AWS AI or Google AI?


Amazon Bedrock vs Google Vertex AI — Who’s Winning the AI Race?

AI is no longer a buzzword — it’s the new business backbone. Whether you’re automating a contact center, building customer analytics, or integrating natural language chat into your apps, the question is no longer “Should we use AI?” but “Which cloud AI platform should we trust?”

At DrVoIP, we work deep inside the Amazon Web Services (AWS) ecosystem — deploying Amazon Connect contact centers, AI chatbots, and voice automation. But every so often, clients ask, “What about Google AI?” So let’s take a friendly, informative look at how these two giants stack up for developers and business professionals.

AWS AI Services – Built for Builders

Amazon Bedrock and SageMaker form the backbone of AWS AI strategy. Bedrock gives you access to multiple foundation models — Anthropic Claude, Meta Llama, Mistral, Amazon Titan — through a single API. That means developers can experiment and scale without retraining or rebuilding pipelines.

SageMaker powers the entire machine learning lifecycle — from data prep to deployment. Add services like Lex for conversational bots, Comprehend for sentiment analysis, and Kendra for document search, and AWS becomes a full AI ecosystem ready for enterprise workloads.

For business leaders, the key advantage is integration. AI connects seamlessly into AWS’s vast toolkit — S3, DynamoDB, Redshift, Connect — all secured under the same IAM policy framework.

Google AI Services – Designed for Discovery

Google Vertex AI and the new Gemini API represent Google’s unified approach to machine learning and generative AI. Vertex brings together model training, evaluation, deployment, and monitoring under one interface — ideal for data scientists and AI researchers.

Google’s strength is creativity and speed. Vertex AI integrates beautifully with BigQuery, Cloud Storage, Firestore, and Colab notebooks. Developers can test, fine-tune, and deploy models in hours — not days. And the Gemini family models (successor to PaLM and Bard) deliver world-class text and multimodal capabilities for summarization, image reasoning, and code generation.

Head-to-Head Summary

Category AWS AI (Bedrock & SageMaker) Google AI (Vertex & Gemini)
Model Variety Multi-model (Titan, Anthropic, Meta, Mistral) Gemini family + open-source (Gemma, Mistral)
Ease of Use Strong for developers, steeper for business users Very accessible with notebooks and UI tools
Ecosystem Deep enterprise integrations (S3, Connect, Lex) Tight analytics stack (BigQuery, Search, Colab)
Security Enterprise-grade IAM, compliance focus Fine-grained IAM, research-oriented flexibility
Deployment Serverless, multi-model endpoints Edge and cloud endpoints, rapid prototyping

The DrVoIP Takeaway

For production-scale enterprise AI deployments — especially where security, governance, and integration matter — AWS Bedrock and SageMaker are the clear choice. They’re built for scale, built for control, and built to integrate into your existing AWS architecture.

For fast prototyping, experimentation, and data-driven innovation, Google Vertex AI shines. If you’re already running on Google Workspace or BigQuery, Vertex offers the shortest path from concept to prototype.

Our Recommendation

Most organizations don’t need to pick a side. The smartest strategy is multi-cloud AI: use AWS Bedrock for enterprise workloads and Google Vertex for innovation labs. The two can complement each other beautifully when designed with the right architecture.

Need help navigating AI services for your contact center or enterprise app? DrVoIP can help you design, deploy, and manage a secure, cost-effective AI strategy on AWS — complete with proof-of-concept, data pipeline, and integration guidance.

Contact us today: Grace@DrVoIP.com


DrVoIP — Delivering Amazon Connect, AI, and voice automation on time, on budget, with the highest customer satisfaction.

“Pick and Park” feature?

Pick & Park™ for Amazon Connect – Finally, Real Agent Call Control

Amazon Connect is amazing – but sometimes, agents need more control than a routing flow can provide.

That’s why we created Pick & Park™, a pair of features designed by DrVoIP to give agents the power to:

  • See who’s waiting in queue – and choose the right caller
  • Park any live call – and return to it later without losing context
  • Toggle between multiple calls – like a trading desk or dispatcher

What is Pick & Park™?

