Christopher Comodikes

The team’s routing and scheduling software works well for day-to-day operational planning, but comparing service levels across several days felt cumbersome and slow.

• AI made the comparison quick to set up.
• They also identify limits planners should weigh: no road-map view, only scatterplots of geolocated stops; no manual fine-tuning through a graphical planning interface; and the risk that a new dataset and prompt could send the model in a very different direction, making it hard to redirect without restarting.
• Their takeaway: AI is a credible backup if supply chain software goes down and a quick way to explore advanced concepts without lengthy setup.

Comparing a standard service level with more frequent customer deliveries, Gemini said route clustering and optimisation required specialised software.

• ChatGPT clustered deliveries into routes, sequenced stops and produced finish times to check feasibility.
• Its outputs aligned surprisingly well with the routing system’s base cases, though Comodikes notes that hallucinations can still lurk in the results.

Read original

They added a missing depot address, corrected vehicles’ return to the depot and updated finish times, imposed a 17h00 return with a 30-minute grace period, confirmed an assumed average travelling speed, added delivery durations and merged multiple orders to the same address into a single delivery duration. The tool also suggested applying customer delivery windows, which they declined.

After correcting the finish times to account for vehicles returning to the depot, Comodikes added a 17h00 return constraint with a 30-minute grace period. The tool adapted without restarting.

Read original

The priority is selection and clarity, not covering every topic. They floated a podcast and asked the audience which domains to unpack first.

The post asks readers to like it if they think the idea would benefit the industry, share existing supply chain podcasts and suggest which supply chain domains should be unpacked first.

Read original

In a related warehouse manager post, they describe business growth as adding complexity, not simply volume to move, and frame the role as an opportunity to take on interesting challenges and build supply chain experience. They also recommend a supply chain graduate finishing a BCom, available from the end of November and open to any role that builds experience.

The planner requirement calls for advanced Excel analytics and an enjoyment of advanced planning tools. In a warehouse post, Comodikes describes the pet industry as resilient through the pandemic and the company’s growth as a reason to embrace added complexity.

Read original · Original 2 · Original 3

Start here when deciding whether Christopher Comodikes merits a closer look.

Comodikes describes giving a real delivery dataset to AI instead of using the routing and scheduling software the team already owns.

• The goal was to compare a standard service level with a more demanding one across several days, a task the standard software made slow and awkward.
• Gemini replied that route clustering and optimization needed specialized software.
• ChatGPT clustered deliveries into routes, sequenced the stops, and produced finish times a planner could check.
• Comodikes then built the analysis step by step: adding a missing depot address and rerunning it; correcting the assumption that vehicles did not have to return to the depot, which changed the finish times; adding a "return by 17h00, with a 30-minute grace period" constraint; declining the tool's offer to apply customer delivery windows; answering its prompt about an assumed average travelling speed; and telling it to account for delivery times and merge multiple orders to the same address into one delivery duration.
• The stated limits are specific: no road-map view, only scatterplots of geolocated stops; no manual fine-tuning like that available in a graphical planning tool; and a risk that a new dataset and prompt send the model down a very different path that is hard to correct without restarting.
• The takeaways: AI became a credible backup if the supply chain software goes down, a quick tool for scenario testing without lengthy setup, and a system that prompts with suggestions rather than simply following instructions.
• Comodikes also notes that the base-case results aligned surprisingly well with the routing system, while acknowledging that hallucinations can lurk in results.

For brands planning a campaign

For a logistics technology or supply chain software company, this post serves as both a demonstration and a fit test.

• It shows someone who understands what routing software does well, is willing to test an alternative against it, and reports limitations rather than a clean win.
• A discovery call could explore how an AI assistant fits alongside deterministic routing software, which constraints need a stable interface, and how planners should verify AI output against an established solver.

