How Planview constructed a scalable AI Assistant for portfolio and challenge administration utilizing Amazon Bedrock

This submit is co-written with Lee Rehwinkel from Planview.

Companies in the present day face quite a few challenges in managing intricate initiatives and applications, deriving worthwhile insights from huge knowledge volumes, and making well timed choices. These hurdles incessantly result in productiveness bottlenecks for program managers and executives, hindering their capacity to drive organizational success effectively.

Planview, a number one supplier of linked work administration options, launched into an formidable plan in 2023 to revolutionize how 3 million world customers work together with their challenge administration functions. To comprehend this imaginative and prescient, Planview developed an AI assistant known as Planview Copilot, utilizing a multi-agent system powered by Amazon Bedrock.

Growing this multi-agent system posed a number of challenges:

• Reliably routing duties to applicable AI brokers
• Accessing knowledge from varied sources and codecs
• Interacting with a number of software APIs
• Enabling the self-serve creation of recent AI expertise by totally different product groups

To beat these challenges, Planview developed a multi-agent structure constructed utilizing Amazon Bedrock. Amazon Bedrock is a completely managed service that gives API entry to basis fashions (FMs) from Amazon and different main AI startups. This permits builders to decide on the FM that’s greatest fitted to their use case. This method is each architecturally and organizationally scalable, enabling Planview to quickly develop and deploy new AI expertise to fulfill the evolving wants of their clients.

This submit focuses totally on the primary problem: routing duties and managing a number of brokers in a generative AI structure. We discover Planview’s method to this problem throughout the growth of Planview Copilot, sharing insights into the design choices that present environment friendly and dependable activity routing.

We describe personalized home-grown brokers on this submit as a result of this challenge was applied earlier than Amazon Bedrock Brokers was typically out there. Nevertheless, Amazon Bedrock Brokers is now the really helpful answer for organizations wanting to make use of AI-powered brokers of their operations. Amazon Bedrock Brokers can retain reminiscence throughout interactions, providing extra customized and seamless consumer experiences. You may profit from improved suggestions and recall of prior context the place required, having fun with a extra cohesive and environment friendly interplay with the agent. We share our learnings in our answer that can assist you understanding how one can use AWS know-how to construct options to fulfill your targets.

Answer overview

Planview’s multi-agent structure consists of a number of generative AI parts collaborating as a single system. At its core, an orchestrator is liable for routing questions to numerous brokers, gathering the discovered info, and offering customers with a synthesized response. The orchestrator is managed by a central growth staff, and the brokers are managed by every software staff.

The orchestrator includes two most important parts known as the router and responder, that are powered by a giant language mannequin (LLM). The router makes use of AI to intelligently route consumer questions to numerous software brokers with specialised capabilities. The brokers will be categorized into three most important varieties:

• Assist agent – Makes use of Retrieval Augmented Technology (RAG) to offer software assist
• Information agent – Dynamically accesses and analyzes buyer knowledge
• Motion agent – Runs actions inside the software on the consumer’s behalf

After the brokers have processed the questions and offered their responses, the responder, additionally powered by an LLM, synthesizes the discovered info and formulates a coherent response to the consumer. This structure permits for a seamless collaboration between the centralized orchestrator and the specialised brokers, which supplies customers an correct and complete solutions to their questions. The next diagram illustrates the end-to-end workflow.

Technical overview

Planview used key AWS providers to construct its multi-agent structure. The central Copilot service, powered by Amazon Elastic Kubernetes Service (Amazon EKS), is liable for coordinating actions among the many varied providers. Its obligations embody:

• Managing consumer session chat historical past utilizing Amazon Relational Database Service (Amazon RDS)
• Coordinating site visitors between the router, software brokers, and responder
• Dealing with logging, monitoring, and gathering user-submitted suggestions

The router and responder are AWS Lambda capabilities that work together with Amazon Bedrock. The router considers the consumer’s query and chat historical past from the central Copilot service, and the responder considers the consumer’s query, chat historical past, and responses from every agent.

Software groups handle their brokers utilizing Lambda capabilities that work together with Amazon Bedrock. For improved visibility, analysis, and monitoring, Planview has adopted a centralized immediate repository service to retailer LLM prompts.