Pick lets agents view all waiting contacts in real time and click to answer a specific one. This is ideal for VIP customers, repeat callers, or when judgment matters more than FIFO logic.

Park enables agents to place a call into a “staged” state – not HOLD, but a true pause – so they can answer another call, then return to the original conversation when ready. Think of it as call juggling without the chaos.

Together, they form CallBoard™ – a single-agent control dashboard that works with your existing Amazon Connect instance.

Why This Matters

Most Amazon Connect deployments rely entirely on flows and automation. That’s great for scaling – but what about when a human agent needs to act like a human?

Call escalation teams, sales desks, healthcare triage, and dispatch centers all need tools to:

  • Pick the right call at the right moment
  • Switch between calls in progress
  • Park a call while researching or collaborating

🚀 Now Available: CallBoard™ – Built by DrVoIP

Pick & Park™ lives inside CallBoard™, our new agent dashboard designed to extend Amazon Connect functionality without changing your flows.

No rip-and-replace. No complex setup. No need to modify routing profiles. We deploy in under a day.

🎬 Watch the Live Demo:

This short video shows:

  • Real calls entering the queue
  • Agents using Pick & Park™ to control call flow
  • Live toggling between parked calls
  • All powered by Amazon Connect

Pick & Park™ lives inside CallBoard™, our new agent dashboard designed to extend Amazon Connect functionality without changing your flows.

No rip-and-replace. No complex setup. No need to modify routing profiles. We deploy in under a day.

Live Demo Video Coming Soon! We’ll walk you through a working version of CallBoard™, showing live calls entering queue, being picked by agents, parked, and returned – all without breaking the Amazon Connect model.

Use Cases

  • VIP Queue Monitoring – Prioritize high-value contacts manually
  • Escalation Handling – Park and rejoin calls while gathering solutions
  • Trading Desk / Buy-Sell – Switch between multiple parties in real time
  • Dispatch Centers – Queue scanning and fast response to critical contacts
  • Tech Support – Move between sessions without losing context

How to Get It

We’re currently offering Pick & Park™ as a deployable solution for Amazon Connect environments.

Want to see it in your instance? Reach out and we’ll spin up a live demo.

Or visit our Contact Page to request a walkthrough.

Amazon Connect Emergency Notification System

Emergency Notification?

There are a number of use cases for an application that can process a list of contacts to be notified by voice message of an impending event.  The event may be an emergency or  a notification of a community activity of interest to the list subscribers.   Even an appointment reminder falls into this category, so it is not necessary to have an actual ’emergency’ to have a requirement to mass notify a subscriber or customer group.

Basic Functional Spec

We had need of a “campaign dialer” like a solution and we found that all the parts we needed were available in AWS and Amazon Connect.   We hammered out a basic functional statement that had the following features:

  • The application requires the admin to authenticate;
  • A facility exists to record a new message to be played by the dialer;
  • The Dialer had to be able to simultaneously dial a large number of list members by contact group code;
  • The called individuals could enter a response code  to enable a transfer or follow up action (press 1 to transfer to an agent, Press 2 to reschedule,  etc.)
  • All contacts would be coded as notified even if they hung up, to eliminate redials.
  • A facility exists to reset the database in preparation of a new campaign.

The resulting solution contained the following modules:

  • Dialer Setup Module: (Authorization to use the notification system; Menu to reset database,  record new outbound voice message and initiate outbound campaign
  • DynamoDB table to contain the subscriber or customer list with associated fields to support the application (subscriber name, zip code, result code, etc.)
  • Lambda functions to read/write subscriber database and record audio to prompt file
  • Administrator notification of campaign dialer status through SNS

The solution is coded with four contact flows that enable group alerts to be sent simultaneously.  The AWS services used include Amazon Connect, DynamoDB, Lambda, Kinesis and optionally SNS. Depending on your contact center soft limits, the dialer will contact a minimum of 50 subscribers per phone call.  The following video details the configuration and the entire application is available in the DrVoIP.com store including the lambda functions and sample DB configurations.

 

 

Agent Step-by-Step Guide

Amazon Connect Agent Scripts?

