What if you could slash your translation time by 90% — without giving up control?
In this eye-opening interview, Bulgarian translator and journalist Kaloyan Kirilov reveals how he uses CotranslatorAI to speed up complex translation tasks, lock in formatting consistency, and even code custom workflows through language. Whether he’s tackling interviews, contracts, or high-volume journalism, Kaloyan shows how strategic prompt design and smart AI workflows give translators a serious edge — even as clients say, “Why not just use ChatGPT?”
00:00:03 – Introduction and discussion of scripted prompts, pronunciation clarification, and setup.
00:00:52 – Background: translation and journalism experience, boutique LSP, role in news agency.
00:02:03 – Nature of translation work at the news agency: repetitive content, deadlines, and context.
00:03:00 – Use of CotranslatorAI for repetitive datasets (e.g., weather updates), anchor prompts, and speed gains.
00:04:06 – Comparison with DeepL: consistency issues, formatting, following instructions, and limitations.
00:06:39 – Handling longer-form articles and interviews: diarization, formatting, and consistency in CotranslatorAI.
00:09:20 – CAT tool usage vs. non-CAT workflows: Word files, browser systems, distractions, copy-paste efficiency.
00:11:11 – Adapting CotranslatorAI to contexts beyond CAT tools: press releases, journalism requirements, and non-standard workflows.
00:12:09 – Overcoming limitations of obscure CAT tools; building glossaries in CotranslatorAI workflows.
00:13:34 – Parallel use of CotranslatorAI with Trados features: QA, autocomplete, and combining strengths.
00:14:45 – Segment-by-segment workflow vs. export-bilingual workflow in Trados; importance of context.
00:16:31 – Iterative approach using anchor prompts and “gate workflow” for higher quality output.
00:18:00 – Specialized prompts (e.g., GDPR contracts), merging prompts dynamically, and smart prompt design.
00:20:00 – Prompt libraries, shortcuts, and structured workflows compared with default ChatGPT use.
00:22:22 – How customized workflows become intellectual property and provide competitive advantages.
00:23:00 – Professional positioning with AI: why workflows and structured methods distinguish professionals.
00:25:00 – Market realities: client perception, “good enough” work, pressures on rates, and workflow-driven value.
00:26:23 – Discussion of low-resource languages: Bulgarian translation quality over time with GPT models.
00:28:31 – Practical uses: hybrid approaches with CotranslatorAI enhancing DeepL outputs for low-budget clients.
00:30:11 – Expanding beyond translation: text formatting, OCR cleanup with CotranslatorAI, repetitive document tasks.
00:33:36 – OCR + CotranslatorAI workflows for formatted documents including certificates, IDs, and tables.
00:35:25 – Handling interviews and poor transcriptions: using prompts to clean and recover lost meaning.
00:38:03 – Discussion on OCR+translate workflows, skipping steps, project vs. personal preferences.
00:39:25 – Debate on one-step vs. two-step translation workflows; use of iterative prompts for fluency.
00:41:03 – Repeating prompts on output text for elevated quality, terminology consistency, and fluidity.
00:44:00 – Clarifications: re-running prompts directly on the translated text, prompt design considerations.
00:45:58 – Discovery of iterative revision method by mistake; value of experimentation.
00:46:11 – Key advice for beginners: start with simple prompts, small glossaries, and build skills iteratively.
00:49:00 – Personal journey with CotranslatorAI; overcoming complexity, learning incrementally.
00:51:12 – CotranslatorAI improves cognitive load, focus, and quality even in basic workflows.
00:52:07 – Philosophical view: prompting as a form of coding; translators as “prompt engineers.”
00:54:06 – Expanding into multi-language/team workflows: writing prompts for colleagues, creating reusable scripts.
00:56:04 – Importance of writing prompts in target language for better conditioning; observed improvements.
00:57:10 – Smart actions and shortcuts: in-line commands in documents for repetitive elements like stamps.
00:59:59 – Using commands to auto-generate boilerplate and complete repetitive entries (birth certificates, signatures).
01:01:17 – Collation prompts: reorganizing bilingual texts for editing, and managing shortcut overload with remapping.
01:03:18 – Organizing prompt libraries by client and language pair; maintaining intellectual property.
01:04:32 – Use of tabs: when working in CAT tool contexts vs. Word exports; niche but helpful feature.
01:06:00 – Interpreter use cases: cotranslator for last-minute prep, glossaries, and remapped mouse shortcuts.
01:09:11 – Using AI-assisted workflows in live interpreting scenarios, especially with technical subjects.
01:13:14 – Beyond translation: handling client correspondence, estimates, and LinkedIn content with cotranslator.
01:15:26 – On-the-spot commands and in-place editing, including gender and capitalization handling in Slavic languages.
01:17:16 – Replacing global find/replace with AI-based context-aware replacements for gendered or capitalized words.
01:18:11 – Summary reflections: importance of simple-step breakdowns, flexible workflows, and user innovation.