Rob Want

The owner learned each customer's hobbies and background, asked what the machine was for in the short and long term, then made a recommendation.

• That approach produced loyalty and repeat business.
• Want argues that in-person service cannot scale beyond one store, so retailers should extend it with conversational AI: assistants for frontline workers, who already use generative AI in daily life and might otherwise turn to unsanctioned tools, and agents on the retailer's own digital channels, grounded in its data.

A customer whose daughter was heading to law school was steered toward a larger hard drive and told to skip an expensive video card. Many customers kept coming back and paid more than at a cheaper competitor that later opened down the road.

Read original

Want warns that if a third-party AI app makes the journey easier, loyal customers who start on the brand's channel will move there. Running conversational shopping on your own site is framed as a commercial control decision: the retailer keeps the customer relationship and conversation data, meets expectations it has already set, and grounds answers in its unique knowledge.

A Shoptalk session recap notes that retailers must offer natural-language and visual search on their own ecommerce and digital channels, or third-party generative AI apps will own the experience.

Read original · Original 2

The practical recommendation is to start with the decision that needs improving, apply AI selectively, and connect enterprise data to surface insights.

Connect an enterprise data warehouse to surface insights first, then target advanced uses such as inventory optimization and quality control. A related post identifies the operating model and fragmented data as the biggest barriers to factory transformation.

Read original · Original 2

Claimed benefits include lower capital spending through standardized designs, faster modular plant setup, simulation before deployment, and optimized layout, flow, and throughput. Two companion posts widen the argument: one gives improvement ranges and lays out a five-level data-maturity path from lean basics to flexible, fully autonomous plants aligned to ISA-95; another says the biggest barrier is not technology but the operating model and fragmented data, and describes unifying engineering, operations, quality, and maintenance behind one view with recommended actions.

Simulate and fix line issues before they happen, or replace separate functional dashboards with a single unified view and recommended actions.

The earlier post treats digital twins mainly as a tool for capital projects; later posts broaden the case to organization, data maturity, and the shift from reactive operations toward an autonomous enterprise.

Read original · Original 2 · Original 3

The proposed response is agentic AI in three roles: identifying real-time conflict updates and related market data and tying them to forecasting models for the products a company makes; suggesting mitigation strategies that worked in similar past crises; and modelling likely scenarios.

Modelling resolution, prolonged disruption, negotiation breakdown, and infrastructure impacts as separate scenarios. The post does not describe a specific company's recovery, only the risk and the tooling.

Read original

It explains why they see AI as a way to scale human selling rather than replace it.

Rob Want tells a family retail story: in the 1980s and 1990s, their father sold personal computers and office supplies, and would first ask customers about hobbies and background, then about short- and long-term needs.

• The example is specific: for a daughter going to law school, the advice was a larger hard drive and no extra spend on a fancy video card.
• Customers returned and many paid more than at a cheaper competitor down the road.
• Rob Want argues that this consultative approach should be scaled with technology, including conversational assistants for frontline workers and customers, because shopping journeys now often begin in GenAI apps.

For brands planning a campaign

Useful for CPG or retail consulting briefs on frontline enablement, clienteling, and conversational commerce. The post offers a concrete service principle a brand can test against its store or ecommerce experience.

Newcomers get a compact view of where they think AI pays off in supply chain.

Rob Want summarizes a Shoptalk 2026 supply chain session and reports four themes: AI value comes from focus, not scale; data quality and context matter more than models; retail and supply chain lead AI adoption because of direct ROI in fulfillment, demand planning, and operations; and non-technical leaders can now build agents and prototype quickly. The practical insight: start decision intelligence with the decision that needs improvement, then apply AI selectively, often by connecting an enterprise data warehouse to surface insights.

For brands planning a campaign

Good starting point for a CPG supply chain engagement that needs prioritization. The post supports a workshop or assessment focused on a named decision and data readiness, rather than broad AI ambition.

It shows the outcomes they put in front of a manufacturing decision-maker.

Rob Want lists potential CPG manufacturer improvements from digital twins and data-driven operations: innovation +25-40%, service +4-6%, capex -15-20%, opex -20-30%, and sustainability -10-12% CO2.

• The post describes a five-level data maturity journey, from lean basics and digitization to closed-loop operations, autonomous operations, and plants designed for flexible, fully autonomous operations, referencing ISA-95.
• One example: simulating and fixing line issues before they happen.

