thinking sense · Early Conversations
What enterprise teams are telling us.
These are candid quotes from discovery conversations with data and technology leaders across retail, financial services, and consulting. Companies are anonymous. The problems are real.
Fortune 100 Specialty Retailer
Head of Data & Analytics
Brand-led growth · Store operations · Snowflake
“Should I stop doing what I’m doing? Every project requires me to write an ETL pipeline and use a medallion architecture, which I struggle with even to get started. The rest of it is flowing some of our data into Snowflake and get partial answers there. What you’re showing is indeed radical, I need to think about how to approach my CIO. We have so many unanswered business questions as our business moves from bottomline store efficiency to topline brand-led growth.”
What they’re solving
- ETL pipelines required for every project
- Medallion architecture overhead blocks getting started
- Snowflake only surfaces partial answers
- CIO needs a compelling case to shift investment
Insight: The shift from cost-center to growth-driver means the business is asking new questions — questions the current data stack was never designed to answer.
Large Omnichannel Retailer
VP Data Platform Engineering
Data governance · BigQuery · Production system load
“Nearly 2% of data makes it into Google BigQuery, from which we get our analytics. Most business users build their chatbots that increasingly need access to our production data sources — databases, applications — to access the system of record and source of truth. This puts strain on these systems, so security and QoS are big concerns. To begin with we don't want to move data into data warehouses if we can avoid the overhead of ETL pipeline maintenance and governance overhead. What you're building is exactly what solves our problem.”
What they’re solving
- 98% of data never reaches analytics — invisible to the business
- AI chatbots hitting production systems directly, straining QoS
- Security risk from uncontrolled access to source-of-truth systems
- ETL maintenance and governance overhead is unsustainable
Insight: When only 2% of data is analytically reachable, you're not running on data — you're running on a sample. And the fix (move everything to a warehouse) creates the problems thinking sense eliminates.
Global Financial Institution
Head of Risk Technology
Risk management · NL2SQL · Knowledge graph
“We have tons of business users who want to replicate what we do here in risk management — answering questions of business risk when an application server goes down, or an external cyberattack is reported. Today we build layers manually that extract the knowledge from various systems, writing SQL queries using NL2SQL across individual systems and then stitching the knowledge graph. What you're doing in auto-discovery of ontology, building knowledge graph, and using vector search seamlessly to answer questions from a variety of data sources, without ETL, is exactly what we'd like to build our business intelligence platform on.”
What they’re solving
- Manual SQL-layer construction across every system
- Knowledge graph stitched by hand — brittle, slow to update
- Risk questions span systems but queries are siloed
- Can't scale what the risk team does to the broader business
Insight: Risk teams already know what cross-system intelligence looks like — they built it by hand. thinking sense makes that capability available to every business unit without the manual rebuild.
Global Supply Chain Consultancy
VP Supply Chain Technology
Supply chain · SAP · Document-to-ERP
“We have merchants adding new products constantly into our supply chain, and these come in a variety of PDF styles. Joining the data from these documents into ERP systems has required a lot of manual work, and we want AI to solve this. SAP says they can solve this in 2027, and I have customers waiting for this. With what you do, we solve this today.”
What they’re solving
- New product onboarding requires manual document-to-ERP data entry
- PDF formats vary by merchant — no standard to automate against
- SAP's native solution is 12–18 months away
- Clients are waiting now and competitors will move first
Insight: When your ERP vendor's roadmap is the blocker, the enterprise is looking for a layer that can close the gap today — without waiting for a platform upgrade.
What we hear in every conversation.
The same four companies. Two people in the room. Two different sentences — one conclusion.
The CIO says
Control · Time · Exposure
- “I have business questions nobody can answer, and my data team says eighteen months.”
- “Two percent of my data reaches anyone who can ask it a question.”
- “My ERP vendor says 2027. My customers are waiting now.”
- “Every platform I buy wants to be the place my intelligence lives — and the place it can never leave.”
The CIO’s question isn’t which AI. It’s who ends up in control.
The architect says
Pipelines · Stitching · Production load
- “Every project starts with an ETL pipeline and a medallion architecture I have to build before I can ask anything.”
- “Business users’ chatbots are hitting production databases directly. Security and QoS both break.”
- “We hand-stitch NL2SQL across systems into a knowledge graph. It works. It doesn’t scale.”
- “New product PDFs in twenty formats, joined into the ERP by hand.”
The architect’s question isn’t can we build it. It’s why are we building it three times.
Both arrive at the same place.
The intelligence layer has to sit above every system, belong to the enterprise, and start answering in days — not quarters. That is the layer thinking sense is.
Patterns across conversations
Four problems. Every conversation.
⟁
ETL is the bottleneck
Across every conversation, the first blocker is ETL — not strategy, not budget, not will. The overhead of moving, maintaining, and governing data pipelines delays every project before it starts.
◎
Production systems are under strain
As AI agents proliferate, they hit production databases directly for source-of-truth access. Security teams flag it. Ops teams throttle it. No one has solved the access governance problem at scale.
⌬
Cross-system intelligence is built by hand
The teams that do this well — risk management, supply chain ops — built it themselves with SQL, stitching, and years of institutional knowledge. It doesn't scale. It doesn't transfer.
◇
The business can't wait for the vendor roadmap
"SAP says 2027." "Snowflake is working on it." "We're migrating to a new warehouse." These timelines don't match business urgency. Enterprises need answers from the systems they have today.
Recognize your problem here?
We’re working with a small number of enterprise teams on early deployments. If the pattern fits, let’s talk.