Pilots by invitation · The Digital Foundry is the build studio of Digital Edge · Pricing on request
the digital foundry · build studio of digital edge

Agents that work in production, shown live.

We build AI agents, automation and internal tools that run inside real workflows. Scoped fast, proven on your data, and shipped with governance in place. The Digital Foundry is the build engine of the Digital Edge group.

Pilots by invitationPricing on request
intake-agent · pattern
$ foundry run intake-agent --pattern → read new document from shared inbox → extract fields mapped to your schema → validate checked against your business rules ! hold low confidence on one field → route sent to a person for review ✓ approved reviewer confirms the value → post written to the system of record ✓ logged every step, with who and when $

Illustrative capability pattern. Not a client delivery.

how we work

Speed, creation, intelligent delivery.

We build in short, fixed-scope cycles, one workflow at a time. Every build ends in something that runs, and a decision based on evidence.

01 / speed

Short cycles, real decisions

Scope tight, build fast, demo often. A pilot ends with a scale, change or stop decision, agreed before it starts.

02 / creation

Working software, not slideware

Agents, automations and tools that run on your data and inside your systems, so you judge what works rather than what was promised.

03 / intelligent delivery

Governed from the first commit

Access control, logging, human oversight and POPIA designed in, so the security review is a check, not a rebuild.

why production matters

The gap is production, not pilots.

Agents are arriving in enterprise software fast. Most of the value is still being lost between a promising demo and a system people rely on. That gap is where we work.

40%

of enterprise applications will feature task-specific AI agents by the end of 2026, up from under 5% in 2025.

Source: Gartner, August 2025
40%+

of agentic AI projects will be cancelled by the end of 2027, on cost, unclear value or weak risk controls.

Source: Gartner, June 2025
5%

of companies achieve AI value at scale. 60% report little or no value at all.

Source: BCG, The Widening AI Value Gap, 2025
shown live

What our agents do.

Four capability patterns we build to. Each one does a defined job inside a workflow, and each keeps a person at the checkpoint that matters.

pattern/intake

Document intake agent

Reads incoming documents such as applications, claims or invoices, extracts the fields you need, validates them against your rules and posts them to your system of record.

human checkpoint
People review low-confidence extractions and every exception.
pattern/triage

Case triage agent

Classifies each incoming email or case, pulls the related record, sets a priority and routes it to the right queue with a short summary.

human checkpoint
Team leads see every routing decision and can reassign.
pattern/drafting

Proposal and reply drafting agent

Prepares replies or proposal sections from your approved policies, templates and the customer record, ready for review.

human checkpoint
Nothing is sent until a person approves it.
pattern/recon

Reconciliation agent

Matches records across sources, explains breaks in plain language and assembles the recurring report.

human checkpoint
Finance signs off breaks and the final pack.

Capability patterns describe what an agent of each type does. They are not client deliveries and make no claim about results. Build logs from live pilots publish here when pilots open.

packaged offers

Four ways to start building.

Defined packages so you can say yes to a clear scope, not an open-ended engagement. Pricing on request.

offer/01

Rapid MVP Sprint

Turn a defined idea into a working first version that real users can test.

What you get
offer/02

Automation Audit & Redesign

Map one process, find where the time goes, and redesign it before anything is automated.

What you get
offer/03

Internal Agent Build

A task-specific agent built into your systems, with the human checkpoint designed in.

What you get
offer/04

Workflow Modernisation

Move a manual or legacy workflow onto a modern, automated and monitored footing.

What you get
entry offer

Paid proof-of-value pilot

A short, fixed-scope pilot on one real workflow, with success criteria agreed in writing before we start. At the end you have a working agent or automation, evidence against the criteria, and a clear production decision: scale, change or stop.

  • One workflow, one owner, one measure that matters
  • Success criteria and a scale-or-stop decision agreed up front
  • Governance in scope from day one: access, logging, human oversight and POPIA
  • Production path and run cost estimated before the pilot ends
availability
By invitation
scope
Fixed, agreed in writing
fees
Pricing on request
first client

Our first client is our own house.

Before we ask a client to trust a build, we build for the group we belong to. The Digital Foundry builds the Own Your AI Studio products and the OwnID integrations.

client/own-your-ai-studio

Own Your AI Studio products

Own Your AI Studio is the group’s product house: AI products that the people who buy them own. The Foundry builds them.

ownyouraistudio.com
client/ownid

OwnID integrations

OwnID offers identity verification, business verification and AML screening for South African regulated firms, priced per check. The Foundry builds its integrations.

ownid.digital

Our own work runs on the same rules as a client pilot: one workflow, success criteria in writing, governance from the first commit. Build logs publish here when there is something to show, and we claim no results we cannot show. Client pilots are by invitation.

built this week

Shown live, as it ships.

When pilots open, each week’s build is published here: what it does, how it is governed, and what it does not do yet. No mock-ups, and no results we cannot show.

$ foundry log --this-week   [by invitation] No public entries yet. The first build log publishes when pilots open.   Each entry records: what the workflow and the agent built for it governed oversight, access and logging in place limits what it does not do yet
the digital edge group

Imagine. Build. Implement.

The Digital Foundry is part of one house, so an idea can travel from concept to working system to institution without changing hands.

imagines what could exist

Possible Art

Ideas, narratives and future-state thinking on how institutions change.

thepossibleart.com
builds what should exist

The Digital Foundry

AI agents, automation, internal tools and rapid builds that run in production.

How we build
implements what must work

Digital Edge

Vendor-neutral advisory on AI value, AI governance and the operating model around them.

askdigitaledge.com
tools

Estimate before you build.

tools/automation-savings

Automation Savings Estimator

Pick the workflows your teams run, enter your own volumes, time and cost, and see the hours automation could release and the agents worth building first. Your inputs stay in your browser.

Open the estimator
insights/

Why agent pilots stall before production

Most agent pilots do not fail on the model. They fail on scope, ownership, integration and a decision nobody agreed to make.

Read the note
next step

Have a workflow worth proving?

Tell us about one workflow. If it is a good candidate for a proof-of-value pilot, we will say so, and say why.