The Product (in brief)

Picture this

You’re shopping online on a popular fashion app. You start typing “black jacket Men...”. The app auto-suggests "....H&M.". Wow, a black jacket from H&M is exactly what you were looking for, you click the auto-suggested option, and the screen loads.... the excitement is palpable, and you see 186 results that match your search.

You’re eager to browse them all.

Andddddddddd... half of the jackets in those search results are not black, and none of the black jackets are from H&M.

Your shopping experience is ruined.

BUT

What if you had a search assistant that makes information discovery and fulfillment conversational across websites, e-commerce, customer service, and workplaces?

That is exactly what SearchAssist, by Kore.ai, does.

What does SearchAssist do?

SearchAssist helps deliver highly relevant, accurate, contextual, and personalized search results to all users. It uses the Kore.ai proprietary technology, Large Language Models (LLMs), and Generative AI technologies to deliver next-gen search experience to various users, including employees, consumers, customer support executives, and agents.

Some of the key features of a SearchAssist application include:

• It supports ingesting data in different formats from different sources, including web pages, files, and most commonly used third-party content management systems.
• It is highly customizable, allowing you to completely control and customize the application’s response per your business requirements.
• It can be seamlessly integrated into any of your existing applications to provide an engaging user experience.
• It can be easily integrated with the most common large language models for answer generation.

ICPs

The product is built with specific use cases in mind. That means one level of prioritization has been done, allowing us to target all the ICPs via the product use cases i.e, E-Commerce, Website Search, Contact Center, Workplace (employee case)

However, for the onboarding experience, we have applied a few parameters like adoption curve, Frequency of use case to our product's end-users i.e., customers, Employees, and contact center agents. The frequency of use case is low and adoption curve is high for the employee use case, when compared to other ICPs.

Therefore, for the onboarding and teardown exercise, we will broadly classify our ICPs/buyer personas as per the experiences i.e, customer experiences and the agent experiences

Most Used and Valued Features

Jobs To Be Done - For Buyer Persona

Product (and Process) Teardown

We will first look at the acquisition process teardown, and move on to the product itself. Currently, the following acquisition channels are being employed. We will divide them into two categories, i.e., Inbound and Outbound.

• Inbound
• • Through Website (prospects will contact via mail etc from website)
  • Through Analyst Reports (Process driven and controlled by the analysts. no scope to be creative) ❌
  • RFI / RFP / RFQ (part of the sales funnel)

• Outbound
• • Email (by BDRs/AEs) - Part of the sales cycle
  • Industry Events / Summits (Company level, Organizer Defined) ❌
  • Leadership Referrals (highly personal, not defined by a set process) - ❌

For the process teardown, we will only consider those processes that account for majority od deals, are controllable, and where we have visibility. Kore.ai does not do paid ads for SearchAssist specifically. Therefore, we'll broadly classify and look at

• The sales cycle - once the prospect enters the funnel
• The website landing page and the product teardown

The Sales Cycle:

Before we dive into the tasks that happen for each deal stage, it should be noted that the BDRs, and SEs are provided with necessary access to prospecting tools, competitive comparison documents, battlecards, pricing tools, and other enablement processes. The process of sales enablement, while not perfect at this stage, is continuously improving, and must be analysed in that manner.

5% - 10%:

Outbound:

In outbound scenarios, the sales cycle begins with the BDRs sending out outbound emails to the list of clients segregated by the priority levels. For top priority categories, its the account executives (AE) that reach out directly.

Currently, for the SearchAssist product, there are no separate mail templates that are being used. Each mail that is sent out by BDRs pitches either Kore.ai technology as a whole, or the industry-specific solution modules wherever appropriate. Other than that, there is no product-specific communication for SearchAssist. This needs to be improved, and there needs to product specific, and also use case specific mail templates for different prospects.

