Implementation and growth services
We help you turn data from all your systems into AI your team can run
Our engineers build AI-enabled reporting and analytics with your team on the pre-built DataForge platform, starting with the First Mile, a pilot on one slice of what you run today. After the production rollout, your team can add more systems and reports on its own or with us.
At a glance
- First step
- The First Mile, a scoped pilot
- Who builds it
- Our engineers, with your team
- Who runs it after the rollout
- Your team, plus up to 4 hours a month with our technical specialists
- Where your data stays
- Your Databricks or Snowflake account
Getting started
What you need to work with us
We scope each piece of work before it starts, from the First Mile pilot to the production rollout and each project after it.
- Platform license
- Our services require one
- You can also license the platform on its own. Platform pricing is on the pricing page.
- Our services
- A separate agreement from the license
- A fixed scope is our strong preference. We sell outcomes, not people.
- Your cloud
- Your Databricks or Snowflake bill stays your own
- Your data is stored and processed in your own account, and the pipelines run on your compute.
- Your team's time
- At least an hour a day from an analyst during the First Mile and the rollout
- Someone in FP&A or data analytics who knows your current data. How much of their time we need depends on the scope of the project and the size of our team.
Working with us
We strive to be better, faster and cheaper than the competing options at every stage
You license one pre-built platform, and our engineers can work with your team at any stage.
Our implementation services are optional. We are just as happy when your team builds and grows your reporting on its own.
-
Implementation
We build a First Mile pilot with you. When you decide to go ahead, we build the production rollout and test it against your source systems and current reports.
-
Training
We train your team by building with them on your data.
-
Maintenance
DataForge maintains the architecture, and your team manages the business logic. Your license includes up to 4 hours a month with our technical specialists for help using the platform.
-
Growth
We help you add systems and take on new projects whenever you need them, right after implementation or five years later.
60 percent or more saved on the cloud services bill in our case studies
We saved customers that much by improving their architecture and code quality with our product and services.
See one customer's numbersCustomers we have delivered for
Projects
What we can build with you
After the First Mile and the production rollout, we take on any development work on DataForge that we can scope as a fixed outcome: a migration, a new connection, an integration, a new business domain.
Legacy code modernization
We rebuild the business logic in your legacy code as DataForge declarations that are easy to read and adjust, and check the results against what you run today.
Multi-system consolidation
We bring 10+ ERP, EMR and other operational, marketing, sales and HR systems together, with deep financial drill-down across all of them.
Master data management
We set up a repeatable process for customer reconciliation, product mastering and the other core records your reporting depends on, and hand it to your team to run.
Example engagement
An acquisitive global manufacturer
Based on a real engagement, kept anonymous.
-
Their systems
- 4 ERPs
- CRM
- CMS
- Excel files
- Fixed-width files from a third-party FTP site
-
- The pre-built data model
- Master data that consolidates customers, products and other records across their systems
Refreshed near real time, daily or weekly, depending on the data
-
What they got
- Finished reports in Power BI
- An AI semantic layer for agentic tools such as Claude and Codex
- Margin analytics in more detail than their financial consolidation tool could show
- Daily margin detail and forecasting across customers
We support any BI tool, including Tableau and Looker.
How long it took
-
Week 0
Start
-
3 to 4 weeks
First report
From the source data alone, before any new business rules
-
About 3 months
Final reports in production
With detailed analytics, the AI semantic layer and a few thousand new business rules
That was this customer's timeline. We set yours when we scope the work.
Their team's time
At least an hour a day
from an FP&A analyst who knew the current data and could find answers across the business
How we checked the numbers
We tested the outputs against the source systems and existing reports, with AI tools and our standard tests. We boiled the differences down to key business questions. Many were not errors. The old reports used high-level definitions, and the new data is more detailed.
After the build
We trained their team as we built. Support continues under their license, and their team can keep building on its own or with us.
The First Mile
We start with one slice of your business logic
The First Mile is a scoped pilot that modernizes one slice of what you run today, in five phases with one deliverable each.
0 Scope , Week 0
1 Install and Ingest , Week 1
2 Transform , Week 2
3 Publish , Week 3
4 Present , Weeks 4 to 6
- 0 Scope Week 0
- 1 Install and Ingest Week 1
- 2 Transform Week 2
- 3 Publish Week 3
- 4 Present Weeks 4 to 6
We estimate four to six weeks, and we set the actual timeline when we confirm scope in Phase 0. The outlined bar marks the Phase 4 window, weeks 4 to 6.
What each phase delivers, and what you provideAfter the First Mile
What comes next
The pilot ends with a demonstration and our recommendation. What happens next is your call, and some teams carry the work forward themselves.
When you want us to keep going, we scope and build the production rollout with you, and then the systems and reports you add as you grow.
