AI systems thatshow their sources.
Document intelligence, workflow automation, machine learning, data platforms and financial models for healthcare, finance and energy teams. Built on your data and run for you month to month, or built inside your team by an embedded engineer.
We build the document assistant, the workflow, the model, the data pipeline or the financial model on your data, deploy it, and run it for one monthly price, or we put an engineer inside your team to build it with you. Every answer cites its source, and a check that fails stops the run.
Systems we build. Ways we work with you.
Five systems, shown running below with their data and their limits; an engineer embedded in your team; and the briefings that explain the results.
- Document IntelligenceQuestions over records, filings and contracts, every answer cited to the passage.
- Workflow AutomationIntake, drafting and routing with a human review queue where errors are expensive.
- Machine Learning and Decision ModelsForecasting, risk, fraud and ranking models trained on your data, delivered as applications your team uses.
- Data PlatformPipelines and warehouses whose correctness is tested on every run.
- Financial Analysis and ModelingValuation, forecasting, close and reconciliation models for finance and accounting teams, every figure traceable.
- Forward-Deployed AI EngineeringAn engineer embedded with your team, integrating AI into the systems you already run.From $9,800 a month
- AI Communication and EnablementBriefings, results reports, documentation and training that explain what a system does and where it stops.From $2,900
- Fractional Head of AIMethod review, vendor and model selection and hiring help, a few hours a month, for teams that want the judgment before the system.$2,500 a month
- 1Tell us the workflowDescribe the process, the data and what a good result looks like. The reply comes from the engineer who would do the work.Reply within one business day
- 2Discovery SprintA workflow and data audit, a prototype on a sample of your data and a written plan with scope, risks and a quote. You keep both.Two weeks, $4,900 fixed
- 3BuildThe deployed system: assistant, model or pipeline, interface, integration, tests and documentation, accepted against the plan.Typically $12,000 to $40,000
- 4OperateHosted, monitored and maintained, with a monthly usage and quality report. Month to month, cancel any time.From $1,490 a month
A defined result at a published price.
Nine fixed-scope engagements, from $1,900: a written scope, a list of what you receive and a date. Five of them have a sample deliverable you can read now.
- Reconciliation Automation PilotOne reconciliation, automated end to end on your own exports: bank to ledger, subledger to general ledger, or intercompany. A sample deliverable is published.$9,500 fixed, three weeks
- Financial Model ReviewYour model read the way a lender, buyer or auditor will read it. A sample deliverable is published.$1,900 fixed, five business days
- Financial Model BuildAn integrated three-statement or deal model with the assumptions on one sheet, three scenarios and sensitivity tables, built so that someone else can maintain it.From $4,900, two to four weeks
- Pre-LOI Market and Target ReportFor buyers deciding whether a target deserves a letter of intent: market size and growth, competitors, customer and pricing signals, and the risks to the thesis, with every figure cited.$7,500 fixed, ten business days
- Diligence Data AnalysisThe transaction-level data of the target analyzed for the deal team or the accounting firm running diligence: revenue quality, customer cohorts and concentration, margin by product, working capital patterns. A sample deliverable is published.From $7,500, two to three weeks
- Reporting Trust ReviewFor teams whose numbers are disputed in meetings. A sample deliverable is published.$4,900 fixed, two weeks
- Operations and Cost DiagnosticOne function examined from its accounting data and its actual workflow: what it costs, where the time goes, and which steps are worth automating, with the savings estimated under stated assumptions.$7,500 fixed, two to three weeks
- AI System EvaluationFor a team with a generative AI tool in pilot or production and no measured answer to how often it is wrong. A sample deliverable is published.From $7,500, two to four weeks
- AI Team Rollout (Claude or ChatGPT)A working setup for a team of 5 to 50: permissions checked, up to five tools connected, one workflow built for the work the team does most, and a training session on the documents the team works with.From $4,900, about two weeks
Where a confident wrong answer is expensive.
Healthcare, finance and energy share regulated, document-heavy work and data that is already there.
Running systems.
Open them. Each one states what it proves, the data it uses and where it stops.
Chart ExplorerQuestions over a patient chart.
ComputePowerRanks all 21 PJM zones by grid vulnerability as data center load grows.
BlackletterAnswers questions over statutes and contracts with the source passage for every claim.
TickerScopeCompares stocks and ETFs on risk-adjusted terms in one view.
Marketplace Data PlatformA migration verified by source-to-target parity tests.Your workflow, running here.A two-week Discovery Sprint ends with a prototype on your data.Start a conversation Priced on the page.Small first step.
No hourly rates and nothing you cannot leave. The Discovery Sprint is fixed-price, the Build is quoted from it, and the plans run month to month.
- $4,900Discovery Sprint, fixed, credited against the Build
- 2 weeksto a working prototype on your data
- $1,490a month to run one system, all in
- $3,490a month for up to three systems
- Everyanswer cited to its source, or declined
- Nocontract: month to month, and the code is yours
Who does the work?
DF1 is a small firm. A senior engineer leads every engagement and does the core modeling and engineering, with the same method on every project. If a project needs a specialist, the plan names the role before the work starts.
How long until something works?
The Discovery Sprint takes two weeks and ends with a working prototype on a sample of your data. A Build is scoped and dated in the plan that the Sprint produces.
What happens to our data?
It stays in the storage you choose, is used only for your system, and is never used to train a model. DF1 uses model providers under terms that exclude training on customer data.
Can you work with PHI?
Production work on PHI runs under an Enterprise agreement with a business associate agreement chain covering DF1, the model provider and the host. Demos use synthetic records, and discovery uses de-identified samples or runs in your environment.
Is there a contract?
Plans are month to month and can be cancelled at any time. A Build is a fixed-scope project, half on start and half on acceptance.
Who owns the system?
You do. Code, data pipelines, prompts and documentation are yours at the end of the engagement or the plan.
Tell us the workflow. We show you the system.
Describe the process and the data. A reply comes within one business day, from the engineer who would do the work.



