston—through the lens of HI + AI = ECI™
By Carsten Krause — October 24, 2025
If you want to see where AI is actually paying rent, follow retail. The CDAO Summit in Boston made one thing painfully clear: intelligence only compounds when data products are connected across pricing, funding, ordering, and customer touchpoints—and the teams, processes, and guardrails move in lockstep. That’s the spirit of Elevated Collaborative Intelligence™ (ECI): Human Intelligence plus Artificial Intelligence, multiplied by Technology Readiness, minus Risk Impact.
One of the biggest moments at the CDAO Fall Boston Summit this year was the opening panel with heavy hitters from DoorDash, Dtaftkings, Amazon, DiDi and Best Buy.
The discussion ranged from practical applications of AI at these lighthouse companies leading the way and also exploring agentic AI trends.
Here is what the panel actually said (and why I think it matters)
Rajesh Sura (Amazon) cut straight to the point about connected intelligence and metadata as the real AI substrate:
“Your intelligence becomes [most valuable] when all the data products are connected… pricing changes can impact vendor flow funding, in-store ordering, and competitive positioning. The intelligence needs to be connected across systems and acted upon. Interoperability is pushing this shift… metadata should be seen as an AI foundation… you are building lineage, ownership, and knowledge. AI is not just rendering the pixels—it is reasoning over your data.”
That “reasoning over your data” line is the tell. If the metadata isn’t trustworthy and complete, the models hallucinate, and you’re left in dashboard theater—pretty visuals, zero decisions.
Best Buy recommendatins as part of the customer journey mapping exercise:
“real-time recommendations in the Best Buy app, understanding product and customer data with traditional ML and generative AI, and consulting internally to unlock use cases. GenAI lets us capture richer sequential histories—entire browsing patterns across days—and transform that into better personalization.”
On the other side of retail-as-entertainment, DraftKings’ Vatsal Modi described why real-time matters when the “price lever” is promotions, not discounts:
“We optimize how we invest in customers. Real-time signals change the best action… reinforcement learning personalizes experiences on the fly—convert a single bet into a parlay, suggest another game, and so on. Just as important, we detect harmful patterns like chasing losses so we can intervene responsibly.”
When I asked the panel how agentic and multimodal AI are showing up in real use cases, the answer centered on workflow, context, and human fail-safes. The Amazon perspective again:
“Agentic AI builds context from informed metadata and triggers downstream actions. It runs inside the core workflow (not a sidecar). It follows thresholds—below a threshold it resolves autonomously; above it, escalate to a human. Knowing when to stop reasoning and escalate is critical.”
Two realities fell out of that exchange:
You can’t bolt agents onto broken processes. Fragmented data will cap your agent ROI no matter how clever the models.
Another panelist working with Didi the china counterpart of Uber laid out what happens when you take agents seriously at an organizational level:
“In my team of 70 people, we also have 70 agentic AIs—Zoe, Mark, Jamie—embedded in data enablement and governance. We meet with them. It’s changing the org model. Some employees fear replacement, so leadership must frame agents as freeing people to do more creative work.”
That cultural note came up repeatedly. Adoption splits into two tracks: customers adopt AI when service gets better (same-day delivery, instant refunds); employees adopt AI when it shows up inside their existing workflow and measurably reduces friction.
And metadata governance? The closer:
“Informative metadata is core. Without it, AI hallucinates… Leaders shouldn’t see governance as policy—they should see it as a data product that enforces consistency and can block or route workflows when eligibility or constraints change.”
Zooming out: trends that reinforce the panel’s signal
Agentic AI is real—and messy. Gartner’s latest view highlights both hype and trajectory: it warns that more than 40% of agentic AI projects could be scrapped by 2027 due to unclear value and cost overruns, even as it projects that by 2028, 15% of day-to-day business decisions will be made autonomously and 33% of enterprise applications will embed agentic capabilities.
Amazon is doubling down: AWS created a new group focused on agentic AI earlier this year to enable systems that can perform tasks without prompts—automation that fits the panel’s “threshold and escalation” design.
For enterprise context, Gartner placed “Agentic AI” on its Top Strategic Technology Trends list for 2025—firms need architected pathways for autonomy, safety, and oversight.
Where does this intersect with retail? E-commerce is already preparing for a world of shopping agents that search, compare, and even transact, shifting SEO and merchandising from human-facing screens to model-facing schemas. If your product data and policies aren’t “legible” to agents, you won’t be discovered.
Meanwhile, the operational backbone continues to modernize. Best Buy publicly details its use of Vertex AI to power real-time features and humanize support, consistent with the panel’s personalization narrative.
