Problems solved

Real problems. Practical solutions.

Sometimes the right answer is a better process. Sometimes it is AI, automation or a custom tool. These are examples of problems that were identified, simplified and turned into working systems.

These describe practical workflow problems Dekundy Group has worked on. Organisations, clients, colleagues and internal data are not identified, and every demonstration linked from this page uses synthetic data created for that demonstration.

How the work starts

The technology is not the point. The solved problem is.

Dekundy Group works on practical workflow problems, then uses whatever combination actually fits: AI, automation, software you already own, a clearer process, or a small custom tool. The work starts by understanding what is going wrong and ends with the smallest useful thing that fixes it.

Read the examples

Problem solved

TwoCents

A shared place to review a piece of work, see it in context and collect everyone's questions, suggestions and feedback before anything is changed or published.

The problem

When several people need to review a marketing campaign or project, feedback can become scattered across meetings, email, documents, chat and the platform where the work actually lives.

What that creates

People may be reviewing different versions of the same work
Feedback becomes difficult to consolidate
The work and the comments about it are separated
It is difficult to know what has actually been approved
Changes can begin before everyone has had a chance to contribute

What was built

A private review workspace, first called the Two Cents War Room. It gives a team one place to look at a piece of work, understand what the finished material will look like, and leave an opinion against the work itself. It has since grown into TwoCents, a standalone product for collaborative review and discussion in classrooms and organizations.

What it lets people do

See the project or campaign in one place
Review what the finished material will look like
Leave comments in context, against the work
Collect opinions from several people
Review the collected feedback before changes are applied

What changed

Instead of collecting opinions across several disconnected tools, the company now has a centralized review process.

Everyone reviews the same version of the work
Comments sit beside the material they refer to
Collective feedback is read as one set before anything is changed
What has been contributed, and what has not, is visible

Demonstration

Synthetic demonstration dataInteractive

A working demonstration of TwoCents that runs in your browser. Step into a class reviewing an assignment or a team reviewing a policy, choose a section, add a question or a suggestion, and watch it land in that review's jar. Then switch to the teacher's or organizer's seat to approve what is waiting and resolve a discussion. The classes, teams, documents and feedback were invented for the demonstration, and participants are roles rather than people.

Try the interactive demo

Describe a review process that is scattered

Problem solved

Buying Signal Research Agent

A customized research system that looks for fresh public signals that may indicate when an account is becoming worth targeting.

The problem

Good sales opportunities often become more relevant because something changed. Finding those changes early can mean repeatedly searching news, institutional websites, announcements and other public sources. Generic research also produces a lot of information that has nothing to do with the company's actual market.

What that creates

A new building is announced
An institution receives new funding
A department receives a major donation
A new program launches
Leadership changes
A strategic initiative receives investment
Technology modernization is announced

What was built

A customized research agent designed around a specific target market and a specific set of buying signals. It searches fresh public information and turns the relevant findings into a focused research report.

What it lets people do

Search public sources against a defined target market
Keep only the findings that match agreed signals
Record what happened and when it was published
Attach the source to every finding
Produce a focused report rather than a reading list

What changed

Instead of beginning the week with only static account lists or manually searching for developments, the team can start from a current set of organizations showing relevant signals. Each result carries enough context to judge it.

What happened, stated plainly
Why the account surfaced
When the information was published
The source it came from
Why it may be relevant to this market

What a useful signal looks like

For a company selling into medical education, a relevant finding might read like one of these.

A new medical education building is scheduled to open in 2027. That could identify an institution worth beginning to nurture well before purchasing decisions are made.
A university received a significant donation to improve technology across its health sciences programs. That creates a timely reason to investigate the account and determine whether an opportunity exists.

A signal is a reason to investigate, not evidence that an organization will buy anything. The system decides where to look first. People decide whether an opportunity exists.

Demonstration

Synthetic demonstration dataInteractive

A guided simulation of this system. Choose the changes that would make an account worth investigating, configure the agent around what you sell and who you sell it to, pick how often you want to hear from it, and read the report it produces. It takes about a minute, it runs entirely in your browser, and every organization and announcement in it is invented for the demonstration.

Try the interactive demo See Signal, the fuller research application

Describe the research you repeat every week

Your workflow

What is taking more time than it should?

If something in your work feels repetitive, scattered, manual or harder than it needs to be, start with the CA$49 AI & Efficiency Audit. The goal is to identify what is actually worth changing and the simplest useful way to improve it.

A recommendation may be all you need. Building something is optional, and "keep this manual" is a legitimate conclusion.

Get the CA$49 Audit View all examples Discuss setup or a project

Get the CA$49 Audit