Customer Relationship/Success Management is a business strategy and organization-wide process focused on winning, growing, and retaining customers to increase value for both the customer and the company.

Already in the 2010s CRM had become a function nightmare for many large companies. Some argued that failing CRM deployments were a people problem, not a software problem. Companies implementing CRMs were advised to hire better administrators to discipline their salespeople so clients could love their system, whichever this might have been, “higher end (Salesforce.com, Oracle, SAP), mid-range (Microsoft Dynamics CRM, Sage CRM, GoldMine) or lower end (Zoho, Nimble, Highrise, Insight.ly).”

At the core of companies’ strategies for over half a century was building a relationship with clients, knowing them, and building a super-software-structure to avoid rolling the digital Rolodex that had accumulated over the years.

SaaS companies began selling these customer care systems, but implementation required the onboarding of employees whose primary job was to walk other employees through software the company was going to use for customer success. Hence, another separate organizational branch.

For the CEO of Berry , Yan Fu, himself a software engineer and MBA, that was the clear gap AI could solve. Fu had spent years in product and engineering in multinationals like Sony and had at least two startups where this was or could become a problem where more and more customer success managers and employees would need to be hired for growth.

Berry, a Y Combinator-backed company founded in 2023, built a system that ends this disconnection. Now its AI agents don’t just log the customer relationship – they run it. They detect when users are stuck, intervene in real time, handle onboarding, drive adoption, and manage renewals.

How and When This Started

For most of the twentieth century, companies were organized around products. You made something, you shipped it, you moved on. Marketing meant getting the message out to the largest possible audience in the most efficient way. The customer, in practical terms, was an abstraction – a market share number, a demographic segment, a sales forecast.

For the salesperson who actually knew the customer, the memory lived in a Rolodex on their desk. Most of what this person knew of the company’s clients would stay invisible to everyone else in the organization.

Post World War II, large corporations contracted software packages to administer clients and paid upfront perpetual license fees and came with a manual and perhaps a help desk number.

When competition intensified, however, companies noticed that the person at the other end of the sale had specific needs, and that treating those needs as interchangeable was leaving money on the table.

Recognizing clients’ specific needs would enable companies to outcompete those that couldn’t. This is where what we now call CRM was born: not as software, but as a management philosophy built on the premise that decisions about a customer should be made at the point of contact, not passed up the chain of command, and later, diligently maintained and cared for.

And technology continued to adapt to ensure millions of customers could be seen and understood individually.

The Early Wins Became Later Losses

In the 1990s, large corporations , namely banks, retailers, and airlines, invested in tools they needed to learn about their clientele.

In the early 1990s, First Union Bank built an internal system called Einstein. It did something simple in principle and difficult in practice: it gave every front-line worker, in real time, a complete view of a customer’s entire relationship with the bank and their overall profitability. A teller could make an individualized decision such as waive a fee, extend an offer, resolve a complaint without escalating it.

By the late 1990s, the web provided the necessary space to expand. American Airlines launched the largest personalization effort online at the time, using its databases to build custom web pages on the fly for its two million frequent flyers. Instead of showing every visitor the same marketing, the site tailored what each traveler saw.

The challenge then became the huge labor demand these business units required. For twenty years, headcount was how software companies signaled seriousness, as Fu mentioned when we met earlier in 2026 for an interview.

Soon, the 2000s brought the SaaS wave, where software companies could deploy to care for customers to keep them coming back. For years, a SaaS like Salesforce required a separate organization to onboard those managing the software so that subscriptions continued to be renewed.

Is The 21 st Century The Time For Tiny Teams in CRM?

As Fu explained, his startup set out in 2023 to build a system that could cover the entire lifecycle of a customer. Even in Berry’s early days, many pieces were missing, but the release of OpenAI’s models changed everything.

The impact: a SaaS company with ten thousand customers in 2030 may not need a customer success department at all.

This shift creates a slimmer type of organization in the sector. There is minimal communication overhead within the team members, yet greater control over how customer success is defined. The team at Berry (US-headquartered) is small, “Now we’re only 10 people!” Fu noted.

The composition of Berry reflects this technical focus and shift, with 80% to 90% of the company being engineers. In 2026, purposely engineered AI agents can identify needs, frustrations, and signals of intent that were once lost across layers of human management.

Perhaps the issue was, indeed, a software problem after all, just one that does much more than simply remember names and backgrounds.