Finally, on productivity: rigorous studies keep stacking up. Stanford/MIT showed a 14% productivity lift for customer support agents using AI assistants, with the largest gains for less-experienced staff (a classic ECI effect: HI elevates where AI scaffolds). Nielsen Norman Group’s multi-study analysis reports a 66% throughput gain for business users; its companion study cites a 126% boost for programmers using AI pair-programming tools.
From buzzwords to build-able patterns
Let’s translate the panel into four patterns your teams can implement—each mapped to ECI.
Connect intelligence where money actually moves Tie pricing, vendor funding, in-store ordering, and competitive telemetry into a single decision mesh. Use domain-owned data products with explicit contracts (freshness, lineage, ownership, quality SLOs). Treat governance as a data product that can hard-stop transactions when constraints trip. This is textbook data mesh—domain ownership, data-as-product, self-serve platform, and federated governance. Build agents into the workflow, not beside it The Amazon example matters: agents read the dispute, choose an action, and only escalate when thresholds are exceeded. That’s ECI’s risk subtraction (R) in motion—automation plus human judgment. Leaders need escalation playbooks and audit trails from day one. Prioritize metadata that models can reason over “AI is not just rendering the pixels—it is reasoning over your data.” That means descriptive, structural, and policy metadata must be consistent and queryable. Without informed metadata, your agents will hallucinate, your retrieval will miss the mark, and your oversight won’t catch drift. Real-time, sequential context wins personalization DraftKings and Best Buy both point to a shift from batch profiles to streaming context: micro-events, reinforcement-learning policies, and generative summaries of session histories. Infrastructure matters: low-latency feature stores, event buses, and guardrails for responsible actions (e.g., responsible gaming interventions at the edge).
The ECI Perspective: HI + AI = ECI™
The ECI equation is simple on paper—ECI = (HI + AI) × T − R—but demanding in practice.
HI (Leadership & Strategy): clarify decision rights, escalation thresholds, and the north-star outcomes you will measure (not just dashboards, but decisions). AI (Technology & Automation): design agents as decision participants with auditable memory, tool access, and safe-fail behaviors. T (Technology Readiness): harden the substrate—feature stores, lineage, policy enforcement, and observability—so models can reason and teams can trust. R (Risk Impact): bake in responsible-use detections (e.g., DraftKings’ patterns such as chasing losses), policy blocks, and human takeover points.
When you multiply HI+AI by real readiness (T), you create compounding effects: faster cycle times, fewer hand-offs, safer operations. When you ignore R, you accumulate hidden liabilities that erase the gains.
What success looks like (and what it won’t)
What it looks like
• Customer: less friction, more relevance, and faster resolution—customers don’t care that it’s AI; they care that the refund is same-day and the recommendation fits.
• Employee: AI inside the current system of work; fewer copy-paste loops; more time for creative, strategic tasks.
• Platform: clean metadata, agent policy packs, and human-in-the-loop escalation.
• Leadership: experimentation velocity with statistical, not anecdotal, learning.
What it won’t look like
• A sidecar chatbot.
• A single-vendor “agentic platform” flipped on by procurement.
• A hero model duct-taped to 12 brittle pipelines.
• A culture that declares “automation” while starving governance.
Field notes you can apply tomorrow
Embed agents where loss or latency is expensive: disputes, returns, vendor onboarding, price changes, and promo optimization. Treat metadata like code: owners, SLAs, tests, versioning, and promotion gates to production. Move from “search logs + 30-day windows” to streaming context with generative summaries—your personalization teams will discover new signals. Put the escalation thresholds in writing, not tribal memory. Audit everything agents do. Break the “build it and they will adopt” habit: show employees how the agent trims their least-favorite work, inside the tools they already use. Accept the conversion cost: getting from 90% to 99% accuracy takes disproportionate effort—budget and plan for it. (As the panel put it: every “nine” costs about the same as the previous nine.) Separate customer adoption (service quality) from employee adoption (workflow integration). Measure both.
The CDO TIMES Bottom Line
Executive summary
Connected intelligence is now table stakes. The Boston panel underscored what the research confirms: agentic AI pays off only when it’s built on informed metadata, embedded in the core workflow, and bounded by human-in-the-loop escalation. Treat governance as a product, not a PDF. Put GenAI to work where sequential context changes the outcome—returns, disputes, promotions, and high-stakes customer interactions.