Comodikes proposes a supply chain management podcast to dig through the jargon, separate myth from reality, and help C-suite leaders identify which supply chain domains make sense for their businesses to tackle. The post asks readers to like the idea if they think it would benefit the industry, share podcasts that already cover the subject, and suggest which domains would benefit most from being unpacked first.

For brands planning a campaign

This is positioning evidence, not a finished product.

• It shows a preference for explaining supply chain choices to senior decision-makers and a willingness to ask an audience what to cover.
• An engagement built around executive-facing explainers, myth-versus-practice segments, or short educational videos would match the interest Comodikes has declared.
• Ask which domains they would tackle first.

Comodikes posts that Pack Leader is looking for a warehouse manager and describes the pet industry as exciting and caring, with a focus on the well-being of people's pets.

• The post says the pandemic showed the industry is resilient and that growth at the business has created a need to handle additional complexity.
• It presents the role as offering interesting challenges and opportunities to grow in supply chain.
• Written from inside the business, it refers to "our growth."
• Ask Comodikes about their role at Pack Leader and when the post appeared.

For brands planning a campaign

Useful mainly for category context.

• The post shows operational familiarity with warehouse management, pet-industry fulfilment, and the idea that growth brings complexity, not simply more volume.
• A vendor selling warehouse or inventory technology could use it to open a discussion about what additional complexity looks like in a pet-products operation.

Comodikes writes that they are looking for a young supply chain planner, with around three years of experience, completed tertiary education, comfort with analytics in advanced Excel, and a genuine interest in advanced supply chain planning tools.

For brands planning a campaign

This is a small but direct signal about tooling literacy as a hiring and working standard.

• A planning-software company can reasonably expect Comodikes to assess whether a product supports real analytical work and whether planners enjoy using it, not just count features.
• The emphasis on enjoying the tools offers a practical opening for a discovery call about adoption and day-to-day usability.

It says less about how Comodikes approaches supply chain problems and more about their role in a professional network, which still matters to anyone assessing them for a partnership.

Comodikes shares that James Lycett will complete a BCom degree in Supply Chain Management at the end of the year.

• Lycett is looking to join a company and build experience, is eager to take on any role that will support their development, and is available from the end of November.
• Comodikes asks readers to consider Lycett if they are hiring.
• The post recommends a named third party but does not analyse supply chain practice.

For brands planning a campaign

This has little direct campaign use, but shows Comodikes vouching for early-career people in the field.

• For a partner running graduate programmes, apprenticeships, or a talent-facing campaign, that inclination offers a possible angle.
• Ask whether they have other examples of this kind of advocacy before building an engagement around it.

They use routing and scheduling software for daily operational planning but found multi-day scenario comparison cumbersome.

• Gemini replied that route clustering and optimisation needed specialised software; ChatGPT clustered deliveries into routes, sequenced stops and produced finish times to check feasibility.
• They then added the depot address, required vehicles to return to the depot, set a 17:00 return with a 30-minute grace period, specified an assumed average travelling speed and delivery durations, and consolidated multiple orders to one address under a single delivery duration.
• Their takeaways: AI is now a credible backup if supply chain software goes down, a quick analytical tool for testing scenarios without lengthy setup, and a source of suggestions rather than just an instruction-follower.
• In this test, its base cases aligned surprisingly well with their routing system.

Iterative constraint layering: add, correct and rerun one constraint at a time rather than writing a single, complete prompt.

They identify real limitations: no interface for showing routes on a road map, only scatterplots of geolocated stops; planning software allows easier manual fine-tuning; and a new dataset or prompt can lead the model to take a very different approach, making it hard to redirect without restarting. They also note that results may contain hallucinations.

For brands planning a campaign

Relevant for supply chain software, logistics technology or AI vendors.

• They offer a documented hands-on test of where general-purpose AI helps, where dedicated routing software still wins, and how to build resilience into planning.
• Natural topics include evaluating AI tools, route optimisation, scenario planning and software contingency.

Examples: Source 1

One recommends a graduate finishing a BCom in Supply Chain Management at the end of the year, available from the end of November, described as eager and open to any role that builds experience.

