Patents & research impact in agentic AI, NLP, and intelligent interfaces.
Autonomous negotiation agents. AI-driven ideation capture. Answer-synthesis architectures. Specified between 2015 and 2018 — when agentic AI was a research topic, not an industry — by Innovation Mode founder George Krasadakis, most as sole inventor, at Microsoft and Accenture, and cited ever since — in research and in later patents alike. The same inventor architected Ainna.ai.
From data mining to agentic AI.
Seeing early is a habit, not an accident. The timeline below spans twenty years of the same pattern: recognise where technology is heading, specify it, and set it down — in systems, in filings, in print, and now in production. Every step leads to Ainna.
Founded Datamine Decision Support Systems
A 20+ engineer company delivering advanced data mining, credit scoring, and dynamic pricing systems to telecom, banking, retail, and online clients — founded years before “data scientist” was a job title. The direction was set two decades before Ainna.
Filed the artificial intelligence negotiation agent
Goal-driven autonomous AI agents acting on behalf of buyers and sellers — locating deals, running multi-stage negotiation sessions, and optimising terms within user-defined parameters and live market conditions. Filed in March 2016, when autonomous agents were an academic topic, not an industry: six years before ChatGPT, and years before “agentic AI” entered commercial vocabulary. Academia noticed. So did industry — later AI-agent and negotiation patents assigned to IBM, NEC, Visa, Walmart, and Capital One cite the filing as prior art.
Patented the closed-loop natural language response synthesizer
Natural language understanding paired with dynamic answer synthesis: questions answered with coherent, composed responses — “data stories” — assembled from candidate elements and continuously improved through closed-loop feedback. The architecture of the answer engine, filed in 2016 — eight years before synthesised answers became the industry's default interface. Granted twice.
Patented voice-driven ideation session management
Speech-driven NLP that listens to live, multi-speaker ideation — capturing ideas in real time, attributing each to its speaker, autonomously enriching them with supplemental research, and structuring them for action. An AI participant in the innovation room, specified years before AI meeting assistants became a product category. Granted in 2021.
Published the complete corporate innovation methodology
The Innovation Mode: How to Transform Your Organization into an Innovation Powerhouse (Springer) puts the operating model in print — including The Always-On Ideas Channel, the blueprint for permanent, systematic, AI-assisted opportunity capture. The second edition's foreword calls the first edition's anticipation of AI's rise remarkable. Cited in peer-reviewed innovation research since.
Published the agentic innovation architecture
Innovation Mode 2.0: Designing Innovative Companies in the Era of Artificial Intelligence (Springer Nature) documents the first published architecture for agentic innovation in a corporate context — multi-agent orchestration, autonomous opportunity discovery, structured multi-dimensional assessment, and human-in-the-loop governance for innovation management.
Released the AI-native innovation system
Ainna.ai — the agentic innovation management platform, in production: autonomous opportunity discovery, multi-dimensional assessment by AI agents, and board-ready documentation, with self-serve plans and white-labelled enterprise deployment. Patent-pending technologies power Ainna.ai.
Eighteen families. Four themes.
Families deduplicate continuations and jurisdictions: one family, one invention. Citation counts are per family, from Google Scholar, June 2026. Validation runs through two channels: 600+ Google Scholar citations — and the patent record itself, where filings assigned to Apple, Google, Samsung, Sony, IBM, Adobe, Amazon, and Qualcomm cite the portfolio as prior art. The eye-tracking family alone carries 100+ citing patent records; about a third are Apple's.
