EngageAI

EngageAI

Inspiration

We frequently encounter clients with this problem: "Drowning in Data, Starving for Connections" - they are overwhelmed by the sheer volume of data they collect, yet struggle to leverage it effectively for their business needs.

What it does

EngageAI is a proven way to:

1. Find your best customers (Machine Learning - Predictive Customer Analytics):

Harness First-Party CRM Data: Utilize existing CRM data to derive sophisticated predictive metrics tailored to each customer, enhancing precision in targeting and resource allocation.
Strategic Customer Segmentation: Deploy cutting-edge algorithms to segment customers effectively based on predictive analytics, enabling pinpoint targeting of high-potential customer segments.
Key Tool: Google Cloud BigQuery ML

2. Develop lasting relationships (Generative AI - Enhanced Customer Relationship):

Maximize Customer Lifetime Value: Strategically focus efforts to attract customers with the highest potential return and minimize resources spent on less profitable segments. Implement targeted strategies to elevate the lifetime value of existing customers, turning average customers into high-value ones.
Innovative Engagement Campaigns: Utilize tailored marketing campaigns via Telegram/WhatsApp, orchestrated through advanced segmentation based on CLV assessments and LLM Agents.
Key Tools: EngageAI (VertexAI / Gemini)

How we built it

3 days (just check out the Git repo history);
We had the idea;
We broke the project into tasks for everyone; (backend / front / deployment / API integrations / presentation video-ppt-GitHub-DevPost-demo)
We like and enjoy the serverless paradigm a lot, so we decided to go with the GCP serverless stack; (BigQuery, Cloud Run, VertexAI);
Silviu: Backend Laravel, front React, database models, deployment into Cloud Run / Cloud Build;
Christian: Backend Laravel, front React, database models, API integrations (VertexAI & Twilio);
Mike's presentation: PPT, video, demo, DevPost;

Challenges we ran into:

Twilio setup / API integration;
Laravel and Vertex AI SDK for PHP;
We saw that in the EU, WhatsApp Business does not have a payment system;
The solution is based on client first-party data (customer name, email, phone numbers);
For small clients, the cost from Gemini API / Twilio / WhatsApp API can be an issue;

Accomplishments that we're proud of

EngageAI can really have a big impact on business. It's working amazingly good; (we tested it out with dukier.com and it can scale very well);
For the audience part, we used BigQuery and Shopify Data Transfer integration; Then we created a machine learning model using BigQuery ML that will indicate the CLV at user level; and based on the CLV, we can create segments for our audiences;
It's definitely a cookieless solution;
Easy to set up:
1. Import your audience;
2. Define your campaign;
3. Run the campaign;

What we learned

That Gemini 1.5 has a "Function calling" feature 👀👀👀 (we definitely plan to use it in the future);

What's next for EngageAI:

Planning Our Journey Forward (MVP to SaaS)

Data ingestion: Integrate seamlessly the tool with best in market CRMs solutions via API;
Data processing: Think in terms of scale / security / privacy for audience part;
ML & GenAI: Unify in one platform: "Predictive Customer Analytics" and "Enhanced Customer Relationship"; Leverage advanced "Function calling" capabilities from Gemini 1.5 API;
Insights & activation: bypassing intermediaries like Twilio (direct integrations with Whatsapp / Telegram etc; Enhance the "Engage (log)" more visibility, control over actions, workflows, logs and results;

Code on GitHub: (private repo, but we shared access to testing@devpost.com ) https://github.com/seftimie/googleai/