• Another seeks a supply chain planner with about three years' experience, completed tertiary education, comfort with advanced Excel analytics and an interest in advanced planning tools.
• The third recruits a warehouse manager for a pet-industry business, noting that the pandemic showed the industry's resilience and that company growth brought additional complexity.
• The role offers challenges that build supply chain experience.

These are recruitment posts. They describe role requirements, an industry view and a candidate recommendation, not a stated position on hiring policy or workforce strategy.

For brands planning a campaign

Shows a practitioner audience focused on operational hiring and logistics work. A brand could evaluate them for content on talent pipelines, the step from graduate to planner, practical hiring criteria for planning and warehouse roles, and how growth changes warehouse operations.

Examples: Source 1 · Source 2 · Source 3

They asked readers to like the post if they thought it would be good for the industry, share existing podcasts on the subject and suggest which domains to unpack first. Ask whether they plan to develop the series: the post invites audience input but does not present a delivered series.

For brands planning a campaign

Useful for advisory, executive-education or thought-leadership engagements.

• They show interest in translating operational detail for business leaders and in longer-form audio.
• A brand could scope topics such as framing supply chain priorities for executives, cutting jargon and choosing which domains deserve investment attention.

Examples: Source 1

They do not state whether they work for, own, or recruit on behalf of Pack Leader; ask them to clarify their relationship to the business.

View example

The post compares the tools but does not disclose a commercial relationship; ask whether the test was sponsored or otherwise commercially connected.

View example

Comodikes reports the result as part of a comparison between tools; the post does not disclose a commercial relationship.

View example

Like Comodikes, they pair concrete problems with checkable figures: around 5% of UK warehouse roofs carry solar panels, and the remaining roof space could support roughly 15GW of generation.

• They also spell out the limits of the obvious answer.
• Both write about their own service offer as well as the wider industry.

Pickford focuses on energy and physical infrastructure for fleets: rooftop solar, battery storage, smart charging, microgrids, demand-side response and vehicle-to-grid, as well as grid connection constraints.

• Comodikes focuses on software and analysis for planning decisions, plus operational hiring.
• Readers who follow them for AI and routing will find Pickford relevant on depot charging and site power, but not on model behaviour or tool evaluation.

Read original · Original 2

Their posts draw on long operating experience and make the same move as Comodikes: identify what existing systems already hold, then argue for a specific role for AI on top of it.

• White frames the problem around knowledge and answers stored in the TMS, WMS or rate engine—or in the head of someone who remembers the last time the same problem came up.
• Both set out the constraints as well as the benefits.

White focuses on institutional memory and knowledge capture across a career.

• A three-part series follows what they call the three layers of what experienced professionals carry and names a first step readers can take.
• Comodikes reports a hands-on test of a general-purpose AI tool against routing software and the constraints layered onto it.
• White writes as part of a vendor content series; Comodikes writes from an operating and recruitment perspective.

Read original · Original 2

Like Comodikes, Ponce connects that argument to what supply chain professionals should be learning and supports it with evidence.

Ponce focuses on education and events, including webinars, audience polls with response counts and a named academic programme, and addresses workforce and skills issues at scale.

• Comodikes writes short posts about one planning test and a handful of role requirements.
• Readers looking for structured teaching, audience data and a scheduled programme will get more from Ponce; those looking for a practitioner's account of trying a tool will get more from Comodikes.

Read original · Original 2

In two posts, Pickford explains how electric buses can be charged flexibly to absorb surplus renewable power, describes the First Bus trial with Optimo Energy, and lays out what a logistics microgrid needs: rooftop solar, batteries, smart charging, vehicle-to-grid and the right grid connection.