Inbound:

In the inbound scenario, the prospects conduct their research, rely primarily via analyst reports, or their secondary research, and ultimately reach out to our teams via mail provided on our website.

Once received, a BDR is assigned to the prospect, and they address the prospect, and also share additional materials for them to go through.

Next Steps:

Ultimately, in both inbound and outbound scenarios, once the initial interaction is completed via mail, the BDRs conduct an initial qualification meeting with the prospects where the main objective is to ascertain if there is an actual need, pain point to solve for the customer, and if so, is any of our product suitable to meet and exceed the customer requirements. Once this is ascertained, the deal is moved to 10% and the respective account executive for that industry is assigned to the opportunity and same is updated on the HubSpot.

The AE then has another call where they do a deep dive discovery, understand the Budget, Authority, the exact need, and if there is a realistic Timeline to work with. Once determined, the deal is moved to 20%, and a solution engineer (SE) is assigned. Kore.ai has a dedicated team of highly experienced SEs.

20% - 30%:

The SE and the AE discuss internally to prepare a strategy for that account, and they have another discussion with the client, along with the client's broader team. This discussion's purpose is to conduct a technical discovery and showcase an initial standard demo of the solution. Within this call, they also try to understand the evaluation metrics, the decision criteria, process and see if they can identify a champion at this stage.

In this stage, the client typically chooses to either have a Prood of Concept stage or if its a large enterprise, they choose the RFP route, where they then finalize top 3 or 2 participants, and have a PoC conducted.

Kore.ai has a dedicated proposal team, and is shortlisted in the majority of the RFP projects. Once the shortlist is done, the client will request a PoC or a customized demo request based on the use case they will provide. To support the SEs, Kore.ai has a dedicated demo team, that analyses the use cases provided, and builds the demo by working closely with SEs.

The prepared demo is then presented to the client by the SEs on site or on call as agreed. Further iterations, if any, are also accommodated, and are showcased to the customer.

40% - 60%:

Once the PoC or custom demo is completed, the clients provide a confirmation if they are satisfied that the product meets their requirements or not. If they confirm, the deal is moved to the 'Technical Win' stage, and an initial commercial proposal, including the pricing, implementation approach, and anpartner introduction, is done.

If the client is not happy with the demo, they duly communicate that they don't intend to move forward as the product does not meet their technical requirements or expectations.

60% - 75%:

Once the pricing (inclusive of license and implementation) is submitted, several rounds of price negotiations take place, and both parties try to come to an agreement on the BAFO (Best and Final Offer)

If the agreement is reached, the parties start the drafting of contractual documents, and prepare for the contract review. If they do not reach an agreement, the prospect moves on to the second best vendor in the competition.

Please note that this may or may not result in our company being out of the race, but definitely the deal has ran out of the steam. There is always a possibility of the customer coming back, and a revised offer is put on the table

75-100%:

Once the contract documents are drafted, its time for the legal teams to step in. Both parties negotiate the contractual the obligations, and try to arrive at a mutual agreement.

Once the mutual agreement is reached, the information security review is initiated, and the paperwork is completed, and the documents are signed by both the parties.

Once the documents are signed, the sales leadership reviews all the deal related documents to check if they are in place, and only then the deal is marked as Closed and Won

Website and Product Teardown:

Please click the following link or the PDF to view the product teardown slides. You have the comment access.

CLICK TO VIEW PRODUCT AND WEBSITE TEARDOWN

Activation Metrics:

Metrics to Track:

• Total New/Unique Users - Monthly Tracking
• Repeat Users
• Identified and Unidentified User Intents - This helps us understand how well the app is performing or needs training.
• Response TIme
• Agent Transfer Rate - To understand which intents are not being resolved via self-service, and the potential reasons
• CSAT (customer feedback) - Helps us understand the app performance
• Total no. of Queries: Weekly filter. To understand user adoption
• Click Through rate - To understand the relevancy of results
• Search Appearance - To understand which content from the knowledge base is most used