Cognitive Space reposted this
Global C-Suite Executive | COO | Delivering revenue growth and team excellence | Board Member | Executive Coach | Motorcycle Racer | Wild Swimmer | Mensan
Let’s imagine a customer “Analytics Co” needs high-resolution imagery of an AOI for regular monitoring 🛰️ Consider our confidence in Amazon: a purchase translates to 2-day delivery, with multiple status updates. Analytics Co’s customer journey for satellite data tasking is unlike e-commerce transactions - unless you’re a high-priority gov’t customer or very well-funded commercial client, purchase does not guarantee delivery 😬 It’s as if Amazon order response was: “We may or may not deliver.” The tasking order process for Analytics Co extends after order submission. The uncertainty is in the fulfillment phase, where providers execute complex supply chain logistics. Analytics Co sends a request specifying the site location, collection cadence, sensor morphology and resolution to multiple data providers. Some will deem the order infeasible due to various constraints. If deemed feasible, some data providers may accept tasking, but not guarantee data collection and delivery. Analytics Co receives multiple prices quotes and order feasibilities. Excel spreadsheet, anyone? 🙄 The industry lacks a standardized definition of "feasibility”. These assessments are typically static and fail to account for the real-time dynamics inherent within provider supply chains. Analytics Co doesn’t have ongoing visibility into the likelihood of order fulfillment, or a real way to compare across providers. Consider: ➡️ How much better could Analytics Co support customers if there was a comprehensive metric for order fulfillment success across multiple data providers, the “constellation of constellations”? ➡️ How much more effectively could Analytics Co use their data budget with improved predictability and reliability of tasking order feasibility assessments? ➡️ How much better could data providers support Analytics Co with more insight into market and customer demand? To have those, order location, timelines, price, competition, orbits, geometry, sensor type, weather, product type, satellite constraints must be considered, and multiplied by 10,000+ orders being managed 24/7/365. It’s a problem ripe for AI/ML. One we’ve been working for quite some time. Our AI-driven solution for Analytics Co is CNTIENTׄ·Earth ✅ With CNTIENTׄ·Earth, Analytics Co unlocks AI-powered intelligence to understand the capabilities of satellite data providers and plan orders to maximize the confidence of successful data acquisition. ✅ CNTIENTׄ·Earth removes the cognitive burden of complex, diverse interactions required by data providers for feasibility assessment, pricing, ordering, and order responsiveness. ✅ With CNTIENTׄ·Earth Analytics Co efficiently manages highly complex, dynamic order scenarios and demanding mission requirements like custody, monitoring and tip & cue. ✅ And since CNTIENTׄ·Earth is an analytics-ready platform, it feeds directly into Analytics Co’s workflows. Want to know more? 👉 Reach out to me via DM
Based on my conversations with all stakeholders involved in this workflow, this is an on-point summary of what’s needed! I’m so glad someone is working on this! One thing here is: in order to truly standardise feasibility predictions - the AI/ML model would have to be open source or a set of common metrics will have to be agreed upon by the industry like: 1. orbit based: avg. revisit rate, coverage, etc. 2. Weather based 3. sensor based etc. and a score is computed based on it. The heterogeneity of various AI/ML models will non-standardise the feasibility predictions again and Analytics Co. all be back at square one.
This is fantastic, but there's another issue there that I touched on in a post last week. I think this skips over the question of whether or not the data products from two different suppliers are actually interchangeable. They may be analysis ready, but is an image from Planet comparable to an image from Maxar when I'm feeding it into my model? Or do I need to make sure I have a customized version of the model that works for each sensor? Data harmonization is a big area of research right now, and I think we need this if we're really going to commoditize imagery.
I definitely agree with providing end customers with an AI/ML assistant that can help determine which data provider is best suited for their need/timeline/budget. However, sometimes the customer doesn't know what they really need in the product solution. i.e, defining the root issue that they are trying to solve with data products. So extending the AI/ML to help the customer assess what their need is before assessing data provider options, would help deliver the best ROI product solution. Just my opinion.
Interesting industry challenge and solution approach. I can relate this to supply-chain challenges in other industries like automotive where security, reliability, visibility, quality and pricing are paramount for tier-1 suppliers and buyers.
Refreshing and wonderful to read your approach on this Hanna. Our industry partners and customers should welcome this with open arms! love it.
This dovetail's nicely into what Will Cadell was talking about in his pushing pixel's blog a couple of weeks ago. Also, with hundred's of conversations with Bill Dollins Also, Aravind Ravichandran does this track with what you learned at EOSummit?
Insightful!
Global C-Suite Executive | COO | Delivering revenue growth and team excellence | Board Member | Executive Coach | Motorcycle Racer | Wild Swimmer | Mensan
2moLearn more at https://www.cognitivespace.com/cntient-earth/