Contours Forward Deployed Platform

Contours Forward Deployed Model

Services

At Contour, we have a services model that aims to serve implementation teams in onboarding customers in a fraction of the time that professional service organizations or in house implementation teams are able to do. I think services based companies are a new frontier of company that will replace entire verticals. We’re targetting the 3 trillion dollar implementation market, starting with in house midmarket companies and are investigating enterprise PSO’s.

Different New Age Forward Deployed Organizations

Palantir FDE Model

This is the model where organizations pay Palantir 1M+$ a year per FDE + a foundry usage fee + Compute fees and Palantir will deploy engineers on site to the a business and solve their most critical problems. The FDEs will go on site and build a custom platform for customers. For healthcare this might be solving nurse staffing issues. They would go to the hospitals and discover all the stakeholders that are relevant, then they would use foundry to build an app that connects all the stakeholders and relevant actions that the stakeholders can take to interact with each other. This includes organizing the data pipelines to actually work, creating the UI, pushing adoption on their customers and tracking actual ROI.

Decagon Agent PM model

This is where Decagon will deploy agent PMs + FDEs into their customer organizations and understand what the most common cases are that their customer support agents should be able to solve. Then they create agent PRDs that track every single use case for the agent to hillclimb against during the pilot. Decagon has two organizations essentially, the FDE/APM organization that deploys and builds custom integrations for the customer and the core engineering organization that builds the agent infrastructure and core product. For Decagon though, the forward deployed people are really only running discovery, and writing tools for the core engineering agent.

Differences

Now they sound kind of similar right? I mean Palantir has ontology(the core data platform) and Decagon has it’s core agent platform. And they both deploy forward deploy teams to customize the usage of their core platforms - so what’s the difference?

Palantir FDEs create many custom SAAS products across all verticals. They use ontology to streamline the creation of custom SAAS that solves critical bottlenecks - almost like advanced SAAS software consultation as a service.

Decagon on the other hand uses FDE’s to configure their own product to service the customers requirements in a singular vertical. It uses it’s agent platform to become super sophisticated in that singular vertical. Decagon is a true services model that has a manpower replacement argument.

The Contour Model

I believe new age services companies will be a combination of both. For implementation, every different company has different bottlenecks. They each have different workflows that have varying levels of difficulty, pain and serviceability.

So the main question is how do we build a core agent platform that can be repurposed to handle any implementation while having the customizeability per customer to show that we’re catering to their very specific workflows?

My current thoughts are that the contour forward deployed team will be a medium between the Palantir model and the Decagon model. They will need to write tools, procedures and dependency graphs for the customers to make the Contour core engineering agent better, but they will also need to create custom user facing dashboards that will surface to the customer the progress of all their most critical workflows and how they’re improving over time. Implementation isn’t something that customers trust to be fully autonomous on day 1, nor should they trust it because the contour agent will improve as it does more and more implementations. So the perogative for us should definitely be to use forward deployed engineers to create agent + human combined experiences where the agent can surface the confidence levels it has on different parts of the workflows and humans can approve and interact with the agent actions concurrently. For every customer, top line metrics look different, trust levels vary, and how custom the builds are also vary. So at Contour, we’re reimagining the human + agent experience where eventually the human may be removed from the loop but for now they work with the agent to drastically improve efficiency.

What Changes for us in the short term?

I think that there’s an arbitrage right now where nobody knows how to scale new age companies. Palantir patented a way to deploy really smart young teams into legacy companies and actually do real work. I think we’re going to be of the same philosphy, I might go into Berkeley tomorrow and look for 10 kids from ML@B and Launchpad as my FDE’s and poach the most capable ones onto the core platform team. Then I’ll deploy these teams into enterprises and because our core platform for implementation is so flexible they look for creative ways to build on top of the core platform to really do anything. All the while we’ll be improving the core platform, like if one enterprise is blocked because of sandboxing I’ll have the core engineering organization build it.