We have been very impressed with Amazon Connect’s growing move into an embedded CRM.  Customer Profiles and Cases are powerful capabilities that are now integrated in the “pay only for what you use” cloud based contact center.  Recently, Amazon has added “step-by-step” Agent guides!  This powerful new functionality brings “one call” customer problem resolution from concept to reality.    It offers a unified experience for contact center agents to access the tools they need to address customer calls effectively.  Each phone call can generate  a screen pop of information resources.   These information or action resources can be tailored  for each queue.  This sample screen shot illustrates what might be presented to an Agent in an Airline who might handle reservations and claims for lost luggage!

Step-by-Step Agent Guides

Using the existing contact flow designer, you can configure “cards” (picture above”) that prompt the agent through how best to handle a customer request.  You can create detailed views, forms and confirmations that kick off back end actions!  No longer is it necessary to open a separate  browser tab to update your CRM, this can all be handled by the back end functions lunched from the guides.

Give us a call to setup a demo or to discuss how this amazing new set of productivity solutions can assist in achieving your contact center vision! – DrVoIP@DrVoIP.

WTF is a DNIS Map?

Route by DNIS?

A common call center request is to provide a custom greetings or route a call to a Customer Service Queue (aka CSQ) based on the number the caller dialed.   DNIS or “dialed number information service” is usually the solution to this request.   Typically you create a database table in which the index is the DNIS.  In this way we can pull back all the information we need to greet and route the caller.  The solution consists of mapping a DNIS number to a contact flow that has a lambda function that looks up the greeting and routing details in the database referenced the function.DNIS MAP

Assign the incoming DNIS number, actually all incoming numbers should hit this contact flow, and we do the usual setup.  Turn on logging which is very useful during testing and you can turn it off later to save a few pennies.  Set up recording, preferred voice and then invoke the lambda function that will retrieve the desired greeting and routing information.  It is a good practice to set the contact attributes so that you can easily reference the returned variables in subsequent contact flows and to make them available for the CTR records and potential screen pops.

Contact Attributes

Clicking on the “Set Contact Attributes”  call flow step in this example we get the following:

Contact Attributes You can see that the lambda function is pulling back two variables from the associated database tables: queueId and flowId.   in the above contact flow, after we set the contact values, obtained from an “External” database  we”transfer to flow”.   The external flowId points to the flow we want this DNIS to be used for follow on call handling.

The flow will eventually transfer to a queue, which we obtained from the same “External” database. Like the flowId, the queueId is now a “user defined” value. No reason you could skip the “set contact attributes” step and just use the $,External.flowId but by the assignment and the use of $.Attributes.flowId we associate the value with the CTR and make it more easily available for screen pops etc.

(see the blog on should I route to a queue or route to a flow?)

Additionally we could also pull back any other information we might want that would change based on the number the caller dialed. For example a customer greeting.   Did the call the Kin-sue Knife help line or the Hotel reservation line?   Maybe this is an Executive suite application in which the operator has to answer “thank you for calling <yourcompanyname>.

Advanced DNIS Routing

It may be that we want to route the caller to an IVR Menu of options based on the DNIS. You can easily handle this by the flow you route the caller to, but do you really want to do the newbie “copy and save as” to repeat the same contact flow for each DNIS?   There is now reason why the database can point to a generic IVR flow.  The generic IVR menu would have menu options that are variables that might be named “optionOne”, “optionTwo” etc.   These options would be defined in the database and in this way, each DNIS could point to the same IVR based contact flow but the menu prompt and all the options would be obtained via the lambda function from the referenced database.

This solution along with the Lambda function and Video tutorial is also available in our online store!

As always we would be glad to assist with this solution just click or call! – DrVoIP@DrVoIP.com

 

 

 

 

Adding Video to your Amazon Connect Contact Center

Why Video?

If a picture is worth a thousand words then a video is an encyclopedia of information!   I remember a team building exercise in communications in which all the team mates sat with their back to the grease board at the front of the room.  Each team mate  had a blank sheet of paper and a pen on their desktop.   Another team member drew a simple share on the grease board and then described it verbally to the rest of the team.  Instructions included how to orient the paper and a verbal map of how to replicate the grease board drawing was narrated.   It was amazing how many different versions of the illustration were created by the team with verbal descriptions from the team artist!  No two alike!

