The Changelog

We are excited to introduce our first agentic knowledge assistant - your 24/7 Salesforce analyst!

In this write-up, we share details of how we built this agent, along with our framework for extending it to other connectors such as JIRA, Zendesk, and Gmail. We are also introducing "Text to SQL" structured query language capabilities to our Salesforce connector.

Problem Statement

When we first built our Salesforce connector, we implemented advanced fuzzy keyword and semantic search capabilities. While powerful, we quickly discovered our customers needed more. They asked complex, parametric questions that our search couldn't handle effectively.

Our existing system handled basic Salesforce searches effectively, particularly those focused on finding specific opportunities using keywords within our indexed data (title, content, metadata). However, it struggled with more complex user queries that:

Involved relationships between Salesforce objects (Accounts, Opportunities, Contacts, and Tasks)

Examples:

"Show me all opportunities for the United Oil account."

"What tasks are associated with the United Oil account and are due this week?"

Required filtering on specific field values

Examples:

"Show me all opportunities closed won in the last week."

"What are my overdue tasks?"

Needed information not in our search index

Examples:

"What are the details on John Doe?" (Contact details)

"Who is the account owner for Acme Corp?"

Needed a broad search across multiple objects/fields, then filtering

Examples:

"Find all opportunities related to 'renewable energy' that are in 'Negotiation' stage."

"Show me opportunities related to companies or people matching 'cloud solutions'"

These limitations prevented users from asking natural, intuitive questions about their Salesforce data. Additionally, in some cases, as there was no keyword or semantic meaning to search for in the first place, the search would lead to irrelevant retrieval noise in our RAG pipeline.

Enter Text-to-SQL: Precision meets natural language

To solve this, we're introducing Text-to-SQL capabilities in our Salesforce connector. This feature allows our agent to translate natural language questions into precise SQL queries, giving you more control than ever on your search scope.

Instead of using our own SQL database, we picked Salesforce's powerful SOSL (Salesforce Object Search Language) and SOQL (Salesforce Object Query Language) APIs. This allows us to avoid crawling years' worth of Salesforce records & still provide powerful queries over all of the data in seconds.

We will now get into the details of these APIs and how to use them.

Why SOSL and SOQL?

Here's a short crash course on these powerful APIs from Salesforce.

SOSL (Salesforce Object Search Language)

Best for: Broad text searches. Imagine a Google search within Salesforce.

Use it when: You don't know exactly where the information is (which object or field) or you need to search across many objects at once.

Example:

FIND {renewable energy} IN ALL FIELDS (finds "renewable energy" anywhere).

SOQL (Salesforce Object Query Language)

Best for: Structured queries with precise filtering and relationships. Like a particular database query.

Use it when: You know exactly which object and fields you need, and you need to filter based on specific values or relationships.

Example:

SELECT Id, Name FROM Opportunity WHERE StageName = 'Closed Won' AND CloseDate = THIS_YEAR (finds opportunities that are closed won this year).

Sometimes you need both!

Broad Search + Narrowing Down: Use SOSL to find a set of possible records, then use SOQL to filter those records further based on criteria that SOSL can't handle.

Example:

Query: "Find all opportunities related to renewable energy in the negotiation stage."

SOSL: FIND {renewable energy} IN ALL FIELDS RETURNING Opportunity(Id, Name, StageName) (broad search).

SOQL: SELECT Id, Name, StageName FROM Opportunity WHERE StageName = 'Negotiation' AND Id IN (<ids from SOSL>) (filter by stage).

Thankfully, modern LLMs are trained on the language powering these APIs, allowing us to generatively create them on the fly based on user queries.

Enter 24/7 Agent for your Salesforce data

We now describe our agentic search framework, which allows users to speak to their business data. By implementing an AI agent as an intermediary between user queries and various search tools, we've created a system that intelligently determines the most efficient strategy to retrieve relevant information. The agent can also easily be extended to other CRMs and apps like Zendesk, Gmail, Calendar, etc.

Agent Architecture

The agent sits at the forefront of the architecture, acting as an intelligent & flexible middleware between user queries and the backend infrastructure. This design allows for efficient query routing and seamlessly extending the agent's capabilities in the future.

Core Tools

The framework consists of four primary tools available to the agent:

1. Salesforce Object Search Language (SOSL) API from Salesforce
2. Salesforce Object Query Language (SOQL) API from Salesforce
3. Proprietary indexed and embedded vector data for workspace content
4. Native text-search API for historical text search from Salesforce

We document these tools in detail so that the agent can understand when to use them and, if needed, how to use them effectively.