Brokers can work together with functions utilizing varied strategies relying on the use case and knowledge availability:

• Present software APIs – Brokers can talk with functions via their present API endpoints
• Amazon Athena or conventional SQL knowledge shops – Brokers can retrieve knowledge from Amazon Athena or different SQL-based knowledge shops to offer related info
• Amazon Neptune for graph knowledge – Brokers can entry graph knowledge saved in Amazon Neptune to help advanced dependency evaluation
• Amazon OpenSearch Service for doc RAG – Brokers can use Amazon OpenSearch Service to carry out RAG on paperwork

The next diagram illustrates the generative AI assistant structure on AWS.

Router and responder pattern prompts

The router and responder parts work collectively to course of consumer queries and generate applicable responses. The next prompts present illustrative router and responder immediate templates. Extra immediate engineering can be required to enhance reliability for a manufacturing implementation.

First, the out there instruments are described, together with their function and pattern questions that may be requested of every device. The instance questions assist information the pure language interactions between the orchestrator and the out there brokers, as represented by instruments.

instruments=""'

applicationHelp

Use this device to reply software assist associated questions.
Instance questions:
How do I reset my password?
How do I add a brand new consumer?
How do I create a activity?



dataQuery

Use this device to reply questions utilizing software knowledge.
Instance questions:
Which duties are assigned to me?
What number of duties are due subsequent week?
Which activity is most in danger?

Subsequent, the router immediate outlines the rules for the agent to both reply on to consumer queries or request info via particular instruments earlier than formulating a response:

system_prompt_router = f'''

Your job is to determine if you happen to want further info to totally reply the Person's 
questions.
You obtain your aim by selecting both 'reply' or 'callTool'.
You could have entry to your chat historical past in  tags.
You even have an inventory of accessible instruments to help you in  tags.


{chatHistory}


{instruments}


- If the chat historical past incorporates enough info to reply the Person's questions, 
select the 'reply' motion.
- To assemble extra info earlier than responding, select the 'callTool' motion.
- You a lot solely select from the instruments within the  tags.
- If no device can help with the query, select the 'reply' motion.
- Place your chosen motion inside  tags.
- If you selected the 'callTool' motion, present the  and the  you
wish to ask.
- Your  ought to be verbose and keep away from utilizing pronouns.
- Begin by offering your step-by-step considering in  tags.
- Then you'll give your reply in  tags.
- Your reply ought to comply with the format of one among these three examples:
When selecting the 'reply' motion, your reply ought to comply with the under instance EXACTLY:


reply


When selecting the 'callTool' motion for a single Instrument:


callTool


applicationHelp
How do I reset my password?




Executing the above, would produce the next end result:
You can even name a number of Instruments utilizing this format:


callTool


dataQuery
What number of duties are assigned to me?


applicationHelp
How do I add a brand new activity?





'''

The next is a pattern response from the router element that initiates the dataQuery device to retrieve and analyze activity assignments for every consumer:

To find out who has probably the most duties assigned, I might want to question the applying knowledge. The "dataQuery" device appears most applicable for this query.



callTool

    
        dataQuery
        Which consumer has probably the most duties at present assigned to them?   
    

The next is a pattern response from the responder element that makes use of the dataQuery device to fetch details about the consumer’s assigned duties. It experiences that the consumer has 5 duties assigned to them.

Primarily based on the chat historical past, I beforehand known as the dataQuery device to ask "What number of duties are at present assigned to the consumer?". The device responded that the consumer has 5 duties assigned to them.



In line with the info I queried beforehand, you may have 5 duties assigned to you.

Mannequin analysis and choice

Evaluating and monitoring generative AI mannequin efficiency is essential in any AI system. Planview’s multi-agent structure permits evaluation at varied element ranges, offering complete high quality management regardless of the system’s complexity. Planview evaluates parts at three ranges:

• Prompts – Assessing LLM prompts for effectiveness and accuracy
• AI brokers – Evaluating full immediate chains to keep up optimum activity dealing with and response relevance
• AI system – Testing user-facing interactions to confirm seamless integration of all parts

The next determine illustrates the analysis framework for prompts and scoring.

