Need someone who can fix chatbot
Budget: $10 – $30 USD
My website relies on an informational chatbot to guide visitors through common questions, yet lately the answers it returns are wrong or outdated. The flow still triggers, so the underlying trigger logic seems fine—the real problem is the quality of the responses.
I’m looking for someone to audit the current implementation, trace the point where facts are being pulled or generated, and put in place a robust fix. Whether that means retraining intents, cleaning the knowledge base, tightening API calls, or patching code in Dialogflow, Rasa, plain JavaScript, or whatever framework you discover under the hood, I’m open—as long as the end result is a bot that consistently delivers accurate, relevant information.
Deliverables
• Corrected chatbot running on the live website
• Brief change log explaining what you adjusted and why
• Quick test report showing sample queries and the improved responses
Acceptance criteria
The bot must respond accurately to the top-20 FAQs supplied after project start and sustain that accuracy across at least 50 random user-style queries during hand-off testing.
If you have solid NLP debugging experience and can turn this around quickly, let’s get started.
I’m looking for someone to audit the current implementation, trace the point where facts are being pulled or generated, and put in place a robust fix. Whether that means retraining intents, cleaning the knowledge base, tightening API calls, or patching code in Dialogflow, Rasa, plain JavaScript, or whatever framework you discover under the hood, I’m open—as long as the end result is a bot that consistently delivers accurate, relevant information.
Deliverables
• Corrected chatbot running on the live website
• Brief change log explaining what you adjusted and why
• Quick test report showing sample queries and the improved responses
Acceptance criteria
The bot must respond accurately to the top-20 FAQs supplied after project start and sustain that accuracy across at least 50 random user-style queries during hand-off testing.
If you have solid NLP debugging experience and can turn this around quickly, let’s get started.