Agentic Workflow

We then give the Agent a high-level workflow to follow for every user query, but with flexibility driven by its reasoning on how to navigate the workflow. The sequence is:

1. User Intent Analysis

The agent initiates by decoding the user's intent through multi-layered analysis:

• Domain Classification: Determines whether the query relates to Salesforce objects (e.g., accounts, opportunities) or requires external tools.
• Tool Requirement Assessment: Evaluates which Salesforce search tool or mix of tools are necessary based on the user's intent and search scope.

This phase resolves ambiguities early, ensuring downstream processes align with the user's goals.

2. Tool Selection

Leveraging insights from intent analysis, the framework selects optimal tools and strategies:

• Requirement-Tool Matching: Maps query parameters (e.g., date filters, relationship mapping) to available tools.
• Strategy Optimization: Prioritizes execution order—for example, running parametric date filters before semantic keyword searches to narrow the dataset.
• Finalization: Confirms tool sequence and prepares for execution.

This phase eliminates redundant tool usage and ensures resource efficiency.

3. Tool Execution

Once the intent is translated into exact tool or mixture of tools to use, the agent moves towards orchestrating the execution of these tools. The agent has to command our backend code to do this via precise instructions. The agent does this by executing the following steps:

• Entity & relationship extraction: Based on the user query, the agent selects Salesforce objects (accounts, opportunities, tasks & contacts) & the relationships between them ("opportunities linked to Account X")
• Parametric search scope extraction: Likewise, the agent constructs the exact set of criteria the user has specified in terms of date-ranges, personnel, keywords, amounts or other fields.
• Query Construction & Submission: The agent then speaks the language of each of the tools it needs to use. It translates the above entities, relationships, and search filters into the exact parameters, configurations, and/or queries to send to the backend to execute these tools (Salesforce SOSL and SOQL API, Elasticsearch). We use a simple JSON-formatted payload to explain this to the backend.
• Response Handling: The agent then waits for the backend to respond. Once the responses start coming from the various tools, the agent collates responses, validates results, resolves errors & prepares an appropriate response for the user based on the combined response from the tools used.

The key innovation in any agentic framework is that the backend tools are decoupled from the workflow, and the agent is given some level of flexibility to use its reasoning in orchestrating the worfklow and responding to the user accordingly.

Key Benefits

There are several benefits from this release of our Salesforce connector.

• Text to SQL querying via SoSL and SoQL APIs enable extremely powerful analysis, but now possible conversationally
• Analytics become accessible - 24/7 across the sales and management organizations
• Unlimited access to historical data going back years
• Low cost and operational footprint
• Foundation for Agentic search, actions & workflows for various workplace apps

Conclusion

We are excited to introduce this next-generation Salesforce connector & hope this write-up helps explain the technical details behind how to create it in-house.

For more information about implementing this connector or to schedule a demonstration, please get in touch with our sales team via our website - www.ayraa.io

As we close out 2024, we want to share Ayraa's remarkable journey in revolutionizing enterprise search and knowledge discovery. What started as a vision to help employees feel connected to workplace knowledge has evolved into a comprehensive platform that democratizes enterprise search while pushing the boundaries of what's possible with modern AI technology.

The Evolution: From Communication Platform to Search & Knowledge Assistant

Ayraa's story began in late 2021 with a vision to help employees feel connected to workplace knowledge. Initially, we focused on building an uncluttered, reliable communication platform overlaying existing tools like Slack and similar applications. The ChatGPT revolution in late 2022 catalyzed our transition. By early 2023, we had pivoted from a communication-centric approach to building a search and knowledge discovery engine.

This transformation culminated in our December 2023 launch on Product Hunt and debut at TechCrunch Disrupt, where we received exceptional feedback. The foundation was set for what would become a transformative 2024, during which we built 90% of our current platform.

"Recent Mode" Innovation: Balancing Gen AI with Coverage at an Accessible Cost

In 2024, we introduced the "Recent" mode as one of our most practical innovations, embodying our commitment to democratizing enterprise search. We developed this hybrid approach to combine the following:

• Leveraging each app's native Search API for near-zero operational costs for historical data going back years across all apps
• Advanced indexing and vector embedding for recent data (90 days for paid users, 14 days for free users)

In the Recent mode, users work with a recent cache of their workspace to cover 95%+ of their daily knowledge search and discovery needs. Using modern Gen AI capabilities, the Recent mode provides:

• Semantic search functionality to search by semantics/meaning (e.g., sneakers vs shoes)
• Fuzzy keyword matching, which is resilient to typos (e.g., sneakres vs sneakers)
• Blazing-fast, "sub-3-second" search results with immediate streaming of answers

Meanwhile, for historical searches that go back years, users can leave the Recent mode, pick All Time or specific date ranges, and benefit from API-based searches. For date ranges not covered by indexed and embedded content, you lose Gen AI capabilities but still have historical coverage across all apps. 