• They name the commercial constraint plainly: many sites cannot simply ask for more grid capacity because it is slow, costly or constrained.
• They also state the UK statistic that around 5% of warehouse roofs have solar, while the remaining roof space could support roughly 15GW.
• For a buyer seeking depot power, charging strategy or the investment case for electrification, that practical specificity is closer to the task than Comodikes's routing software test.
• They are the better choice when the subject is AI-assisted planning and a working comparison against established routing software.

Pickford's posts are tied to TransitionWise, the business they say helps logistics firms work out how trucks, chargers, solar, batteries and grid connections fit together.

• Both posts end with an invitation to get in touch.
• The content presents a specialist consultancy explaining its own service, while Comodikes compares modelled outputs against an existing routing system and reports where they misaligned.
• Pickford also offers no hands-on AI tool testing, making their posts a weaker match for a campaign about AI capability or model behaviour.
• The choice for supply chain AI content or advisory work is clear: Pickford for energy and infrastructure expertise with a consultancy interest; Comodikes for tested observations about planning AI, with no documented commercial offer.

Examples: Source 1 · Source 2

White states forty years in the business and opens his posts from that vantage point: people who could sense a bearing about to fail across the warehouse floor, and carrier reps who would pick up the phone at ten at night, none of which made it into the succession binder.

• He frames the problem as structured data, trusted analytics and unstructured judgment, a model he calls K³, and says he wrote a three-part RapidCanvas series on moving that judgment from someone's head into operational infrastructure.
• He also describes a recurring delay for logistics providers: the answer already exists in the TMS, WMS or a colleague's memory, but is filed around loads and lanes rather than the way a client asks, so it arrives a day late.
• That diagnosis of knowledge capture and retrieval complements Comodikes's focus on routing systems and makes White a strong starting point for campaigns about institutional memory, customer service responsiveness or bringing a new planner up to a senior planner's standard.

White's posts are built around a vendor relationship: both name RapidCanvas and its "Every Business Deserves Real AI Transformation" series, and both ask readers to follow links in the comments.

• That makes the posts useful for seeing how he writes about applied AI, but the practical claims appear in promotional pieces rather than a self-contained test readers can inspect.
• Comodikes reports the constraints they added, the corrections they made and the limits they hit, including no road-map view and no manual fine-tuning, and compares modelled base cases against their routing system.
• If the buyer wants a creator to put a product through an arms-length trial on camera, White's posts do not show that.
• For thought leadership with a named practitioner and vendor-linked series work, White's forty-year framing and K³ model offer a sharper narrative spine than Comodikes's single routing experiment.

Examples: Source 1 · Source 2

Both posts centre on the MITx MicroMasters Program in Supply Chain Management webinar "From Supply Chain Analytics to Agentic AI" with Vijay Sankararaman of Novant Health, covering the shift from predictive analytics and machine learning to generative and agentic AI, with real-world cases in forecasting, inventory management and procurement.

• The follow-up post reports live poll results from a global audience of almost 900 attendees, including 482 responses on which skills matter most and 538 on where supply chain professionals currently use AI, and closes on lifelong learning.
• For a software company addressing which skills planners will need, where AI is being used today, or how to reach supply chain professionals through an educational format, that combination of institutional framing, named guests and audience data is a better fit than Comodikes's account of a single internal routing comparison.
• They are the stronger choice when the buyer wants someone who has tested an AI assistant against their own planning software and will describe the friction.

Ponce's posts promote and then summarise a webinar rather than presenting original hands-on testing, so a campaign that depends on a creator independently exercising a product would not be served by the same material.

• The framing is institutional — the posts carry MIT Center for Transportation & Logistics, MIT Open Learning and related tags — which brings credibility but also less flexibility than a solo practitioner: a vendor wanting a critical comparison of its own tool against a rival would be asking Ponce to work outside how she is presented in the posts.
• There is also no supply chain AI consultancy offer or product-evaluation record in these two posts.
• The practical split for the same discovery call is whether the company wants educational content with academic weight and audience polling, or an evaluator who reports what a model did when constraints were layered onto a real delivery dataset.

Examples: Source 1 · Source 2