Next steps for leaders
Name your first three agentic workflows (e.g., financial disputes, returns/refunds, vendor onboarding). Define thresholds and escalation. Stand up a “governance as product” squad: metadata ownership, quality SLOs, policy enforcement that can block unsafe actions. Move from batched profiles to streaming context: instrument micro-events; add reinforcement-learning policies where incentives matter. Make adoption two-track: customer service SLAs and employee friction scores. Improve both, weekly. Budget for the “last nine”: 90→99% accuracy takes real time and money; plan staged rollouts with guardrails. Replace “demo theater” with experiment portfolios: small bets, fast cycles, and clear kill criteria.
Resources
• My book, The AI-Ready Leader: HI + AI = ECI™ — practical playbooks for agentic AI, governance-as-product, and measurable transformation (available on Amazon and at https://cdotimes.com).
• CDO TIMES ongoing coverage and executive toolkits: frameworks, canvases, and case studies updated weekly at https://cdotimes.com.
If you want your org to ride this wave instead of getting dragged by it, build ECI into the bones: human judgment plus machine autonomy, multiplied by real readiness, minus unmanaged risk. That’s how connected intelligence turns into compounding outcomes.
Research links mentioned (beyond the panel)
Gartner via Reuters on agentic AI project outcomes and 2028 outlook, https://www.reuters.com/business/over-40-agentic-ai-projects-will-be-scrapped-by-2027-gartner-says-2025-06-25/
AWS agentic AI group (Amazon), https://www.reuters.com/technology/artificial-intelligence/amazons-aws-forms-new-group-focused-agentic-ai-2025-03-04/
McKinsey, State of AI 2023 (genAI adoption), PDF, https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our%20insights/the%20state%20of%20ai%20in%202023%20generative%20ais%20breakout%20year/the-state-of-ai-in-2023-generative-ais-breakout-year_vf.pdf
Stanford HAI/MIT: 14% productivity in call centers, https://hai.stanford.edu/news/will-generative-ai-make-you-more-productive-work-yes-only-if-youre-not-already-great-your-job
Nielsen Norman Group: 66% throughput gain; 126% for programmers, https://www.nngroup.com/articles/ai-tools-productivity-gains/ and https://www.nngroup.com/articles/ai-programmers-productive/
McKinsey primer, https://www.mckinsey.com/capabilities/quantumblack/our-insights/demystifying-data-mesh
DraftKings responsible gaming page (loss-chasing guidance), https://www.draftkings.com/responsible-gaming-about
Love this article? Embrace the full potential and become an esteemed full access member, experiencing the exhilaration of unlimited access to captivating articles, exclusive non-public content, empowering hands-on guides, and transformative training material. Unleash your true potential today!
Order the AI + HI = ECI book by Carsten Krause today! at cdotimes.com/book
Subscribe on LinkedIn: Digital Insider
Become a paid subscriber for unlimited access, exclusive content, no ads: CDO TIMES
Do You Need Help?
Consider bringing on a fractional CIO, CISO, CDO or CAIO from CDO TIMES Leadership as a Service. The expertise of CDO TIMES becomes indispensable for organizations striving to stay ahead in the digital transformation journey. Here are some compelling reasons to engage their experts:
- Deep Expertise: CDO TIMES has a team of experts with deep expertise in the field of Cybersecurity, Digital, Data and AI and its integration into business processes. This knowledge ensures that your organization can leverage digital and AI in the most optimal and innovative ways.
- Strategic Insight: Not only can the CDO TIMES team help develop a Digital & AI strategy, but they can also provide insights into how this strategy fits into your overall business model and objectives. They understand that every business is unique, and so should be its Digital & AI strategy.
- Future-Proofing: With CDO TIMES, organizations can ensure they are future-proofed against rapid technological changes. Our experts stay abreast of the latest AI, Data and digital advancements and can guide your organization to adapt and evolve as the technology does.
- Risk Management: Implementing a Digital & AI strategy is not without its risks. The CDO TIMES can help identify potential pitfalls and develop mitigation strategies, helping you avoid costly mistakes and ensuring a smooth transition with fractional CISO services.
- Competitive Advantage: Finally, by hiring CDO TIMES experts, you are investing in a competitive advantage. Their expertise can help you speed up your innovation processes, bring products to market faster, and stay ahead of your competitors.
By employing the expertise of CDO TIMES, organizations can navigate the complexities of digital innovation with greater confidence and foresight, setting themselves up for success in the rapidly evolving digital economy. The future is digital, and with CDO TIMES, you’ll be well-equipped to lead in this new frontier.
Do you need help with your digital transformation initiatives? We provide fractional CAIO, CDO, CISO and CIO services, do a Preliminary ECI and Tech Navigator Assessment and we will help you drive results and deliver winning digital and AI strategies for you!
Subscribe now for free and never miss out on digital insights delivered right to your inbox!