So now put yourself in a technical support contact center listening to your caller describing how they have the yellow cable plugged into the thing next to the other thing?   How more effective would it be if the help desk technician could see the wifi router the caller was trying to install.    One call resolution time would be reduced,  average holding time would be reduced and SLA’s would improve as agents spend less time listening to caller descriptions and more time seeing issues enabling them to handle more calls and solve more problems more quickly.

Mobile Video Chat!

Enabling a link on the company website that opens a chat session with a customer service representative in your call center has always been possible with Amazon Connect.   Now with the aid of AWS Chime, you can escalate a keyboard chat to a video call complete with audio!  You can also push the chat link out to IOS and Android devices, making it possible for mobile video solutions.    Insurance companies could see real time accident reports.    Contractors could quote jobs more quickly, reducing time by eliminating site visits in favor of  video ‘meet ups”.   Telemedicine is also a rich vertical for video in the call center.

Video for Amazon Connect Contact Centers

Integrating video into your Amazon Connect contact center is relatively straight forward.   The solution makes use of a custom CCP, Chime SDK and a bit of serverless code on the back end.    Agents are alerted to an incoming request through the Amazon Connect chat facility using LEX to orchestrate the dialog and gather preliminary data from the caller.  Once the chat is connected between the caller and the agent, both have buttons to escalate to video.

REQUEST  A DEMO

If you would like a demo of this functionality, send us a request and we will send you a link. DrVoIP@DrVoIP.com

 

 

 

 

 

 

 

Amazon Connect Call Center Frequently Asked Questions!

What is Amazon Connect?

Amazon Connect is a modern, cloud-based contact center platform from AWS that enables businesses to deliver seamless customer experiences across voice, chat, and other channels. Unlike traditional systems, Amazon Connect has no per-seat licensing fees — you pay only for what you use.

At DrVoIP, we’ve been designing, deploying, and optimizing Amazon Connect environments since 2017. We specialize in rapid deployments, advanced integrations, and fixed-fee pricing.

What does it cost to deploy?

Amazon Connect is a consumption-based service. You are billed by the minute, message, and feature used. Here’s a quick snapshot:

  • Voice minutes: ~$0.018 per minute (inbound/outbound)
  • Direct dial numbers: $1 per number/month
  • Chat sessions: ~$0.004 per message
  • Optional features: Amazon Lex bots, Lambda integrations, and Contact Lens

DrVoIP Fixed-Fee Packages start at $6,000 for a 5-day quickstart implementation and scale based on complexity. You’ll always know the cost before you begin — no surprises.

How long does it take to go live?

Our standard deployment timeline is:

  • 5 business days for a core voice contact center
  • 7–10 business days for multichannel or AI-enhanced solutions

We’ve launched entire centers in under a week, including agent training, call flows, and queue configurations.

What is included in a DrVoIP Quickstart?

Each quickstart includes:

  • Amazon Connect instance setup (or optimization)
  • Inbound call routing, queues, and prompts
  • Agent configuration with email and softphones
  • Custom IVR flows and menus
  • Live dashboards for real-time and historical reporting
  • Admin training and documentation

Can Amazon Connect support my business size?

Yes. Amazon Connect is used by companies ranging from startups to global enterprises. It scales instantly, requires no on-premises hardware, and is supported globally by AWS infrastructure.

What makes DrVoIP different?

DrVoIP delivers:

  • Rapid implementation (days, not months)
  • Flat-fee pricing with no surprises
  • Full AWS integration (Lex, Lambda, Pinpoint, S3)
  • Proven success with contact centers large and small
  • US-based engineering support — not outsourced

What about SMS, AI bots, screen pops, and CRMs?

We can enable any or all of the following:

  • SMS and text messaging integrations
  • AI voice bots using Amazon Lex
  • CRM integrations (Salesforce, HubSpot, Zoho, and others)
  • Screen pops with contact attributes on inbound calls
  • Call recording, transcription, and sentiment analysis
  • Secure call recording storage and compliance configuration

Is my data secure?

