Menlo Ventures just released an interesting report on The State of Generative AI in the Enterprise. The stand out number is that the enterprise market went from spending $2.3B on generative AI in 2023 to $13.8B in 2024 – an over 6x increase in a single year.
With enterprise embracing products built around AI, if you’ve been sitting on a product idea now is the time to jump into the market.
We’ll give you a quick run through on where enterprise is spending money on AI and follow that with the steps your Customer Success team will need to take to guarantee your customers embrace your product.
The AI-based products enterprise is spending money on
If you’re targeting the enterprise you have to ask yourself if you’re going cross-vertical or narrowing in on a specific vertical you have experience in.
If you look at the chart above you can “see search and retrieval” as well as “Data extraction and transformation”. They are both broad use cases. Do you build a cross-vertical solution for those use cases, or do you say “Let’s extract this data in X vertical and transform it to do Y”? (Personally, we’d prove the tech in the vertical and then move into other vertical from there.)
Cross-Vertical Products
Some AI-based products are naturally cross-vertical. Businesses are all built on the same tech and share common processes. Being the preferred supplier for universal processes like “math” (looking at you, Excel), or “workflow optimisation” or “coding” opens up a huge potential market.
Some of these broad tools include:
- Coding assistants: Tools like GitHub Copilot and Codeium speed up the boring, and sometimes repetitive nuts-and-bolts of programming: test cases, boilerplate, and debugging. They can shorten development cycles and allow for more ambitious projects. It is no wonder they are becoming indispensable for IT and engineering teams in any industry. Given that you can fork an open source code editor like Visual Studio Code, add some interface logic and custom prompts to create your own coding assistant, it is no wonder options in this area keep growing.
- Support Chatbots: There are hundreds of players in the chatbot field. But enterprise has special needs and is demanding. Platforms like Aisera and Cognigy help enterprises automate customer service 24/7, and do it by offering scalable solutions and deep feature sets.
- Enterprise Search + Retrieval: Products like Glean enable semantic knowledge searches across internal documents, CRMs, Slack channels, Google documents, databases – all of an enterprise’s data. And these products provide the fine-grained data access permissions enterprise requires on top of secure and regulatory compliant indexing and ingestion. Permissions, security and compliance – they’re the big challenges. Once that is done these systems are still just chat interfaces to RAG systems with bespoke prompts and AI workflows.
- Data Extraction and Transformation: This is quite similar to 3, and in that way Sana offers similar features to Glean. Data extraction is a form of search and retrieval, but AI, via RAG (and sometimes combined with traditional search tech), brings new ways to search and filter unstructured data, or turn messy data, like scans of handwritten forms, into accessible structured data. This “new” or newly accessible data can then be fed into decision-making processes or further AI applications.
- Meeting Summarisation Tools: That this category, which is Text-to-speech with smarts, ranks so highly speaks to how much time is lost in meetings. Fireflies.ai and Otter.ai are examples of this category. Not only do these products save people time, the summarisations they produce can be fed back into the Enterprise Search and Retrieval tooling so answers and decisions are globally accessible.
Vertical-Specific Products
Cross-vertical tools are the general solutions – toolkits like Glean and Sana where the effort and expense is in security and making them available to an entire enterprise.
For verticals, the products are specialised versions of the general solutions. Here, you need to know the challenges, the data, the reports, and the processes of the vertical. If you can recognise where you can leverage AI and understand how to integrate it and how to convince the stakeholders to integrate it, you’re on your way to a successful product.
Here are some verticals and the kind of AI-based products finding success. Note – this is the US market, which is quite different to the Australian market in how industries like healthcare and financial services operate. Also, all numbers are in $USD.
Healthcare ($500 million in AI enterprise spend):
- Ambient Scribes: Tools like Heidi or Eleos Health (or Lyrebird in Australia) automate medical documentation, reducing the administrative burden for healthcare practitioners.
- Clinical Lifecycle Automation: Solutions targeting triage, intake, coding, and revenue cycle management streamline intricate processes unique to healthcare workflows.
Legal ($350 million in AI enterprise spend):
- Contract Analysis Tools: Products like Harvey provide AI-assisted legal research, let lawyers “query” documents instead of scanning through them, assist in contract drafting much like they do in coding – filling in boilerplate, catching simple errors and types, and some forms of analysis.
- Litigation Support: Platforms such as Everlaw add AI-based search and retrieval to e-discovery processes and speed up trial preparation.