A · Agentic & AI-Powered Innovation| Invention | Priority | Records | Assignee | Citations |
|---|---|---|---|---|
| Artificial intelligence negotiation agentGoal-driven autonomous AI agents running multi-stage negotiations on behalf of buyers and sellers — agentic AI within defined parameters, elasticity thresholds, and market conditions. Cited as prior art by 50+ later filings, including IBM's multi-agent conversation framework and Visa's AI-to-AI communication. | 2016 | US20170287038A1 | Microsoft | 67 |
| Voice-driven ideation session managementSpeech-driven NLP capturing ideas from live multi-speaker sessions in real time — attributing each to its speaker, autonomously enriching with research, structuring for action. | 2018 | US10891436 · granted | Accenture | — |
| Closed-loop NL query pre-processor & response synthesizerNatural language understanding with dynamic answer synthesis — coherent, composed “data stories”, continuously improved through closed-loop feedback. | 2016 | US10275539 + US11250063 · granted ×2 | Accenture | 36 |
| Candidate participant recommendationMachine-learned people-to-project matching — assembling teams for collaborative engagements from skills, context, and availability signals. | 2016 | US20180007100A1 | Microsoft | 42 |
| Smart building team identificationComputer vision and sensor fusion detecting people gathered in smart buildings — and predicting the project teams and projects they belong to. | 2018 | US10985938 · granted | Accenture | 8 |
| Invention | Priority | Records | Assignee | Citations |
|---|---|---|---|---|
| Unknown word predictor & content-integrated translatorPersonalised NLP: per-user vocabulary modelling with a machine classifier predicting unknown words — adaptive in-line annotation and translation. | 2016 | US10409903 + US11188711 · granted ×2 | Microsoft | 27 |
| Content moderation & document reliability *Machine-assessed content moderation and trust scoring — reliability indication for electronic documents at scale. | 2017 | US11106718 · granted | Accenture | 16 |
| Content recognition & communication systemEnterprise intelligence from distributed sensing — organisation-wide activity signals processed against learned baseline metrics. | 2017 | US10310471 · granted · EP family | Accenture | 14 |
| CMS content relevance to published documents *Semantic and structural analysis of published documents — surfacing related content automatically across content management systems. | 2016 | US10762140 · granted | Microsoft | 7 |
| Media organisation via viewing-environment signalsContext-aware AI organising, retrieving, and annotating media collections — driven by signals captured in real time from the viewing environment. | 2016 | US10235367 + US11188586 · granted ×2 | Microsoft | 1 |
| Invention | Priority | Records | Assignee | Citations |
|---|---|---|---|---|
| Implicitly adaptive eye-tracking user interfaceGaze-tracking interface control: implicit, hands-free navigation and focus of multi-window interfaces — human-computer interaction without commands. Cited by 100+ subsequent patent filings, including Apple's gaze-interaction patents. | 2016 | US10775882 · granted + continuation | Microsoft | 105 |
| Networked public multi-screen content delivery *Computer-vision recognition driving personalised, real-time content selection across networked public display screens. | 2016 | US20170289596A1 | Microsoft | 74 |
| Interactive online music experienceSynchronised real-time shared listening — collaborative selection and control with ML-based music recommendation. | 2016 | US20170289202A1 | Microsoft | 58 |
| Augmented experience of media presentation eventsReal-time NLP identifying entities, activities, and moments in live events — automatically synchronising second-screen content. | 2015 | US20170006356A1 | Microsoft | 29 |
| Live audience experience techniquesAudience emotional-state capture connecting in-venue and online audiences in real time — affective signals augmenting live broadcasts. | 2015 | US9854410 + EP3304920B1 · granted (US, EP) | Microsoft | 26 |
| Contextual athlete performance assessmentUser-weighted, multi-variable machine scoring of athlete performance — blending statistics, physical attributes, and social-media signals. | 2015 | US20170109015A1 | Microsoft | 33 |
| Gamified adaptive digital disc jockeyAudience-state sensing and gamification driving real-time, adaptive optimisation of live media presentations. | 2015 | US20160357498A1 | Microsoft | 4 |
| Invention | Priority | Records | Assignee | Citations |
|---|---|---|---|---|
| Respiratory therapy device data transfer *Data connection between a respiratory therapy device and a docked portable device — connected medical devices in digital health. | 2023 | WO2025024220A1 · pending (WO, CN) | ResMed | — |
* Co-invented. All other families: sole inventor. Citation counts: Google Scholar, June 2026, summed across family records.
The record continues — at Innovation Mode.
The pace has only increased. Innovation Mode Limited keeps inventing at the intersection of AI and innovation management — patent-pending architectures spanning agents, assessment, and autonomous systems, applied directly in Ainna. See it running, or talk to the inventor.
The patents listed on this page are owned by their assignees — Microsoft, Accenture, and ResMed — and appear here as the founder's invention record from senior roles at those organisations, not as intellectual property of Innovation Mode Limited. Ainna is built independently of them; its architecture is protected by separate patent applications filed by and assigned to Innovation Mode Limited, currently pending. The complete record is public — view the patents on Google Patents and the citations on Google Scholar.