Yes. Amazon Connect runs on AWS infrastructure and meets strict security and compliance standards including PCI, HIPAA, and SOC 2. All data is encrypted in transit and at rest. We implement role-based access controls and audit trails as part of our deployments.

What support options does DrVoIP offer?

We provide on-demand and managed support plans:

  • Business hours support: 7 AM – 7 PM EST
  • After-hours support (premium rates apply)
  • Prepaid blocks starting at $1,500 for 10 hours
  • Time tracked by ticket — calls and emails billed in 15-minute increments

All support is logged through our secure ticketing portal. No email black holes, no vague billing.

Where can I learn more?

📞 Voice: 844-437-8647 (844-437-DrVoIP)
💬 SMS: 844-798-3504
📧 Email: sales@drvoip.com for a detailed Amazon Connect PLanning Guide

If you’re thinking about Amazon Connect, you’re already ahead of the curve. Let us help you get there faster.

Amazon Connect – Start/Stop/Pause/Resume voice recording?

Recording and Analytics

While configuring Amazon Connect contact flows you will find a “SET Recording and Analytics?  step that you can make use of to set the voice recording behavior.   When you set the recording behavior you select recording only one channel or both the Customer and the Agent.

LENS Speech Analytics

In addition to setting the recording behavior you can switch on LENS a transcription and comprehension service.  This service is typical of call centers as they manage to improve Agent performance and customer satisfaction.   You also make a choice between post call transcriptions or real time subscriptions that can be used to alert supervisors to a real time need for interaction.

 

Start/Stop/Pause/Resume

 

Often for reasons of compliance  HIPA or PIC for example you might want to turn off voice recording while personal data, like credit card information, is being provide.   To do this, it will be necessary to create an extension to the CCP to enable the buttons need to effect the feature.    The clip below reviews a customized CCP we created to handle several features including the start/stop recording API as published.

Give us a call and let us help you with this requirement!  DrVoIP@DrVoIP.com 

 

Amazon Connect – Play Pre-Recorded messages to remote Voice Mail?

Recorded Message Use Case

In an outbound  call center an agent may make several hundred phone calls each and every working day.     When the person they dialed does not answer and the call is forwarded to voice mail, the Agent leaves a message requesting a call back indicating who to call and at what number?    The message is most likely very similar for each called party, requiring the Agent to be enthusiastic and engaging for each message left!

Is there a better way?

There is a cheap and dirty way to do this!   Each Amazon Connect Instance has a “Default Outbound” contact flow.   This contact flow is used for all outbound calls even if you do not specify it.  In fact you cant specify it, or an alternative.  Usually in contact flows, you can do a “save as” on any of the sample or default contact flows and then reference them in your own contact flow designs.   The Default Outbound contact flow however can not easily be used n this way.

What you can do is modify the outbound contact flow and add a “set disconnect” step in that contact flow.    This step is activated when the Agent hangs up and the outbound line is still up.   In this use case, the “set disconnect flow” point to a simple contact flow that has  “play prompt” which contains the message you want to leave in the called parties voice mail box.  The agent, hearing the call has been forwarded to a voice mail, simply hangs up!  The contact flow then plays the prompt asking for a call back!  Cheap but very workable!   (This is a video clip that reviews this configuration).

Note: The Lambda function in this configuration is to support a different function, that of selecting the Caller ID the agent wants to display to the called party.  You can see the details of that configuration at this link.

How about multiple message play options?

What about enabling the Agent to select from multiple messages?  Can each Agent have their own pre-recorded messages? The use case would be the same as configuration as above, but with a couple of changes.  The first change is we need to create a custom CCP that would have the buttons we need to indicate which message the Agent wants published to the called party.  This might be individual buttons or it might be a drop down list of available numbers.    Secondly, we would need a lambda function to retrieve the messages from a Dynamodb table prepared for this purpose.

The following video reviews how this configuration might work with a customized CCP.

Give us a call or email DrVoIP@DrVoIP.com if we can help you with this function!