This innovation enables Ayraa to maintain a disruptively low pricing while delivering high-quality Gen AI capabilities for 95%+ of the use cases.

Search 2.0: Making Workspace Knowledge Accessible

In 2024, we transformed search into a powerful default landing place for users, making your workspace more accessible than ever. We enhanced functionality with:

• Click-to-query or summarize functionality for every search result - ensuring docs, files, long threads, or complex tickets are not only searchable across the workspace but, once found, can be easily understood
• Assist-like summaries at the top of search results to provide quick answers when users don't need to look at all results
• Dual-view options: Grid view for app-specific results and List view for confidence-based ordering

Comprehensive Connector Coverage

In 2024, we expanded our integration capabilities to include:

• 15 specialized application connectors (Notion, Confluence, Salesforce, OneDrive, Box, etc.)
• Slack bot for seamless in-platform interaction
• Browser extension that serves as a web co-pilot for the millions of web pages, including Wikipedia pages, blogs, Reddit threads, etc.
• Complete Microsoft ecosystem support (SharePoint, OneDrive, Outlook Calendar, Teams, etc.)

Expanding Knowledge Capture & Search to Meetings

We launched our Meetings app in mid-2024 with support for all major meeting applications (Zoom, Slack, Microsoft Teams, Google Meet).

We initially launched with a browser extension approach that leveraged closed captioning in web interfaces of meeting platforms. While highly effective for specific use cases, this approach had limitations. Users often preferred native desktop apps over web interfaces, and the browser-based solution couldn't serve users who couldn't attend meetings personally.

This led us to develop our bot-based approach, which became our primary solution. The bot was designed to synchronize with users' calendars to automatically attend meetings on their behalf, providing a more seamless experience. This approach solved both the native app limitation and the meeting attendance constraint.

However, we recognized that bot attendance sometimes created unnecessary friction - each bot has to "knock" to join meetings, potentially disrupting hosts and participants. In 2025, we will build a botless alternative to address such needs. This hybrid approach allows the bot-based approach to serve scenarios where users are absent, while the new botless recording will cater to use cases where the user prefers uninterrupted meeting flows.

Discovery & Insights Engine Debut

In 2024, we developed our discovery engine, powered by what we call "reverse GPT," representing a fundamental shift in how workplace content is processed and understood. As we crawl through platforms like Slack, we don't just index and vectorize content – we intelligently classify it into meaningful categories such as process changes, announcements, escalations, releases, milestones, scrum notes, and meeting minutes.

The Insights app, launched in beta in 2024, was developed to treat workspace activity as a continuous stream of events. By using text-to-SQL technology, we enabled powerful queries across this event stream, extracting insights that aren't possible through traditional snapshot-based keyword or semantic searches. This allowed users to ask questions like:

• "What did I work on today?"
• "What happened in the workspace while I was gone last week?"
• "What Jiras did John work on?"
• "Show me all company announcements from the past 30 days."

Looking ahead to Insights 2.0 in mid-2025, we're exploring two promising approaches to enhance our querying capabilities:

• Multiple SQL-based phased queries
• Spreadsheet-based querying that replaces some layers of SQL

This evolution will transform Ayraa from a pure search platform into a powerful analytical tool, enabling users to derive meaningful insights from any email, CRM, ticketing system, or chat application by converting notable activities into queryable events. We've already seen strong results in specific use cases and are excited to expand these capabilities across our platform.

Collections: Building a Personal or Collaborative Second Brain

In early 2024, we introduced Collections, representing a significant advancement in how teams organize and share institutional knowledge. This technology was developed to enable teams to build a trusted, organized second brain that can be easily replicated and shared across the organization.

Collections was developed with a comprehensive approach to knowledge management:

• Flexible Content Integration: We built the ability for users to create collections through multiple sources - uploading official documents (Word, PDFs), linking trusted resources from supported platforms (Notion, Confluence, Google Docs), creating information cards, or directly adding bookmarked websites through drag-and-drop functionality
• Smart Search & Query Integration: We integrated Collections seamlessly with Ayraa's search capabilities, allowing users to constrain their queries to specific curated knowledge sets. This meant searches and assist queries could focus exclusively on verified, curated content rather than searching the entire workspace
• Enterprise-Ready Verification: We implemented an expert validation workflow where subject matter experts can verify content and timestamp its relevance, ensuring collections maintain their accuracy over time

The impact of Collections has been particularly notable across different organizational roles:

• Executive Teams: Streamlining access to strategic plans, OKRs, and compliance documentation
• Engineering & Product: Centralizing technical guides and project specifications to reduce repetitive questions
• HR Teams: Revolutionizing employee onboarding through organized training and policy documentation
• Sales Enablement: Providing 24/7 access to product documentation, pricing sheets, and competitive analysis

The power of Collections was demonstrated in its ability to transform individual expertise into shared organizational knowledge, ensuring teams have immediate access to accurate, current information without hunting through multiple platforms or channels.

