Niliunda GenAI Chatbot yenye nguvu ya titan lakini kununua Amazon Q inaweza kuwa na akili zaidi

Niliunda chatbot ya Gen AI ili kutafuta kupitia kadi ya nyaraka za ndani kwa kampuni ya huduma za kifedha ili kuruhusu watumiaji na chombo cha kutafuta, kupeleleza, na kuzingatia nyaraka kwa kasi zaidi. Kuunda ufanisi wa 80%, niliitumia Retrieval-Augmented Generation (RAG), kuingizwa nyaraka katika kadi (kila kikundi cha mapendekezo), na niliitumia Titan baada ya kulinganisha LLMs kama Claude, Titan, na Llama-3 kutoa majibu ya mazingira, sahihi kwa utafutaji wa nyaraka.

Kwa ajili ya kazi ya sprints tatu ya kulinganisha mifano, kubadilisha vifaa, kujaribu kuboresha utendaji, na kupiga marekebisho - ilifanya kazi vizuri sana kwenye mfano wa Titan (Titan inafanya kazi bora kati ya mifano iliyochaguliwa kwa sababu imeundwa kwa mfano wa maandishi na kesi yangu ya matumizi iko chini ya kiwango cha uwezo wake). Watumiaji wote niliyoonyeshwa na kuonyeshwa walipenda. Ilikuwa kuokoa masaa ya kutafuta na kupeleleza nyaraka kwa FAQs na T & C ya kiwango kwa nyaraka za bidhaa za kifedha (sayansi kubwa na kamili ya alama za miguu, nk). Ingawa ninajivunia matokeo ya kazi yangu, nilikuwa na swali moja katika akili yangu: Nina uhakika 100% kwamba hii ni suluhisho na bei bora na

matatizo ya mafanikio

When you build something that works well to test and validate an idea, it feels like a great achievement —until you realize the bar is not just “working.” Although customers are overly cautious about using GenAI in highly regulated industries such as financial services, it still doesn’t seem enough for me to charge a higher price than enterprise-ready solutions like Amazon Q, Microsoft Copilot, and Google Gemini. Recognising that my solution with RAG that are increasing confidence on responses, eliminating hallucinations, and quality with many guardrails, I wasn’t satisfied that it is outperforming what's already out there. Enterprise-ready platforms are cheap, well-looking, and deeply integrated with ecosystems that businesses already live in. Amazon Q Business, for example, can index your S3 docs, handle access control, and costs next to nothing compared to the infrastructure I’d need to support high-volume GenAI queries on Titan or Claude.

Katika kesi yangu, Amazon Q Business Enterprise inachukua $ 0.264 kwa saa kwa kiwango kimoja (20K hati au 200Mb ya maandishi yaliyotolewa), wakati nililinganisha gharama yangu kwa shughuli kwa hati kama $ 0.23.

Mimi nilikuwa na kiburi cha chatbot niliyojenga. Lakini kwa masharti ya kupanua na ufanisi wa gharama? Mimi nilikuwa ghafla katika mahali gumu sana.

Uamuzi mgumu: kujenga vs. kununua katika Era ya GenAI

Huu sio tatizo jipya, lakini na GenAI, mambo yamekuwa magumu zaidi. Hapa ni muhtasari wa kile nilichoweza kujifunza, ikilinganishwa wakati wa maendeleo ya chatbot ya ndani ambayo ni sawa na AWS Q Business (kwa baadhi kuimarisha):

Control

Wewe kuwa na customization kamili (RAG, LLM tuning,
Uhandisi wa haraka)

Uwezekano mdogo kwa uwezo wa jukwaa - lakini rahisi sana

Data Privacy

Unaweza kuzuia encryption ya kibinafsi, anonimization, au kuongeza ngazi mpya

Sera zilizoelezwa na ushirikiano - mahitaji ya usanidi

Cost

Kiwango cha juu sana (hasa ikiwa unatumia mazingira ya aina ya Bedrock)

Ufanisi wa gharama kwa ufumbuzi wa kiwango cha biashara

Setup Time

Wiki za uteuzi wa mfano, maendeleo halisi, QA, iterations

Hours, sometimes minutes (can have account support)

LLM Model Options

Chagua mfano wako (Claude, Llama-3, Titan, nk)

Kufungwa katika uchaguzi wa jukwaa (Amazon = Titan / Q, Microsoft = GPT-4)

Maintenance

unahitaji kusimamia mwenyewe - kupanua, uptime, latency tuning

Kuendeshwa na mtoa huduma

Je, ni thamani ya kujenga ndani ya nyumba?

Ndiyo na siyo.

Ikiwa wateja wangu wanataka kuwa na udhibiti wa 100% na kunilazimisha kutekeleza kanuni zote zilizojulikana (hata kama hazipatikani kabisa na zinahitajika) juu ya usindikaji wa hati, utafutaji, na ufafanuzi wa mfano, kujenga ndani ina maana. Kwa sababu udhibiti wa mara mbili wa mahitaji ya udhibiti na miundo ya maombi ya kibinafsi inaweza kuunda thamani kwa biashara; wanaweza kuwa tayari kulipa zaidi kwa thamani ya kuongeza. Hata hivyo, kama nilihitaji kupanua haraka? Au kama nilihitaji kitu kizuri (85-90% ya matukio ya matumizi) bila kuwekeza fedha nyingi mapema, Amazon Q Business ghafla inakuwa ya kuvutia sana.

Unapaswa kujenga ikiwa:

• Unahitaji udhibiti zaidi juu ya data na tabia ya mfano
• Unahitaji kufanya kazi ndani ya mchakato wa kazi wa kipekee kama vile nyaraka za kifedha ngumu, RAG ya vyanzo vinginevyo.
• Unataka kuongeza vipengele vingine vya AI ndani ya bidhaa kuu
• Shirika lako ni tayari kuwekeza mapema kwa ajili ya maendeleo, utunzaji na gharama za infra

Unapaswa kununua ikiwa:

• Unataka ushindi wa haraka kwa scope ndogo na showcases
• Mfano wako wa matumizi ni wa kawaida (kwa mfano, hati ya Q&A, usafiri wa sera, nk)
• Wewe ni kiwango cha gharama
• Wewe tayari katika mazingira ya wauzaji (kwa mfano AWS, Microsoft, Google)

Kufunga mawazo ya

GenAI chatbot creation can be rewarding and create quick wins for the business. It is also relatively easier to start experimenting with GenAI tooling within operations, upskilling the team. But the market moves super fast. Tech giants like AWS are lowering the barrier even further for tools like AWS Q Business. So, it always worth asking, “Should we build or buy?” Cause we’re not just competing with code in this era. We’re competing with commoditized GenAI tools created by giants with billion-dollar infrastructure, talent, and polished UI/UX as well as tooling. And that’s a different kind of battle.