Financial Services ($100 million in AI enterprise spend):
- Risk Assessment and Compliance Tools: Products like Greenlite provide real-time compliance monitoring, while tools like Arkifi use LLMs to accelerate financial research.
- Back Office Automation: Solutions that automate complex reconciliation and reporting workflows can transform accounting processes.
Choosing and Developing Your Product Strategy
Let’s look at one type of product—Enterprise Search + Retrieval. It’s a good place to start as most AI products are built around search and retrieval.
We’ll walk through what a startup creating such a product needs to undertake to ensure its successfully adopted by customers.
Ensuring Integration and Delivering ROI
Your Customer Success strategy needs to ensure that enterprises can integrate your product into their workflows while realising enough measurable value. Here, loosely, are the steps you’re going to need to take.
Step 1: Strategic Alignment and Goal Setting
Before implementation starts you need to work with customers to align your product with their strategic priorities. For an enterprise search product, consider the following actions:
- Identify specific pain points (e.g., fragmented search across emails, messengers, or document repositories).
- Define KPIs such as search accuracy improvement (e.g., from 60% to 90%), query response times (e.g., reducing average time from 10 seconds to 3 seconds), or overall employee productivity gains.
- Ensure stakeholders across IT, HR, and knowledge management buy into the solution’s potential impact.
Step 2: Baseline Assessment
Establish a clear baseline to measure improvements once your product is being used:
- Measure existing search query success rates (e.g., percentage of queries resolved on the first attempt).
- Assess the average time employees spend searching for information daily or weekly.
For example:
- Current query success rate: 60%
- Average weekly search time per employee: 5 hours
These metrics will serve as comparison points post-deployment.
Step 3: Pilot Testing with Focused Use Cases
Run a targeted pilot with a single department or team before rolling out your product enterprise-wide. Keep the scope manageable with focused use cases—for example, deploying within HR for talent management or within IT for knowledge-base retrievals.
Actions to take:
- Collect real-time feedback from users about usability and result quality.
- Begin quantifying benefits such as time savings in accessing key documents.
- Look at how they use it and if there are opportunities for new, specific queries to support their processes.
Step 4: Integration and Scalability Design
Successful generative AI tools must integrate smoothly with existing systems (like Slack, Google Drive, or Microsoft Teams) while maintaining scalability for future needs. Particular attention should be paid to data privacy compliance when dealing with sensitive internal information.
Strategies for success include:
- Partnering with IT stakeholders to ensure seamless integration into existing data sources.
- Creating onboarding documentation or customer training guides to help teams maximise adoption.
- Provide experienced technical support that can assist directly with the integration.
Step 5: Real-Time Monitoring and Iteration
Once your product moves into enterprise-wide use, monitor key metrics and track performance against the benchmarks established during the baseline assessment phase.
Iterate based on feedback by optimising features like query suggestions or integrating additional data sources.
Step 6: Qualitative Impact Assessment
Beyond quantitative metrics, gather qualitative feedback through surveys and interviews. Assess how your product affects employees’ ability to collaborate or make decisions faster based on relevant insights.
Use feedback to refine features and add context-specific improvements.
Step 7: ROI Reporting and Stakeholder Communication
As results accrue, create clear reports that quantify both tangible (e.g., time savings) and intangible (e.g., employee satisfaction) benefits. Focus on linking these metrics back to strategic goals identified earlier.
Example ROI Report Structure:
- Baseline Metrics: Itemize pre-launch performance (e.g., search accuracy at 60%).
- Post-Go-Live Impact: Document achievements (e.g., accuracy improvement to 90%, time savings worth $500K annually).
- Customer Testimonials: Include quotes demonstrating how employees value the tool’s impact on their work.
- Long-Term Predictions: Highlight how ongoing adjustments will deliver scaling advantages across departments.
Step 7 is obviously the most important step, but you should know around step 3 or 4 how you are tracking on delivering value. That will give you opportunities to find and implement the targeted features and applications on top of your product that will win you the customer.
Now is the best time to start building on AI
As generative AI gains momentum in enterprise operations, startups have the opportunity to become embedded within enterprise workflows as the essential solutions providers.
As a startup, you can capture this momentum by either developing wide-reaching cross-vertical products like enterprise search or by addressing industry-specific needs in the vertical of your choosing.
But building a great product is only the first step. Ensuring your enterprise customers realise its potential is where the growth is. Enterprise wants every step of the implementation process to be focusing on delivering value. By investing in Customer Success, and taking all the necessary steps to reduce risk, like pilot testing and monitoring, you can lock customers in and position yourself as the go-to for the market.