Proof
What a First Mile has produced
The First Mile page describes one pilot in detail, and the work that came after it. Every figure is traced to project documents.
Read what the pilot producedCost
We reduce the custom work behind implementation and growth
One customer's numbers
- Before
- $500k+
- a year for their old stack, built natively on Databricks or Snowflake
- The project
- code and architecture work with our product and services
- After
- ~$200k
- a year for their new stack, including the DataForge license
We price based on total cost of ownership across both capital and operating budgets.
A custom implementation
A contract and a bill from each
- Ingestion tool
- Transformation tool
- Orchestration tool
- Observability and lineage tools
- Outside firm for the build, and again for each new project
- Databricks or Snowflake
On DataForge
You can buy the platform and the build from one company, with a separate agreement and bill for each
One view of your pipeline spend
DataForge
- The platform license, with ingestion, transformation, orchestration, observability and lineage built in
- Up to 4 hours a month with our technical specialists, included in the license
- Implementation and growth services when you want them
DataForge's cost report puts the platform fee and the compute your pipelines use in one view, from the monthly total down to each source. You get it in the DataForge app, as a Databricks dashboard, or built to your spec.
The alternative
A pre-built platform instead of tools assembled by hand
On the usual path, you buy a tool for each layer of the stack and hire an outside firm to connect them and build the warehouse. Then your team runs the result. DataForge licenses the platform as one product, and our engineers can help you build your business logic on it.
-
Assemble a tool for each layer
The usual path
- Tool vendors
- Your team
On DataForge
- The platform
On DataForge: DataForge already has ingestion, transformation, orchestration, observability and lineage built in.
-
Connect the parts
The usual path
- Outside firm
On DataForge
- The platform
On DataForge: DataForge is sold only as one complete platform, with zero integration needed between the parts.
-
Design the architecture
The usual path
- Outside firm
- Your team
On DataForge
- The platform
On DataForge: DataForge makes the decisions about layers, keys, processing order and failure handling for you, or removes the underlying problem.
-
Write the design documents
The usual path
- Outside firm
- Your team
On DataForge
- The platform
On DataForge: The standard DataForge documentation fills this role, as every DataForge project is built on the same design.
-
Build the pipelines
The usual path
- Outside firm
On DataForge
- With us
On DataForge: We modernize your existing pipeline logic as DataForge declarations and build the target tables, starting with Phases 2 and 3 of the First Mile.
-
Deploy to production and train resources
The usual path
- Outside firm
On DataForge
- With us
On DataForge: We build the datasets and reports together with your team, so they learn the platform hands-on, and then hand them the running platform.
-
Maintain the design
The usual path
- Your team
On DataForge
- The platform
- Your team
On DataForge: DataForge maintains the architecture and technical design. Your team manages the business logic definitions as declarations that are easy to read and adjust over time.
-
Grow, expand, enhance
The usual path
- Outside firm
- Your team
On DataForge
- Your team
- With us
On DataForge: Your team can take on new projects on its own or with us. Your license includes up to 4 hours a month with our technical specialists to fill in any gaps your team has in expertise on the platform and its evolution.
Architecture
Decisions and expertise built into the platform
Building a warehouse means making a familiar set of architectural decisions. On the usual path, each one is designed for your project and left for whoever maintains it afterwards.
For most of them, DataForge either makes the decision for you or removes the underlying problem completely. This applies to the build or migration, and to long-term development and maintenance as well. Your team keeps the ones that decide whether the numbers are right.
The architectural decisions a warehouse build requires, on the usual path and on DataForge
On DataForge
Crossed out means built into the platform
- 1
Layers, and where each calculation goes, built into the platform - 2
Keys and join paths between tables, built into the platform - 3
Processing order, data loads and compute, built into the platform - 4
Failure handling and reprocessing, built into the platform - 5
Run history and lineage, built into the platform - 6 What each number means, decided by your team
- 7 Which source table and key each report uses, decided by your team
- 8 The fiscal calendar, exclusions and anomalies, decided by your team
How DataForge eliminates decisions
- 1 Layers, and where each calculation goes. The platform keeps data in five layers, each with one job, and a rule computes one column on the source it belongs to.
- 2 Keys and join paths between tables. Relations join sources on columns they already carry, such as a shared customer ID. When a rule reads a related source, the platform picks the path to it, and you can change that path.
- 3 Processing order, data loads and compute. Steps run in a fixed order, and declared dependencies make a source wait for the sources it reads from. Each source has settings for how new and changed records load and which compute runs them.
- 4 Failure handling and reprocessing. Each failure is logged against the pull of data it happened on, and once the cause is fixed, the failed step reruns and every later step follows. You can also rerun a chosen step, or reprocess the rules you added or changed.
- 5 Run history and lineage. Every pull of data and every process run on it is logged with its status, timing and record counts. The lineage graph follows any column downstream to the rules and outputs that use it.
