Generative AI can produce campaign lines, product names, packaging concepts and business ideas for a Georgian company in seconds. The speed is real. The risk is that competitors using the same models, similar prompts and the same public information may reach the same polished answer faster.
Research from Wharton illustrates the tension. AI-generated product ideas can score well on purchase intent and appear among top-ranked concepts, yet show higher similarity and lower novelty. In another experiment involving a toy made from a brick and a fan, 94% of ChatGPT-assisted concepts overlapped, while the unaided human ideas were unique. These are international results, not measurements of Georgian firms.
Why Georgia needs an originality system
Georgia is a small, highly connected market. Companies often follow the same platforms, global trends, agencies and formats. A generic instruction such as “create a modern campaign for Georgia” supplies little proprietary material. The model therefore gravitates toward the category median: safe emotion, familiar wording and a visual direction that any competitor can reproduce.
The relatively smaller digital footprint of the Georgian language and local culture adds a second challenge. Grammatically correct Georgian does not automatically mean Georgian thinking. Regional detail, generational speech, local humour and sector experience must enter the process through people and proprietary evidence.
BTU researchers’ Originality Index
BTU researchers propose a 100-point internal audit. Ideas receive up to 20 points for each of the five dimensions: local truth, brand ownership, conceptual distance, human evidence and resistance to imitation. The index is a management tool rather than a national statistic. A low score signals that the idea could travel unchanged across countries or competitors and therefore needs more research and reframing.
Measure directions, not wording
Thirty outputs are not thirty ideas if 22 repeat the same problem, emotion and format. BTU researchers propose a simple portfolio diversity coefficient: distinct conceptual directions divided by total outputs. Thirty outputs clustered into eight directions produce 26.7% diversity; 18 directions produce 60%. The measure does not assess quality. It reveals how widely the team searched.
The break-even calculation
Suppose a Georgian company spends an extra GEL 12,000 on interviews, local research and human concept development. The campaign reaches 200,000 potential customers and earns GEL 30 contribution margin from each incremental customer. It needs 400 additional customers to cover the investment, equal to 0.2% of reached consumers. If baseline conversion is 1.0%, break-even is approximately 1.2%. This is an illustrative BTU scenario, not a market forecast.
Use AI after divergence, not before it
The weakest workflow is a blank page, a generic prompt and the first acceptable output. A stronger workflow begins with customer language, sales objections, regional realities, the founder’s choices, failed experiments and proprietary assets. Humans then form several genuinely competing frames. AI expands, combines and prototypes them, challenges assumptions and helps test versions. The decision remains human-owned.
Conclusion
AI can lower the cost of ideation and raise average quality. But once competent average output becomes cheap and universal, it stops being an advantage. BTU researchers conclude that Georgian companies should use AI as an amplifier of distinct human starting points, not as a substitute for originality. The objective is not more answers; it is more independent directions grounded in private knowledge, local experience and accountable human taste.