Demonstration systems / Signal
Demonstration system · Research Intelligence Agent
Signal. Turn recurring research into a system instead of starting from zero.
Monitors approved sources for the companies that matter, scores what changed, keeps structured company records and produces decision-ready briefings and on-demand research runs with sources attached.
Useful intelligence on a schedule, with every finding traceable to its source.
Built with
- Research Agent
- Source Management
- AI Summarisation
- Structured Data
- Monitoring
- Scoring
- Briefings
- Dashboard
A Dekundy Group demonstration build set in a fictional business, Harborline Systems, a B2B firm tracking competitors and target accounts. These examples use synthetic data and are not presented as client case studies. Every business, contact, figure and message shown is fictional, created by Dekundy Group to demonstrate what can be built.
Interactive demonstration
Try Signal.
Open a record, approve a draft, run the agent. Everything is local to your browser and resets on reload.
The business problem
What this solves.
Research restarts from zero
Every proposal, account review or competitor check begins with the same open tabs, the same copying and the same notes nobody can find afterwards.
Important changes go unnoticed
A competitor launches, a target account hires an operations lead, a tender closes. The information was public. Nobody was looking that day.
Findings are not comparable
Two people research the same company and produce two different documents in two formats, so nothing can be compared or reused.
From published research
20%
McKinsey Global Institute has estimated that the average knowledge worker spends nearly a fifth of the work week looking for internal information or for the colleague who has it, and more than a quarter managing email.
Source: McKinsey Global Institute, The social economy: Unlocking value and productivity through social technologies (2012)
47%
A 2023 Gartner survey of 4,861 employees found 47% struggle to find the information or data they need to do their jobs, while the average desk worker now juggles 11 applications.
Source: Gartner, Digital Worker Survey press release, 10 May 2023 (2023)
Industry findings, not results of this demonstration.
What the system does
Input to outcome, with a person in the loop.
Each stage names who does the work: the agent, business rules, automation, or a person.
Input
01Sources
Approved sources are registered per company: sites, filings, job boards, news.
Automation
02Monitoring
Sources are checked on a schedule and new items are captured as records.
AI agent
03Extraction
The agent extracts developments and compares them with what is already known.
Business logic
04Scoring
Importance is scored against defined criteria and the signal is filed.
AI agent
05Briefing
A recurring briefing is drafted, with each finding linked to its sources.
Human approval
06Review
A person reviews before anything is circulated or acted on.
Output
07Record
Company records stay current and research runs are stored for reuse.
- Input
- AI agent
- Business logic
- Automation
- Human approval
- Output
Who does what
What the AI does, what is automated, what you control.
The AI agent
What the AI actually does
- Reads captured source items and extracts the development
- Compares each development with the existing company record
- Scores importance against agreed criteria and explains why
- Suggests an action for account signals
- Drafts recurring briefings with sources attached
- Runs structured research on demand: profile, developments, model, risks, questions
Automation
What gets automated
- Checks registered sources on a schedule
- Captures new items as structured source records
- Files signals against the right company
- Assembles the briefing on the agreed day and time
- Archives research runs for reuse
Human control
What a person controls
- Chooses which companies and sources are tracked
- Sets the scoring criteria
- Reviews briefings before they are circulated
- Decides what action, if any, follows a signal
- Corrects or annotates any record the agent produced
Potential outcomes
Designed to help
- Replace repeated manual research with a maintained record per company
- Surface meaningful changes soon after they become public
- Make findings comparable across people and time
- Attach a source to every claim in a briefing
- Shorten preparation for proposals and account reviews
No percentages are claimed for this demonstration. Where published research applies, it is cited above as industry evidence.
Components
The building blocks
- 01Research Agent
- 02Source Registry
- 03Scheduler
- 04Structured Database
- 05Scoring Rules
- 06Briefing Generator
- 07Dashboard
- 08Human Review
Adam Dekundy
Founder, Dekundy Group
Built as a demonstration by Dekundy Group
This is not an off-the-shelf product. It is the kind of system built around a business’s actual workflow.
Not every situation needs something this size, and the CA$49 audit is the honest way to find out which one yours is. Where a build is warranted, every implementation begins with agreed success criteria and is covered by the Dekundy Group Build Guarantee.
Not sure yet
Not sure something like this is what you need?
That is exactly what the CA$49 Personalized AI & Efficiency Audit is for. It will show you whether something like this is worth building, whether a tool you already have could solve the problem, or whether something much smaller would do the job.
Other demonstration systems
AI Lead & Sales Agent
LeadFlow
No enquiry waits without an owner, a summary and a scheduled next action.
Client Intake & Operations System
ClientFlow
Every new client has a record, an owner, a task list and a clear next step within minutes of submitting.
Business Command Center
Command
One place to see what is happening across the business and what deserves attention next.