AI automation

Automation for legacy software without APIs

Steadyhand · Screen-reading agents that finish legacy portal work, with approvals and audit trails

WatchingDraft scoreMature market · since 2003 · AI estimate

33PWAS score

Confidence

Based on 8 of 9 factors

Overview

AIThis description was written by a model from public signals. The facts behind it are listed under Evidence.

Problem: Teams need to automate workflows in legacy desktop and web applications that have no API, incomplete APIs, or integrations that take months to build. Traditional RPA breaks when interfaces change or unexpected states occur, and automating these workflows reliably enough for production use—especially when they involve sensitive data—is difficult.
Solution concept: A computer-use agent platform that can observe screen state, decide and execute actions, verify results, and recover from unexpected conditions—without needing DOM access, selectors, or application-specific integrations. The platform would enforce invariants and approval gates to prevent silent errors in sensitive workflows, and provide full audit trails and replay capability for compliance.
Who has it: Operations teams in healthcare, accounting, and other regulated industries who manage workflows across legacy applications, payer portals, EHRs, and remote desktops.
How they handle it today: Today, teams either manually complete workflows through legacy applications, use record-and-replay RPA systems that are brittle, or spend months building integrations. Many workflows still involve manual data entry between systems, spreadsheets, and remote desktops.
Why now: Vision-based AI models have reached the capability needed for production computer-use agents, and the gap between demo reliability and production-grade reliability (handling unexpected states, verifying data, preventing silent errors, providing audit trails) is now being addressed by dedicated infrastructure.
How it could charge: Per-agent runtime and workflow volume, with separate pricing for dedicated environments and enterprise deployments.

Score breakdown

A model sorted each factor into a rubric bucket; code turned buckets into numbers. Unknown factors are left out of the score and lower confidence instead.

Market validation

Confidence 95 · 4 evidence · Multiple signals show clear demand: a Hacker News post with strong engagement (100 points, 84 comments), evidence of existing use cases (property manager using Kampala), and founder experience automating legacy system workflows in healthcare operations.

Revenue evidence

Unknown

No evidence provided about whether customers pay, pricing acceptance, or any revenue generated by existing solutions in this space.

Growth

Confidence 95 · 2 evidence · Engagement on a single Hacker News post (44 and 100 points across posts) shows interest but is insufficient to establish momentum; no data on growth trajectory of computer-use automation or related categories.

Problem severity

Confidence 70 · 4 evidence · The evidence describes critical pain: production workflows in regulated industries with high cost of failure, organizations spending multiple hours daily on manual reconciliation, and reliability requirements that existing tools cannot meet.

Economic proximity

Confidence 70 · 1 evidence · Workflows in healthcare operations and accounting involve direct cost savings (eliminating hours of manual labor daily) and reducing failure risk in revenue-critical processes, placing the product close to money saved.

Distribution

Confidence 30 · Judged from the description · The product is an internal automation tool that operations teams run on their own systems; however, it likely produces audit trails, reports, and outputs reviewed by other stakeholders including compliance and management.

Buildability

Confidence 30 · Judged from the description · Building a production-grade computer-use agent requires handling complex challenges: session management, anti-bot detection evasion, state tracking, error recovery, and compliance-grade audit logging—more complex than standard integrations but not requiring breakthrough research.

Differentiation

Confidence 70 · 4 evidence · A clear gap exists that users and incumbents recognize: existing RPA breaks with interface changes, browser automation is detected by anti-bots, and computer vision approaches are brittle; the solution addresses this with deterministic session-token-based automation and reliability infrastructure.

Retention and timing

Confidence 70 · 2 evidence · Workflows are recurring (daily or ongoing data reconciliation across legacy systems), and the timing is favorable given recent advances in vision-based AI agents and growing pressure to automate legacy-system workflows in regulated industries.