For brands planning a campaign

Useful for a CPG operational efficiency brief or business case. The maturity levels can structure a phased digital twin roadmap, while the metrics offer a first-pass value hypothesis to validate.

It explains the before-you-build use case.

Rob Want describes leading CPG companies adopting model-based engineering and digital twins.

• Instead of testing line performance after installation, they simulate and validate the line virtually first.
• Stated benefits include reduced capex through standardization, faster plant setup through modular design, performance simulation before deployment, and optimization of layout, flow, and throughput.
• The outcomes they emphasize are faster time-to-market and lower risk.

For brands planning a campaign

Relevant for manufacturing clients planning a new line, plant, or significant retrofit. A practical entry point: use virtual validation to test design and throughput before committing capital.

It explains what they think must change before autonomous operations are realistic.

Rob Want argues that the biggest barriers to transformation are the operating model and data fragmentation, not technology.

• The post says leading CPG companies unify engineering, operations, quality, and maintenance; build twin-centric intelligence; move to composable, event-driven platforms; and create a single source of truth.
• The example is a shift from separate functional dashboards to one unified view with recommended actions.
• The end state they describe is a move from reactive operations to an autonomous enterprise.

For brands planning a campaign

Useful for consulting engagements that include operating model and data governance, not just tool selection. A brand can use the unified-view example to test whether its teams share one operational picture.

It shows how Rob Want links market shocks to AI-supported forecasting and mitigation.

Rob Want warns that CPG products face short-term cost and downstream price pressure because of polyethylene dependency in plastics and packaging.

• The post notes that supplier contracts and inventory cycles can be short, roughly 2 to 8 weeks.
• They propose using agentic AI to identify real-time conflict updates and related market data, connect them to forecasting models for a company's products, suggest mitigation strategies that worked in similar crises, and model scenarios such as resolution, prolonged disruption, negotiation breakdown, and infrastructure impacts.

For brands planning a campaign

Relevant for CPG resilience and procurement briefs. The post suggests a concrete use case: combine external disruption signals with product-level forecasting and scenario planning for packaging materials.

At a Shoptalk session, Rob Want reports that user behavior is shifting toward multimodal and natural-language interactions, as well as visual search and discovery. Retailers risk ceding the experience to third-party GenAI apps unless they offer these experiences through their own ecommerce and digital channels, where they can retain the customer relationship and data.

For brands planning a campaign

Useful for retail digital experience reviews. It frames personalization and search as a data-ownership decision, not just a feature upgrade.

One post frames the shift as replacing post-installation line testing with virtual validation, listing benefits such as lower capital spending through standardized designs, faster modular plant setup, simulation of performance before deployment, and optimization of layout, flow, and throughput; the stated outcome is faster time to market with lower risk.

• A related post puts numbers behind the case: innovation improvement potential of 25–40%, service 4–6%, capital spending down 15–20%, operating expense down 20–30%, and CO2 down 10–12%, alongside a five-level data-maturity path from lean basics to fully autonomous flexible plants aligned to ISA-95.
• A third post says the biggest barrier is the operating model and fragmented data, not technology, and describes unifying engineering, operations, quality, and maintenance behind a single view with recommended actions and composable, event-driven platforms.

Uses “Instead of X, you get Y” contrasts, stage lists (the five levels of data maturity), and emoji-labelled benefit bullets.

The improvement percentages are not tied to a named plant, client, or study. Ask for the underlying evidence before presenting them as measured results for a specific facility.

For brands planning a campaign

For a CPG brand weighing transformation consulting, these posts lay out a clear sequence: standardize plant designs, simulate before building, then unify data across functions. The percentage ranges work as campaign hooks, but brands should verify them against their own sites before using them in client-facing material.

Examples: Source 1 · Source 2 · Source 3

It notes that supplier contracts and inventory cycles run only 2 to 8 weeks. Rob Want proposes three agentic AI uses in that situation: automatically identifying real-time conflict updates and market data tied to forecasting models for the products a company makes; suggesting mitigation strategies that worked in similar past crises; and modelling scenarios such as resolution, prolonged disruption, negotiation breakdown, and infrastructure impacts.

For brands planning a campaign

Suits a brief about supply-chain resilience. A brand could ask for a crisis-response framework built around those three agent roles — detect, recall prior mitigations, and model scenarios — while framing the posts as discussion of risk and tooling, not a documented client recovery.