To conduct these evaluations, Planview makes use of a set of rigorously crafted take a look at questions that cowl typical consumer queries and edge circumstances. These evaluations are carried out throughout the growth part and proceed in manufacturing to trace the standard of responses over time. At present, human evaluators play a vital position in scoring responses. To help within the analysis, Planview has developed an inside analysis device to retailer the library of questions and monitor the responses over time.

To evaluate every element and decide probably the most appropriate Amazon Bedrock mannequin for a given activity, Planview established the next prioritized analysis standards:

• High quality of response – Assuring accuracy, relevance, and helpfulness of system responses
• Time of response – Minimizing latency between consumer queries and system responses
• Scale – Ensuring the system can scale to hundreds of concurrent customers
• Value of response – Optimizing operational prices, together with AWS providers and generative AI fashions, to keep up financial viability

Primarily based on these standards and the present use case, Planview chosen Anthropic’s Claude 3 Sonnet on Amazon Bedrock for the router and responder parts.

Outcomes and affect

Over the previous 12 months, Planview Copilot’s efficiency has considerably improved via the implementation of a multi-agent structure, growth of a sturdy analysis framework, and adoption of the newest FMs out there via Amazon Bedrock. Planview noticed the next outcomes between the primary era of Planview Copilot developed mid-2023 and the newest model:

• Accuracy – Human-evaluated accuracy has improved from 50% reply acceptance to now exceeding 95%
• Response time – Common response occasions have been diminished from over 1 minute to twenty seconds
• Load testing – The AI assistant has efficiently handed load exams, the place 1,000 questions have been submitted simultaneous with no noticeable affect on response time or high quality
• Value-efficiency – The price per buyer interplay has been slashed to at least one tenth of the preliminary expense
• Time-to-market – New agent growth and deployment time has been diminished from months to weeks

Conclusion

On this submit, we explored how Planview was in a position to develop a generative AI assistant to deal with advanced work administration course of by adopting the next methods:

• Modular growth – Planview constructed a multi-agent structure with a centralized orchestrator. The answer permits environment friendly activity dealing with and system scalability, whereas permitting totally different product groups to quickly develop and deploy new AI expertise via specialised brokers.
• Analysis framework – Planview applied a sturdy analysis course of at a number of ranges, which was essential for sustaining and bettering efficiency.
• Amazon Bedrock integration – Planview used Amazon Bedrock to innovate quicker with broad mannequin alternative and entry to numerous FMs, permitting for versatile mannequin choice based mostly on particular activity necessities.

Planview is migrating to Amazon Bedrock Brokers, which permits the mixing of clever autonomous brokers inside their software ecosystem. Amazon Bedrock Brokers automate processes by orchestrating interactions between basis fashions, knowledge sources, functions, and consumer conversations.

As subsequent steps, you may discover Planview’s AI assistant characteristic constructed on Amazon Bedrock and keep up to date with new Amazon Bedrock options and releases to advance your AI journey on AWS.

About Authors

Sunil Ramachandra is a Senior Options Architect enabling hyper-growth Unbiased Software program Distributors (ISVs) to innovate and speed up on AWS. He companions with clients to construct extremely scalable and resilient cloud architectures. When not collaborating with clients, Sunil enjoys spending time with household, operating, meditating, and watching motion pictures on Prime Video.

Benedict Augustine is a thought chief in Generative AI and Machine Studying, serving as a Senior Specialist at AWS. He advises buyer CxOs on AI technique, to construct long-term visions whereas delivering quick ROI.As VP of Machine Studying, Benedict spent the final decade constructing seven AI-first SaaS merchandise, now utilized by Fortune 100 corporations, driving vital enterprise affect. His work has earned him 5 patents.

Lee Rehwinkel is a Principal Information Scientist at Planview with 20 years of expertise in incorporating AI & ML into Enterprise software program. He holds superior levels from each Carnegie Mellon College and Columbia College. Lee spearheads Planview’s R&D efforts on AI capabilities inside Planview Copilot. Outdoors of labor, he enjoys rowing on Austin’s Woman Chook Lake.