Risk penalties

Heavy regulationThe product targets healthcare, accounting, and regulated industries managing sensitive workflows and compliance-critical processes; automation of such workflows faces audit, liability, and regulatory compliance demands.−15
Questionable data accessAutomating workflows in legacy applications, payer portals, and remote desktops requires the platform to intercept, observe, and replicate data flows that may be governed by patient privacy laws, financial regulations, or terms of service.−15
Extreme competitionThe landscape includes established RPA vendors (UiPath, Blue Prism, Automation Anywhere), browser automation tools, and custom API integration services; however, none solve the production-reliability problem identified, creating a differentiated niche.−6

Score history

24 · confidence 56 · config v1First score · 10 Oct 2026
33 · confidence 58 · config v1Evidence changed · 10 Oct 2026

Evidence · 13 items

Market validation

Hacker News post received 44 points and 26 comments
Verified
Hacker NewsRetrieved 10 Oct 2026
Hacker News post received 100 points and 84 comments
Verified
Hacker NewsRetrieved 10 Oct 2026
A property manager automated 2-3 hours of billing information entry using Kampala
Public claim
I even sent it to my property manager mom, who (with a lot of help from me lol), automated 2-3 hours of billing information entry in Yardi.
Hacker NewsRetrieved 10 Oct 2026
The founders used this technology to integrate with insurance payer dashboards, PMS systems, and other legacy solutions during dental tech operations
AI inference
The model stated this without a quote found in the source.
Hacker NewsRetrieved 10 Oct 2026

Problem severity

Many important applications used in regulated industries lack usable APIs or have incomplete ones
Public claim
A lot of important software is still difficult to automate. Healthcare teams submit prior authorizations through payer portals, accounting teams enter data into desktop applications, and operations teams move information between internal systems, spreadsheets, and remote desktops. Many of these applications have no API, incomplete APIs, or integrations that take months to build.
Hacker NewsRetrieved 10 Oct 2026
Production automation of sensitive workflows requires reliability, state tracking, and verification that existing tools struggle to provide
Public claim
Getting an agent to complete a task once is fairly straightforward. Getting it to repeat that task, recover from unexpected states, avoid silently entering incorrect data, and produce evidence of what it did is much harder.
Hacker NewsRetrieved 10 Oct 2026
Organizations spend multiple hours per day manually reconciling data across platforms and legacy systems
Public claim
Many people spend hours per day in legacy dashboards and on-prem solutions reconciling data across platforms.
Hacker NewsRetrieved 10 Oct 2026

Economic proximity

Healthcare operations workflows combine multiple difficult automation challenges with high cost of failure
Public claim
We are initially working with healthcare operations teams because their workflows combine many of the hardest conditions: payer portals, EHRs, PDFs, spreadsheets, remote desktops, and actions where quiet mistakes are expensive.
Hacker NewsRetrieved 10 Oct 2026

Differentiation

RPA fails when interfaces are not perfectly predictable
Public claim
That works when the interface and workflow are predictable, but it often breaks when a button moves, a pop-up appears, a page loads slowly, or the application enters an unexpected state.
Hacker NewsRetrieved 10 Oct 2026
Kampala can automate workflows deterministically in seconds by leveraging session tokens and anti-bot cookies
Public claim
Because Kampala is a MITM, it is able to leverage existing session tokens/anti-bot cookies and automate things deterministically in seconds.
Hacker NewsRetrieved 10 Oct 2026
Existing MITM proxies and browser automation tools are detected by anti-bot measures and fail
Public claim
They manipulated the TLS and HTTP2 fingerprint over the wire which was detected by strict anti-bots.
Hacker NewsRetrieved 10 Oct 2026
Existing browser automation and computer vision approaches are brittle and nondeterministic
Public claim
Current attempts at automation use browser automations or computer use agents which are brittle, slow, and nondeterministic.
Hacker NewsRetrieved 10 Oct 2026

Retention and timing

Vision-based computer-use models exist but lack production-grade reliability infrastructure
Public claim
We saw that models were getting better at vision but kept seeing a gap between computer-use demos and the reliability needed for production workflows.
Hacker NewsRetrieved 10 Oct 2026

Automation for legacy software without APIs

Overview

Build this business

Verdict

Score breakdown

Risk penalties

Score history

Validation gate

Evidence · 13 items

Market validation

Problem severity

Economic proximity

Differentiation

Retention and timing

Market

Existing products

Build

Legal reconstruction check

Community