Your business logic
We help your team work out what the numbers should mean
Our engineers help your team find the business logic in what you run today, such as what a status code in the source means, and build it on the platform. After the build, your team owns that business logic and can read and adjust it over time.
We build with coding agents like Claude Code and Codex through the DataForge MCP server, and with other tools as needed. Every change an agent makes goes through the same checks as a change made by hand.
After the engagement
Your team takes it from here
We build the datasets and reports together with your people, so they learn the platform hands-on. Support for the running platform continues under your license, through the support page.
Your Databricks or Snowflake account
-
Ingested data
Stored in your account
-
Pipelines
Run on your compute, orchestrated by DataForge
-
Target tables
Normal Databricks or Snowflake tables and views
Stays with you
Our engineers
Work inside your cloud account, on your data, for the initial build
Your team
Runs the platform after the handover
Your data
Your data stays in your cloud
Your data is stored and processed in your own Databricks or Snowflake account, on AWS, Azure or Google Cloud.
DataForge orchestrates the pipelines that run there, on your compute.
We host your metadata and the DataForge services for you. On the Enterprise plan, you can host both in your own cloud account instead.
Yours to query, read and export
Normal tables
The tables DataForge builds are normal Databricks or Snowflake tables and views in your own account, queryable with the SQL your team already writes.
Vanilla SQL
The code DataForge manages is vanilla SQL, stored in a database your team can easily query. Lineage traces every column back to its source. Every data pull and process run is logged with its record counts and status, so you can audit how any number was built.
Yours by contract
Under your agreement, your data, configuration and results stay yours, and your team can retrieve them through the platform at any time during the term.
A path to native Databricks, Snowflake or dbt
Moving to native Databricks, native Snowflake or dbt is a straightforward path. You keep the tables, the vanilla SQL and your configuration. Your team rebuilds the orchestration and the business logic in native code, starting from that SQL, and we can help. Going from a well-formed architecture and a declarative framework to hand-written, step-by-step code is much easier than the reverse.
Direct answers
Do you do the implementation yourselves?
Yes, if you want us to. Our forward-deployed engineers do the initial build inside your cloud account, on your data, and then hand the running platform to your team. Until then, we are accountable for running the build.
Can we keep our systems integrator?
Yes. We work alongside your partner and review their designs, code and configuration. You keep the decisions and the partner relationship. If you would rather work with one company, you can buy the platform and the build from us.
We already run Databricks or Snowflake. What does DataForge add?
A complete data platform that works with the account you already have. Ingestion, transformation, orchestration, observability and lineage are built in, with zero integration needed between the parts. DataForge maintains the architecture and technical design, and your team manages the business logic.
Where does our data stay?
In your own Databricks or Snowflake account, on AWS, Azure or Google Cloud. Your data is stored and processed there, and DataForge orchestrates the pipelines that run on your compute. We host your metadata and the DataForge services for you. On the Enterprise plan, you can host both in your own cloud account instead.
How do we get started?
You can license the platform and start building on your own, or start with us on the First Mile, a pilot that modernizes one slice of what you run today. We estimate the First Mile at four to six weeks, and we set the actual timeline when we confirm scope in Phase 0. It also depends on how quickly we get access to your systems and answers from your team. At the end we check the DataForge output against your current output, side by side, with spot checks and aggregate comparisons.
What happens after the First Mile?
That is your call. The First Mile ends with a demonstration and our recommendation, and it proves the approach on your data. Some teams carry on from there themselves, and we are just as happy when they do. When you want us to keep going, we build the production rollout with you and test it against your source systems and current reports. You approve the rollout before it goes live. We also help you add systems and reports as you grow, right after the rollout or five years later. We write each one into a statement of work before it starts.
Who runs the platform after the handover?
Your team. We build the datasets and reports together with your people, so they learn the platform hands-on. The platform handles the tooling and the architecture, so a data-literate person working as part of a team can do the building, and your people decide what the numbers should mean. Your team can also build with coding agents such as Claude Code and Codex through the DataForge MCP server, and every change an agent makes is validated the same way. Support continues under your license, including up to 4 hours a month with our technical specialists. Most customers have one or two of their own people who keep adding systems and reports, and we can provide a full outsourced team if you want one.
What does it cost?
Platform pricing is on the pricing page. You can license the platform on its own, and our services require a license. The license and the services are separate agreements with separate bills, and your Databricks or Snowflake bill stays your own. For services, a fixed scope is our strong preference. We sell outcomes, not people.
Can we talk to a customer?
Yes. Ask us on a call and we will set it up, with the customer's agreement.
Tell us what you run today and what you need it to answer
We will tell you whether DataForge is a fit, the likely scope, and what we need from you to start.
Platform pricing is on the pricing page.