Examples: Source 1

The post also states that Decision Intelligence starts with the decision that needs improving, then applies AI selectively, beginning by connecting enterprise data to surface insights, with inventory optimization and quality control cited as more advanced uses. A related post reinforces the same point from the other direction, naming the operating model and data fragmentation as the biggest barrier to factory transformation.

For brands planning a campaign

This is the clearest “start narrow” argument in the material and can anchor a consulting campaign positioned against broad AI transformation pitches. The point that non-technical leaders can build agents supports workshop, prototype, and live-demo formats over theory-led content.

Examples: Source 1 · Source 2

Customers returned, and many paid more than at a cheaper competitor that opened nearby.

• Rob Want argues that this consultative model cannot scale in person beyond one store, so retailers should extend it with conversational AI agents grounded in the right data: for front-line workers, who already use generative AI apps and may otherwise reach for unsanctioned ones, and on the retailer's own digital channels.
• The post gives three reasons to run conversational shopping on your own site: retailers keep the customer relationship and the conversation data; loyal customers starting on the brand's channel expect more and will move if a third-party app is easier; and retailers can ground answers in their own unique knowledge.
• A second post reports from a Shoptalk session that personalization expectations are shifting toward multimodal and natural-language interaction and visual search and discovery, with the same warning that third-party generative AI apps could otherwise own the experience.

For brands planning a campaign

This is the strongest first-person material in the set: a concrete retail story joined to a channel-ownership argument. A brand can ask for customer-experience content that starts with the consultative sale and ends in conversational commerce, or for a piece aimed at retailers deciding whether to build their own assistant rather than cede the shopping journey.

Examples: Source 1 · Source 2

Two cover a Shoptalk supply-chain session and a Shoptalk personalization session, one looks ahead to Shoptalk the week of 23 March, and two cover NRF 2026: one promotes hands-on retail AI demos across merchandising, customer engagement, operations, and supply chain at a booth and offers to schedule an AI lounge demo tour, while another describes a lounge experience for AI thought leadership across marketing, customer engagement, store operations, and supply chain. Together, they show a working pattern of attending, summarizing sessions, and inviting readers to meet on the floor.

For brands planning a campaign

Useful for event-adjacent content: pre-event agenda previews, on-floor notes, or booth and demo-lounge promotion in the retail and CPG calendar. Ask for confirmation of Rob Want's role or employer before naming a partner or company; a hashtag alone does not establish either.

Examples: Source 1 · Source 2 · Source 3 · Source 4 · Source 5

Numbers carry the argument, such as the improvement ranges for digital twins, the 2 to 8 week contract cycles in the packaging-cost post, and the improvement percentages in the manufacturing post.

• A recurring teaching move is the contrast pair — simulate line performance before installation rather than test it afterward; use one unified view with recommended actions rather than separate dashboards by function.
• The sequence is consistent: name the decision or constraint first, then the tools that address it, then the expected result.

For brands planning a campaign

Best suited to short, list-driven sponsored formats — single-post explainers, carousels, or session takeaways — with one concrete example per point. For a long-form asset, the brand should supply the structure rather than expect it to emerge from the existing posts.

Examples: Source 1 · Source 2 · Source 3 · Source 4 · Source 5

A second post amplified the lounge as "AI thought leadership" across marketing, customer engagement, store operations, and supply chain, with links and the #AvanadeDoWhatMatters tag. The posts don't specify the relationship; ask whether it's employment, client work or another commercial arrangement.

View example

Both treat a simulated plant as the practical route to less unplanned downtime and lower processing time.

Santi Adavani stays inside the engineering taxonomy, working through named levels of digital twins from descriptive modeling upward and using automotive examples such as physics-based vehicle simulation and VR showrooms. Rob Want frames factory simulation as an operating-model and capital-planning argument, covering data maturity and org structure but rarely going into the technical stack or the reference model behind it.

Read original · Original 2

Both warn that clean context, not a better model, is what makes AI pay off.

Mark Bidgood drives the point into data governance and ontology, naming MCP connections and boardroom-level agreement on revenue, customer, product, inventory, and margin.

• Their direct, argument-and-question style centers on one idea per post.
• Rob Want ranges wider across retail customer experience, event coverage, and plant operations, and usually keeps the data argument to a sentence or two inside a broader topic.

Read original · Original 2

Dani Goldman works at market level, tying AI to hyper-personalized consumer experiences and to sustainability and traceability, and explains architecture concepts such as the context layer, retrieval-augmented generation, and knowledge graphs. Rob Want is more concrete and event-anchored: a specific retail story, a session recap, or a named material and its contract cycle.