Go Links: Simplifying Resource Access

In 2024, we developed Go Links to address a critical need we identified: simplifying access to frequently used enterprise resources. We built Go Links to transform complex URLs into human-friendly shortcuts anyone can remember and share.

The technology became foundational to workplace knowledge sharing, with teams across organizations adopting shortcuts like:

• "go/benefits" for HR's benefits documentation
• "go/sales-assets" for sales enablement resources
• "go/engineering" for technical processes and release docs
• "go/support" for customer service playbooks

This elegant solution to resource organization was developed to complement our broader search capabilities while providing quick access to frequently used resources. Teams could now organize and share their knowledge through intuitive, standardized shortcuts rather than wrestling with lengthy, complex URLs.

The Slack Co-pilot Experience

Throughout 2024, we focused on making Ayraa seamlessly accessible within Slack, where many users spend their workday. We evolved our Slack bot beyond simple search functionality to become a faithful workplace co-pilot, enabling users to harness Ayraa's full capabilities without context switching.

The impact was transformative for team productivity. We enabled users to mention Ayraa within any Slack conversation to instantly access workplace knowledge, turning the bot into a 24/7 consultant who provides contextual clarifications during ongoing discussions. This proved particularly valuable during decision-making processes, where teams could quickly reference historical context or related documentation.

We developed our thread summarization capability as a crucial time-saver - rather than reading through lengthy discussions, users could get key points or ask specific questions about the thread's content right where they work.

Moreover, we equipped the bot with the ability to detect and highlight stalled discussions, helping teams maintain momentum on essential conversations that might otherwise have been forgotten.

Perhaps most impactfully, we enhanced our discovery (Insights) engine to proactively bring essential updates to users through the Slack bot. By identifying and delivering crucial escalations, milestone achievements, and releases directly to relevant team members, we helped ensure that critical information doesn't get lost in the noise of busy workspaces.

Our Slack integration developments in 2024 included:

• Instant workplace search within Slack
• Thread summarization and querying
• 24/7 contextual workspace assistance
• Notifications of significant updates across your workspace
• Stalled discussion identification and follow-up

The Browser Co-pilot Experience

Early in 2024, we developed our browser extension, initially designed to help users rapidly summarize and search through lengthy web content like Wikipedia articles, blogs, and Reddit discussions.

Throughout the year, we evolved the extension into a faithful browser co-pilot. Built on standard HTML/CSS parsing capabilities, we enabled it to seamlessly process content from millions of websites. The technology proved particularly valuable for:

• Web research - allowing users to summarize/query pages and save content directly to their Collections
• PDF documents - extracting and analyzing text from vector-based documents without specialized readers
• YouTube videos - providing instant transcript access, summaries, and timestamp-based navigation

Quality at Scale: Building Trust Through Automation

In 2024, we invested heavily in automated quality systems to ensure reliability at scale. A cornerstone of this investment was our automated regression suite for Assist, which we developed to evaluate a comprehensive set of search queries and responses daily. This system was designed not only to catch potential regressions in search quality and answer accuracy but also to help measure improvements as we fine-tune our platform's weights and configurations.

To measure our RAG (Retrieval Augmented Generation) effectiveness, we integrated the open-source Tonic Validate tool into our testing framework. This integration provided quantifiable metrics for response quality, allowing us to make data-driven decisions about our search and assist features.

Enterprise-Grade Security Focus

Security remained core to our DNA, with our founding team's experience from McAfee. 2024 achievements included:

• SOC 2 Type 2 compliance across all five trust principles
• CASA Tier-2 certification from PWC
• Grade A (100% pass) in third-party penetration testing by Halo Security

Looking Forward

As we enter 2025, we're excited to continue innovating in enterprise search and knowledge discovery. Our commitment to democratizing access to workplace knowledge remains stronger than ever, supported by our technical innovations and user-centric approach.

This journey wouldn't be possible without our users' trust and feedback. We look forward to bringing even more innovations to workplace knowledge discovery in 2025.