What we build for you

Eight ways we build with you.

Strategy to support: every service family, grouped by the outcome you need.

FileWhat we build

Outcomes
4
Service families
8

every claim links to its source

Four outcomes

Start from the outcome, not the technology.

Two service families deliver each outcome.

  1. Outcome A: Understand your knowledge

    Recurring research across documents nobody can search well.

  2. Outcome B: Make sense of your data

    Fragmented data and reporting rebuilt by hand every month.

  3. Outcome C: Build better workflows

    Disconnected tools, with the real process in spreadsheets and chat.

  4. Outcome D: Build and operate useful AI products

    Pilots that impress in a demo and stall before everyday use.

The eight service families

01Outcome D

AI strategy and product discovery

Find the one or two AI opportunities worth building, and prove it.

What we build

We interview the people who do the work, audit the data and rank opportunities by value, feasibility and risk.

Input output
  1. In Workflow interviews + data audit
  2. Then: prioritise by value and feasibility
  3. Out prototype brief

Example outputs

  1. A ranked use-case shortlist
  2. A prototype brief: user, inputs, expected output, evaluation questions

Proof from our products

The same method decided what we built in our own products.

02Outcome A

Data preparation and document intelligence

Turn scattered PDFs, spreadsheets and records into clean, searchable assets AI can work from.

What we build

We ingest documents and tables, standardise names, codes and units, and make the result searchable. It is the work behind Dataful’s 18,000+ public datasets[1].

Input output
  1. In PDF filings + spreadsheets
  2. Then: extract, standardise, index
  3. Out searchable, documented tables

Example outputs

  1. An ingestion pipeline to standardised, documented tables
  2. Fields extracted from documents, source page attached

Proof from our products

Dataful’s curation pipeline[1], the India Data Dictionary[2] and the BRSR extraction rules[3] are this work, in production.

03Outcome A

Knowledge and research assistants

Ask questions of your own documents and get answers that show their sources.

What we build

Assistants that retrieve only from your approved corpus, with a citation a reviewer can open on every answer.

Input output
  1. In Reports, circulars, archives
  2. Then: retrieve from an approved corpus
  3. Out cited answer, ready for review

Example outputs

  1. A cited research assistant over your reports, datasets and policy documents
  2. Policy search that compares versions and flags what changed

Proof from our products

Tagore AI answers from 2.5 million+ official documents[4] with citations; SACH searches a curated archive of hundreds of thousands of reviewed claims[5] from the browser.

04Outcome B

Conversational analytics, dashboards and reports

Ask a question of your data in plain language and get a chart, a table or a report you can check.

What we build

Built on GoPie: tables and the documents that explain them in one project, answers with the query and evidence attached[6], dashboards from a prompt and reports you approve before sharing[7].

Input output
  1. In Tables + the documents that explain them
  2. Then: query, chart, draft
  3. Out reviewed dashboard or report

Example outputs

  1. An ask-your-data workspace with evidence on every answer
  2. Recurring reports delivered to an inbox or chat channel for approval

Proof from our products

CSR Insights[8] and BRSR Insights[9] are GoPie running on real filings.

05Outcome B

Interactive data maps and knowledge portals

Organise a complex domain so people find what they did not know to search for.

What we build

We design the taxonomy, then build guided maps, collections and search. Dataful’s agriculture Data Map arranges 990 datasets as a journey in ten stages[10].

Input output
  1. In A large catalogue
  2. Then: taxonomy + information architecture
  3. Out guided map people can explore

Example outputs

  1. A guided sector catalogue or knowledge hub
  2. A data portal with preview-first pages and downloads

Proof from our products

Data Maps and Dataful are live, and both are built this way.

06Outcome C

Custom CRM, task and operations tools

Focused software around the way your team already works, from enquiries to tasks and records.

What we build

Small, well-made operational tools: a CRM that mirrors your pipeline, a delivery workspace that links proposals to projects and tasks.

Input output
  1. In Enquiries, proposals, tasks in five places
  2. Then: one focused workspace
  3. Out a team that works from one record

Example outputs

  1. A CRM shaped around your enquiry-to-proposal pipeline
  2. A delivery workspace linking proposals, projects, tasks and documents

Proof from our products

Factly’s own CRM and task tools, in daily use since we built them.

07Outcome C

Workflow automation and integrations

Connect your data, systems and approvals; automate the repeatable steps.

What we build

We connect the systems you have and automate the steps in between, with approval points where judgement is needed and a record of what ran.

Input output
  1. In Systems that do not talk
  2. Then: connect, schedule, approve
  3. Out steps that run themselves, with a record

Example outputs

  1. Reports and alerts to Slack, Microsoft Teams, Google Chat or email, with an approval step
  2. Integrations through APIs and the Model Context Protocol (MCP)

Proof from our products

GoPie ships scheduled reports[7], chat integrations[11], a REST API and an MCP server[12].

08Outcome D

Deployment, evaluation and ongoing support

Make it reliable, measure whether it works, and keep improving it after launch.

What we build

Test questions agreed before build, deployment to your hosting with monitoring and training, and measurement of task completion and correction effort.

Input output
  1. In A working prototype
  2. Then: evaluate, deploy, monitor
  3. Out a tool your team relies on

Example outputs

  1. An evaluation set and failure analysis against agreed thresholds
  2. A production deployment with monitoring, documentation and training

Sources (15)

  1. [1] Dataful: home · dataful.in (opens external site) accessed 2026-11-15
  2. [2] India Data Dictionary · dictionary.dataful.in (opens external site) accessed 2026-11-15
  3. [3] Dataful: BRSR dataset page · dataful.in (opens external site) accessed 2026-11-15
  4. [4] Tagore AI corpus figures, provided by Factly (November 2026) · tagore.ai (opens external site) accessed 2026-11-16
  5. [5] SACH browser extension · factlymedia.com (opens external site) accessed 2026-11-15
  6. [6] GoPie documentation: how the agent works · gopie.ai (opens external site) accessed 2026-11-15
  7. [7] GoPie documentation: reports · gopie.ai (opens external site) accessed 2026-11-15
  8. [8] CSR Insights: home · csr.dataful.in (opens external site) accessed 2026-11-15
  9. [9] BRSR Insights: home · brsr.dataful.in (opens external site) accessed 2026-11-15
  10. [10] Agriculture Data Map catalogue (bundle of 2026-10-16) · datamaps.dataful.in (opens external site) accessed 2026-11-15
  11. [11] GoPie documentation: integrations · gopie.ai (opens external site) accessed 2026-11-15
  12. [12] GoPie documentation: MCP server · gopie.ai (opens external site) accessed 2026-11-15
  13. [13] Tagore AI documentation: limitations · docs.tagore.ai (opens external site) accessed 2026-11-15
  14. [14] GoPie documentation: security overview · gopie.ai (opens external site) accessed 2026-11-15
  15. [15] GoPie documentation: AI data privacy · gopie.ai (opens external site) accessed 2026-11-15

Discuss a project

Tell us what you want to build.

Describe the information problem. We will say what we would build first.