Read original · Original 2

The posts lay out a level-by-level model, starting with L1 descriptive twins as the 'blueprint phase': asset modeling of every component and subsystem, physics-based engineering design for vehicle dynamics and battery performance, and immersive VR showrooms for remote collaboration.

• The factory post explains what a twin ingests and how it is used, naming live data from PLCs, MES systems, IoT sensors and inventory databases.
• It lists use cases including production scheduling, bottleneck identification, layout optimization, predictive maintenance and workforce planning, and adds adoption context: 86% of manufacturing executives see practical applications and 44% are already implementing them.
• For a CPG manufacturer that accepts digital twins as the direction and needs help deciding scope — design-phase modeling versus live-floor scheduling versus maintenance prediction — this offers a more granular starting point than Rob Want’s roadmap-level framing.

The tradeoff is concentration and distance from the commercial problem.

• Both posts focus squarely on manufacturing and factory digital twins; neither addresses the retail-facing questions that matter to many CPG companies — personalization, conversational commerce, trade spend or retail media.
• Rob Want’s method starts by naming an operational decision and checking whether the operating model and fragmented data can support it.
• Santi Adavani’s posts move faster to twin capabilities and benefit claims, saying less about organizational readiness.
• The benefits are presented at industry level rather than tied to a named plant or client, so treat them as directional and validate them against the buyer’s own line data.
• This comparison covers two posts from May 2025 and characterizes a working position on digital twins, not the creator’s full range of work.

Examples: Source 1 · Source 2

The posts diagnose a direct failure mode: consumer goods companies deploy AI while still disagreeing internally about what 'product,' 'customer' and 'margin' mean, with marketing, supply chain, finance and retail partners each using different hierarchies.

• Mark Bidgood argues that layering AI onto that inconsistency produces trade spend optimized against the wrong signals, forecasts that are 'confidently wrong,' untrustworthy retail media ROI and agents that technically work but miss business intent.
• The recommended sequence is explicit: create unified definitions across customer, product, promotion and margin; connect retail, supply chain, ERP, CRM and trade systems in one governed architecture; give AI context rather than raw data; and build a trusted operational layer before deploying autonomous agents.
• A second post takes the argument to the boardroom, framing trusted data as a governance concern and noting that connected systems alone will not settle meaning.
• For a CPG company whose transformation has stalled on conflicting master data, or whose finance and supply chain teams report different numbers, this is a sharper, more diagnosis-oriented fit than a factory-roadmap conversation.

The tradeoff is that this is a foundation argument, not a delivery path.

• The posts identify what must be fixed and why, but offer no staged sequence for a manufacturer to act on next quarter: no maturity ladder, simulation pilot, capex or plant-commissioning angle comparable to Rob Want’s before-you-build framing.
• They are essay-style analyses of CPG data strategy, not implementation accounts; no named client, system or measured outcome appears.
• A buyer who agrees that definitions are messy but wants to know which plant decision to start with would need to bring that scoping themselves.
• This comparison covers two posts from May 2026.

Examples: Source 1 · Source 2

The reviewed trend post covers four shifts: AI moving from experimentation and dashboards into demand forecasting, supply-chain optimization, personalization and product innovation; consumers expecting hyper-personalized experiences and smarter promotions; supply-chain resilience becoming a competitive advantage as over half of CPG executives report rising challenges, from labor shortages to geopolitical disruption; and sustainability tied directly to business performance through traceability expectations from consumers, retailers and regulators.

• The second post supplies the architecture vocabulary behind those shifts: the 'context layer,' built on retrieval-augmented generation, knowledge graphs and data platforms, grounds a model in a specific business through governance and access controls.
• It also argues that agentic AI needs continuous structured context to execute workflows such as demand planning.
• For a CPG team that needs a shared frame across commercial and supply-chain stakeholders before committing to an advisor, this is broader and more current than a single-decision scoping session.

The tradeoff is altitude.

• The posts stay at trend and architecture level: they describe what leading companies are doing and what separates them, but do not walk buyers through a specific plant or line decision, a readiness assessment or a staged pilot with cost and service ranges.
• Rob Want turns a theme into a decision-first sequence; Mark Bidgood pins the failure to definitions; Dani Goldman explains the landscape and the plumbing.
• Buyers would still need to supply concrete project scoping.
• The posts are forward-looking commentary, not accounts of delivered work, and include no named client or measured result.
• This comparison covers two posts from May and June 2026.

Examples: Source 1 · Source 2